lectures/09_classification-metrics.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "![](img/330-banner.png)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "-"
    }
   },
   "source": [
    "# Lecture 9: Classification Metrics\n",
    "\n",
    "UBC 2022-23\n",
    "\n",
    "Instructor: Varada Kolhatkar"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Imports, Announcements, and LO"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "skip"
    }
   },
   "source": [
    "### Imports"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "slideshow": {
     "slide_type": "skip"
    }
   },
   "outputs": [],
   "source": [
    "import os\n",
    "import sys\n",
    "\n",
    "sys.path.append(\"code/.\")\n",
    "\n",
    "import IPython\n",
    "import matplotlib.pyplot as plt\n",
    "import mglearn\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "from IPython.display import HTML, display\n",
    "from plotting_functions import *\n",
    "from sklearn.dummy import DummyClassifier\n",
    "from sklearn.linear_model import LogisticRegression\n",
    "from sklearn.model_selection import cross_val_score, cross_validate, train_test_split\n",
    "from sklearn.pipeline import Pipeline, make_pipeline\n",
    "from sklearn.preprocessing import StandardScaler\n",
    "from utils import *\n",
    "\n",
    "%matplotlib inline\n",
    "pd.set_option(\"display.max_colwidth\", 200)\n",
    "\n",
    "from IPython.display import Image"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "slideshow": {
     "slide_type": "skip"
    }
   },
   "outputs": [],
   "source": [
    "# Changing global matplotlib settings for confusion matrix.\n",
    "plt.rcParams[\"xtick.labelsize\"] = 18\n",
    "plt.rcParams[\"ytick.labelsize\"] = 18"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "### Announcements\n",
    "\n",
    "- Autograder problems in Lab 2 and Lab 3: I'll be looking at autograding results of hw2 later this week.  \n",
    "- HW4 is released.\n",
    "- Please make use of OH and tutorials to interact with the teaching team and ask your questions.  \n",
    "- Midterm is coming up soon (Oct 27th during class time)."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "skip"
    }
   },
   "source": [
    "### Learning outcomes \n",
    "\n",
    "From this lecture, students are expected to be able to:\n",
    "\n",
    "- Explain why accuracy is not always the best metric in ML.\n",
    "- Explain components of a confusion matrix. \n",
    "- Define precision, recall, and f1-score and use them to evaluate different classifiers. \n",
    "- Broadly explain macro-average, weighted average.\n",
    "- Interpret and use precision-recall curves. \n",
    "- Explain average precision score.\n",
    "- Interpret and use ROC curves and ROC AUC using `scikit-learn`.  \n",
    "- Identify whether there is class imbalance and whether you need to deal with it.\n",
    "- Explain and use `class_weight` to deal with data imbalance."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "skip"
    }
   },
   "source": [
    "<br><br><br><br>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "## Evaluation metrics for binary classification: Motivation "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "### Dataset for demonstration \n",
    "\n",
    "- Let's classify fraudulent and non-fraudulent transactions using Kaggle's [Credit Card Fraud Detection](https://www.kaggle.com/mlg-ulb/creditcardfraud) data set."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "slideshow": {
     "slide_type": "-"
    }
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Time</th>\n",
       "      <th>V1</th>\n",
       "      <th>V2</th>\n",
       "      <th>V3</th>\n",
       "      <th>V4</th>\n",
       "      <th>V5</th>\n",
       "      <th>V6</th>\n",
       "      <th>V7</th>\n",
       "      <th>V8</th>\n",
       "      <th>V9</th>\n",
       "      <th>...</th>\n",
       "      <th>V21</th>\n",
       "      <th>V22</th>\n",
       "      <th>V23</th>\n",
       "      <th>V24</th>\n",
       "      <th>V25</th>\n",
       "      <th>V26</th>\n",
       "      <th>V27</th>\n",
       "      <th>V28</th>\n",
       "      <th>Amount</th>\n",
       "      <th>Class</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>64454</th>\n",
       "      <td>51150.0</td>\n",
       "      <td>-3.538816</td>\n",
       "      <td>3.481893</td>\n",
       "      <td>-1.827130</td>\n",
       "      <td>-0.573050</td>\n",
       "      <td>2.644106</td>\n",
       "      <td>-0.340988</td>\n",
       "      <td>2.102135</td>\n",
       "      <td>-2.939006</td>\n",
       "      <td>2.578654</td>\n",
       "      <td>...</td>\n",
       "      <td>0.530978</td>\n",
       "      <td>-0.860677</td>\n",
       "      <td>-0.201810</td>\n",
       "      <td>-1.719747</td>\n",
       "      <td>0.729143</td>\n",
       "      <td>-0.547993</td>\n",
       "      <td>-0.023636</td>\n",
       "      <td>-0.454966</td>\n",
       "      <td>1.00</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>37906</th>\n",
       "      <td>39163.0</td>\n",
       "      <td>-0.363913</td>\n",
       "      <td>0.853399</td>\n",
       "      <td>1.648195</td>\n",
       "      <td>1.118934</td>\n",
       "      <td>0.100882</td>\n",
       "      <td>0.423852</td>\n",
       "      <td>0.472790</td>\n",
       "      <td>-0.972440</td>\n",
       "      <td>0.033833</td>\n",
       "      <td>...</td>\n",
       "      <td>0.687055</td>\n",
       "      <td>-0.094586</td>\n",
       "      <td>0.121531</td>\n",
       "      <td>0.146830</td>\n",
       "      <td>-0.944092</td>\n",
       "      <td>-0.558564</td>\n",
       "      <td>-0.186814</td>\n",
       "      <td>-0.257103</td>\n",
       "      <td>18.49</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>79378</th>\n",
       "      <td>57994.0</td>\n",
       "      <td>1.193021</td>\n",
       "      <td>-0.136714</td>\n",
       "      <td>0.622612</td>\n",
       "      <td>0.780864</td>\n",
       "      <td>-0.823511</td>\n",
       "      <td>-0.706444</td>\n",
       "      <td>-0.206073</td>\n",
       "      <td>-0.016918</td>\n",
       "      <td>0.781531</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.310405</td>\n",
       "      <td>-0.842028</td>\n",
       "      <td>0.085477</td>\n",
       "      <td>0.366005</td>\n",
       "      <td>0.254443</td>\n",
       "      <td>0.290002</td>\n",
       "      <td>-0.036764</td>\n",
       "      <td>0.015039</td>\n",
       "      <td>23.74</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>245686</th>\n",
       "      <td>152859.0</td>\n",
       "      <td>1.604032</td>\n",
       "      <td>-0.808208</td>\n",
       "      <td>-1.594982</td>\n",
       "      <td>0.200475</td>\n",
       "      <td>0.502985</td>\n",
       "      <td>0.832370</td>\n",
       "      <td>-0.034071</td>\n",
       "      <td>0.234040</td>\n",
       "      <td>0.550616</td>\n",
       "      <td>...</td>\n",
       "      <td>0.519029</td>\n",
       "      <td>1.429217</td>\n",
       "      <td>-0.139322</td>\n",
       "      <td>-1.293663</td>\n",
       "      <td>0.037785</td>\n",
       "      <td>0.061206</td>\n",
       "      <td>0.005387</td>\n",
       "      <td>-0.057296</td>\n",
       "      <td>156.52</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>60943</th>\n",
       "      <td>49575.0</td>\n",
       "      <td>-2.669614</td>\n",
       "      <td>-2.734385</td>\n",
       "      <td>0.662450</td>\n",
       "      <td>-0.059077</td>\n",
       "      <td>3.346850</td>\n",
       "      <td>-2.549682</td>\n",
       "      <td>-1.430571</td>\n",
       "      <td>-0.118450</td>\n",
       "      <td>0.469383</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.228329</td>\n",
       "      <td>-0.370643</td>\n",
       "      <td>-0.211544</td>\n",
       "      <td>-0.300837</td>\n",
       "      <td>-1.174590</td>\n",
       "      <td>0.573818</td>\n",
       "      <td>0.388023</td>\n",
       "      <td>0.161782</td>\n",
       "      <td>57.50</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 31 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "            Time        V1        V2        V3        V4        V5        V6  \\\n",
       "64454    51150.0 -3.538816  3.481893 -1.827130 -0.573050  2.644106 -0.340988   \n",
       "37906    39163.0 -0.363913  0.853399  1.648195  1.118934  0.100882  0.423852   \n",
       "79378    57994.0  1.193021 -0.136714  0.622612  0.780864 -0.823511 -0.706444   \n",
       "245686  152859.0  1.604032 -0.808208 -1.594982  0.200475  0.502985  0.832370   \n",
       "60943    49575.0 -2.669614 -2.734385  0.662450 -0.059077  3.346850 -2.549682   \n",
       "\n",
       "              V7        V8        V9  ...       V21       V22       V23  \\\n",
       "64454   2.102135 -2.939006  2.578654  ...  0.530978 -0.860677 -0.201810   \n",
       "37906   0.472790 -0.972440  0.033833  ...  0.687055 -0.094586  0.121531   \n",
       "79378  -0.206073 -0.016918  0.781531  ... -0.310405 -0.842028  0.085477   \n",
       "245686 -0.034071  0.234040  0.550616  ...  0.519029  1.429217 -0.139322   \n",
       "60943  -1.430571 -0.118450  0.469383  ... -0.228329 -0.370643 -0.211544   \n",
       "\n",
       "             V24       V25       V26       V27       V28  Amount  Class  \n",
       "64454  -1.719747  0.729143 -0.547993 -0.023636 -0.454966    1.00      0  \n",
       "37906   0.146830 -0.944092 -0.558564 -0.186814 -0.257103   18.49      0  \n",
       "79378   0.366005  0.254443  0.290002 -0.036764  0.015039   23.74      0  \n",
       "245686 -1.293663  0.037785  0.061206  0.005387 -0.057296  156.52      0  \n",
       "60943  -0.300837 -1.174590  0.573818  0.388023  0.161782   57.50      0  \n",
       "\n",
       "[5 rows x 31 columns]"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "cc_df = pd.read_csv(\"data/creditcard.csv\", encoding=\"latin-1\")\n",
    "train_df, test_df = train_test_split(cc_df, test_size=0.3, random_state=111)\n",
    "train_df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(199364, 31)"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_df.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "- Good size dataset \n",
    "- For confidentially reasons, it only provides transformed features with PCA, which is a popular dimensionality reduction technique. "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "### EDA"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "slideshow": {
     "slide_type": "-"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "Int64Index: 199364 entries, 64454 to 129900\n",
      "Data columns (total 31 columns):\n",
      " #   Column  Non-Null Count   Dtype  \n",
      "---  ------  --------------   -----  \n",
      " 0   Time    199364 non-null  float64\n",
      " 1   V1      199364 non-null  float64\n",
      " 2   V2      199364 non-null  float64\n",
      " 3   V3      199364 non-null  float64\n",
      " 4   V4      199364 non-null  float64\n",
      " 5   V5      199364 non-null  float64\n",
      " 6   V6      199364 non-null  float64\n",
      " 7   V7      199364 non-null  float64\n",
      " 8   V8      199364 non-null  float64\n",
      " 9   V9      199364 non-null  float64\n",
      " 10  V10     199364 non-null  float64\n",
      " 11  V11     199364 non-null  float64\n",
      " 12  V12     199364 non-null  float64\n",
      " 13  V13     199364 non-null  float64\n",
      " 14  V14     199364 non-null  float64\n",
      " 15  V15     199364 non-null  float64\n",
      " 16  V16     199364 non-null  float64\n",
      " 17  V17     199364 non-null  float64\n",
      " 18  V18     199364 non-null  float64\n",
      " 19  V19     199364 non-null  float64\n",
      " 20  V20     199364 non-null  float64\n",
      " 21  V21     199364 non-null  float64\n",
      " 22  V22     199364 non-null  float64\n",
      " 23  V23     199364 non-null  float64\n",
      " 24  V24     199364 non-null  float64\n",
      " 25  V25     199364 non-null  float64\n",
      " 26  V26     199364 non-null  float64\n",
      " 27  V27     199364 non-null  float64\n",
      " 28  V28     199364 non-null  float64\n",
      " 29  Amount  199364 non-null  float64\n",
      " 30  Class   199364 non-null  int64  \n",
      "dtypes: float64(30), int64(1)\n",
      "memory usage: 48.7 MB\n"
     ]
    }
   ],
   "source": [
    "train_df.info()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Time</th>\n",
       "      <th>V1</th>\n",
       "      <th>V2</th>\n",
       "      <th>V3</th>\n",
       "      <th>V4</th>\n",
       "      <th>V5</th>\n",
       "      <th>V6</th>\n",
       "      <th>V7</th>\n",
       "      <th>V8</th>\n",
       "      <th>V9</th>\n",
       "      <th>...</th>\n",
       "      <th>V21</th>\n",
       "      <th>V22</th>\n",
       "      <th>V23</th>\n",
       "      <th>V24</th>\n",
       "      <th>V25</th>\n",
       "      <th>V26</th>\n",
       "      <th>V27</th>\n",
       "      <th>V28</th>\n",
       "      <th>Amount</th>\n",
       "      <th>Class</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>199364.000000</td>\n",
       "      <td>199364.000000</td>\n",
       "      <td>199364.000000</td>\n",
       "      <td>199364.000000</td>\n",
       "      <td>199364.000000</td>\n",
       "      <td>199364.000000</td>\n",
       "      <td>199364.000000</td>\n",
       "      <td>199364.000000</td>\n",
       "      <td>199364.000000</td>\n",
       "      <td>199364.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>199364.000000</td>\n",
       "      <td>199364.000000</td>\n",
       "      <td>199364.000000</td>\n",
       "      <td>199364.000000</td>\n",
       "      <td>199364.000000</td>\n",
       "      <td>199364.000000</td>\n",
       "      <td>199364.000000</td>\n",
       "      <td>199364.000000</td>\n",
       "      <td>199364.000000</td>\n",
       "      <td>199364.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>94888.815669</td>\n",
       "      <td>0.000492</td>\n",
       "      <td>-0.000726</td>\n",
       "      <td>0.000927</td>\n",
       "      <td>0.000630</td>\n",
       "      <td>0.000036</td>\n",
       "      <td>0.000011</td>\n",
       "      <td>-0.001286</td>\n",
       "      <td>-0.002889</td>\n",
       "      <td>-0.000891</td>\n",
       "      <td>...</td>\n",
       "      <td>0.001205</td>\n",
       "      <td>0.000155</td>\n",
       "      <td>-0.000198</td>\n",
       "      <td>0.000113</td>\n",
       "      <td>0.000235</td>\n",
       "      <td>0.000312</td>\n",
       "      <td>-0.000366</td>\n",
       "      <td>0.000227</td>\n",
       "      <td>88.164679</td>\n",
       "      <td>0.001700</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>47491.435489</td>\n",
       "      <td>1.959870</td>\n",
       "      <td>1.645519</td>\n",
       "      <td>1.505335</td>\n",
       "      <td>1.413958</td>\n",
       "      <td>1.361718</td>\n",
       "      <td>1.327188</td>\n",
       "      <td>1.210001</td>\n",
       "      <td>1.214852</td>\n",
       "      <td>1.096927</td>\n",
       "      <td>...</td>\n",
       "      <td>0.748510</td>\n",
       "      <td>0.726634</td>\n",
       "      <td>0.628139</td>\n",
       "      <td>0.605060</td>\n",
       "      <td>0.520857</td>\n",
       "      <td>0.481960</td>\n",
       "      <td>0.401541</td>\n",
       "      <td>0.333139</td>\n",
       "      <td>238.925768</td>\n",
       "      <td>0.041201</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>-56.407510</td>\n",
       "      <td>-72.715728</td>\n",
       "      <td>-31.813586</td>\n",
       "      <td>-5.683171</td>\n",
       "      <td>-42.147898</td>\n",
       "      <td>-26.160506</td>\n",
       "      <td>-43.557242</td>\n",
       "      <td>-73.216718</td>\n",
       "      <td>-13.320155</td>\n",
       "      <td>...</td>\n",
       "      <td>-34.830382</td>\n",
       "      <td>-8.887017</td>\n",
       "      <td>-44.807735</td>\n",
       "      <td>-2.824849</td>\n",
       "      <td>-10.295397</td>\n",
       "      <td>-2.241620</td>\n",
       "      <td>-22.565679</td>\n",
       "      <td>-11.710896</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>54240.000000</td>\n",
       "      <td>-0.918124</td>\n",
       "      <td>-0.600193</td>\n",
       "      <td>-0.892476</td>\n",
       "      <td>-0.847178</td>\n",
       "      <td>-0.691241</td>\n",
       "      <td>-0.768512</td>\n",
       "      <td>-0.553979</td>\n",
       "      <td>-0.209746</td>\n",
       "      <td>-0.642965</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.227836</td>\n",
       "      <td>-0.541795</td>\n",
       "      <td>-0.162330</td>\n",
       "      <td>-0.354604</td>\n",
       "      <td>-0.317761</td>\n",
       "      <td>-0.326730</td>\n",
       "      <td>-0.070929</td>\n",
       "      <td>-0.052819</td>\n",
       "      <td>5.640000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>84772.500000</td>\n",
       "      <td>0.018854</td>\n",
       "      <td>0.065463</td>\n",
       "      <td>0.179080</td>\n",
       "      <td>-0.019531</td>\n",
       "      <td>-0.056703</td>\n",
       "      <td>-0.275290</td>\n",
       "      <td>0.040497</td>\n",
       "      <td>0.022039</td>\n",
       "      <td>-0.052607</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.029146</td>\n",
       "      <td>0.007666</td>\n",
       "      <td>-0.011678</td>\n",
       "      <td>0.041031</td>\n",
       "      <td>0.016587</td>\n",
       "      <td>-0.052790</td>\n",
       "      <td>0.001239</td>\n",
       "      <td>0.011234</td>\n",
       "      <td>22.000000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>139349.250000</td>\n",
       "      <td>1.315630</td>\n",
       "      <td>0.803617</td>\n",
       "      <td>1.028023</td>\n",
       "      <td>0.744201</td>\n",
       "      <td>0.610407</td>\n",
       "      <td>0.399827</td>\n",
       "      <td>0.570449</td>\n",
       "      <td>0.327408</td>\n",
       "      <td>0.597326</td>\n",
       "      <td>...</td>\n",
       "      <td>0.186899</td>\n",
       "      <td>0.529210</td>\n",
       "      <td>0.146809</td>\n",
       "      <td>0.439209</td>\n",
       "      <td>0.351366</td>\n",
       "      <td>0.242169</td>\n",
       "      <td>0.090453</td>\n",
       "      <td>0.078052</td>\n",
       "      <td>77.150000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>172792.000000</td>\n",
       "      <td>2.451888</td>\n",
       "      <td>22.057729</td>\n",
       "      <td>9.382558</td>\n",
       "      <td>16.491217</td>\n",
       "      <td>34.801666</td>\n",
       "      <td>23.917837</td>\n",
       "      <td>44.054461</td>\n",
       "      <td>19.587773</td>\n",
       "      <td>15.594995</td>\n",
       "      <td>...</td>\n",
       "      <td>27.202839</td>\n",
       "      <td>10.503090</td>\n",
       "      <td>22.083545</td>\n",
       "      <td>4.022866</td>\n",
       "      <td>6.070850</td>\n",
       "      <td>3.517346</td>\n",
       "      <td>12.152401</td>\n",
       "      <td>33.847808</td>\n",
       "      <td>11898.090000</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>8 rows × 31 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                Time             V1             V2             V3  \\\n",
       "count  199364.000000  199364.000000  199364.000000  199364.000000   \n",
       "mean    94888.815669       0.000492      -0.000726       0.000927   \n",
       "std     47491.435489       1.959870       1.645519       1.505335   \n",
       "min         0.000000     -56.407510     -72.715728     -31.813586   \n",
       "25%     54240.000000      -0.918124      -0.600193      -0.892476   \n",
       "50%     84772.500000       0.018854       0.065463       0.179080   \n",
       "75%    139349.250000       1.315630       0.803617       1.028023   \n",
       "max    172792.000000       2.451888      22.057729       9.382558   \n",
       "\n",
       "                  V4             V5             V6             V7  \\\n",
       "count  199364.000000  199364.000000  199364.000000  199364.000000   \n",
       "mean        0.000630       0.000036       0.000011      -0.001286   \n",
       "std         1.413958       1.361718       1.327188       1.210001   \n",
       "min        -5.683171     -42.147898     -26.160506     -43.557242   \n",
       "25%        -0.847178      -0.691241      -0.768512      -0.553979   \n",
       "50%        -0.019531      -0.056703      -0.275290       0.040497   \n",
       "75%         0.744201       0.610407       0.399827       0.570449   \n",
       "max        16.491217      34.801666      23.917837      44.054461   \n",
       "\n",
       "                  V8             V9  ...            V21            V22  \\\n",
       "count  199364.000000  199364.000000  ...  199364.000000  199364.000000   \n",
       "mean       -0.002889      -0.000891  ...       0.001205       0.000155   \n",
       "std         1.214852       1.096927  ...       0.748510       0.726634   \n",
       "min       -73.216718     -13.320155  ...     -34.830382      -8.887017   \n",
       "25%        -0.209746      -0.642965  ...      -0.227836      -0.541795   \n",
       "50%         0.022039      -0.052607  ...      -0.029146       0.007666   \n",
       "75%         0.327408       0.597326  ...       0.186899       0.529210   \n",
       "max        19.587773      15.594995  ...      27.202839      10.503090   \n",
       "\n",
       "                 V23            V24            V25            V26  \\\n",
       "count  199364.000000  199364.000000  199364.000000  199364.000000   \n",
       "mean       -0.000198       0.000113       0.000235       0.000312   \n",
       "std         0.628139       0.605060       0.520857       0.481960   \n",
       "min       -44.807735      -2.824849     -10.295397      -2.241620   \n",
       "25%        -0.162330      -0.354604      -0.317761      -0.326730   \n",
       "50%        -0.011678       0.041031       0.016587      -0.052790   \n",
       "75%         0.146809       0.439209       0.351366       0.242169   \n",
       "max        22.083545       4.022866       6.070850       3.517346   \n",
       "\n",
       "                 V27            V28         Amount          Class  \n",
       "count  199364.000000  199364.000000  199364.000000  199364.000000  \n",
       "mean       -0.000366       0.000227      88.164679       0.001700  \n",
       "std         0.401541       0.333139     238.925768       0.041201  \n",
       "min       -22.565679     -11.710896       0.000000       0.000000  \n",
       "25%        -0.070929      -0.052819       5.640000       0.000000  \n",
       "50%         0.001239       0.011234      22.000000       0.000000  \n",
       "75%         0.090453       0.078052      77.150000       0.000000  \n",
       "max        12.152401      33.847808   11898.090000       1.000000  \n",
       "\n",
       "[8 rows x 31 columns]"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_df.describe(include=\"all\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "-"
    }
   },
   "source": [
    "- We do not have categorical features. All features are numeric. \n",
    "- We have to be careful about the `Time` and `Amount` features. \n",
    "- We could scale `Amount`. \n",
    "- Do we want to scale time?\n",
    "    - In this lecture, we'll just drop the Time feature. \n",
    "    - We'll learn about time series briefly later in the course. "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "Let's separate `X` and `y` for train and test splits."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "slideshow": {
     "slide_type": "-"
    }
   },
   "outputs": [],
   "source": [
    "X_train_big, y_train_big = train_df.drop(columns=[\"Class\", \"Time\"]), train_df[\"Class\"]\n",
    "X_test, y_test = test_df.drop(columns=[\"Class\", \"Time\"]), test_df[\"Class\"]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "- It's easier to demonstrate evaluation metrics using an explicit validation set instead of using cross-validation. \n",
    "- So let's create a validation set. \n",
    "- Our data is large enough so it shouldn't be a problem. \n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train, X_valid, y_train, y_valid = train_test_split(\n",
    "    X_train_big, y_train_big, test_size=0.3, random_state=123\n",
    ")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "### Baseline"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "fit_time       0.010542\n",
       "score_time     0.001104\n",
       "test_score     0.998302\n",
       "train_score    0.998302\n",
       "dtype: float64"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "dummy = DummyClassifier()\n",
    "pd.DataFrame(cross_validate(dummy, X_train, y_train, return_train_score=True)).mean()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "### Observations \n",
    "\n",
    "- `DummyClassifier` is getting 0.998 cross-validation accuracy!! \n",
    "- Should we be happy with this accuracy and deploy this `DummyClassifier` model for fraud detection? "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "-"
    }
   },
   "source": [
    "What's the class distribution? "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "slideshow": {
     "slide_type": "-"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0    0.9983\n",
       "1    0.0017\n",
       "Name: Class, dtype: float64"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_df[\"Class\"].value_counts(normalize=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "- We have class imbalance. \n",
    "- We have MANY non-fraud transactions and only a handful of fraud transactions. \n",
    "- So in the training set, `most_frequent` strategy is labeling 199,025 (99.83%) instances correctly and only 339 (0.17%) instances incorrectly. \n",
    "- Is this what we want? \n",
    "- The \"fraud\" class is the important class that we want to spot. "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "Let's scale the features and try `LogisticRegression`.   "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "fit_time       0.397143\n",
       "score_time     0.003884\n",
       "test_score     0.999176\n",
       "train_score    0.999235\n",
       "dtype: float64"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pipe = make_pipeline(StandardScaler(), LogisticRegression())\n",
    "pd.DataFrame(cross_validate(pipe, X_train, y_train, return_train_score=True)).mean()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "-"
    }
   },
   "source": [
    "- We are getting a slightly better score with logistic regression.  \n",
    "- What score should be considered an acceptable score here? \n",
    "- Are we actually spotting any \"fraud\" transactions? "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "- `.score` by default returns accuracy which is \n",
    "$$\\frac{\\text{correct predictions}}{\\text{total examples}}$$\n",
    "- Is accuracy a good metric here? \n",
    "- Is there anything more informative than accuracy that we can use here? \n",
    "\n",
    "Let's dig a little deeper."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<br><br><br><br>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "## Confusion matrix"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "One way to get a better understanding of the errors is by looking at \n",
    "- false positives (type I errors), where the model incorrectly spots examples as fraud\n",
    "- false negatives (type II errors), where it's missing to spot fraud examples "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "slideshow": {
     "slide_type": "-"
    }
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "from sklearn.metrics import ConfusionMatrixDisplay\n",
    "\n",
    "pipe.fit(X_train, y_train)\n",
    "disp = ConfusionMatrixDisplay.from_estimator(\n",
    "    pipe,\n",
    "    X_valid,\n",
    "    y_valid,\n",
    "    display_labels=[\"Non fraud\", \"fraud\"],\n",
    "    values_format=\"d\",\n",
    "    cmap=plt.cm.Blues,\n",
    "    colorbar=False,\n",
    ");"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 2000x600 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "from sklearn.metrics import confusion_matrix\n",
    "\n",
    "predictions = pipe.predict(X_valid)\n",
    "TN, FP, FN, TP = confusion_matrix(y_valid, predictions).ravel()\n",
    "plot_confusion_matrix_example(TN, FP, FN, TP)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "-"
    }
   },
   "source": [
    "- Perfect prediction has all values down the diagonal\n",
    "- Off diagonal entries can often tell us about what is being mis-predicted"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "### What is \"positive\" and \"negative\"?\n",
    "\n",
    "- Two kinds of binary classification problems \n",
    "    - Distinguishing between two classes\n",
    "    - Spotting a class (spot fraud transaction, spot spam, spot disease)\n",
    "- In case of spotting problems, the thing that we are interested in spotting is considered \"positive\". \n",
    "- Above we wanted to spot fraudulent transactions and so they are \"positive\". "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "You can get a numpy array of confusion matrix as follows: "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Confusion matrix for fraud data set\n",
      "[[59700     8]\n",
      " [   38    64]]\n"
     ]
    }
   ],
   "source": [
    "from sklearn.metrics import confusion_matrix\n",
    "\n",
    "predictions = pipe.predict(X_valid)\n",
    "TN, FP, FN, TP = confusion_matrix(y_valid, predictions).ravel()\n",
    "print(\"Confusion matrix for fraud data set\")\n",
    "print(disp.confusion_matrix)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "### Confusion matrix with cross-validation "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "-"
    }
   },
   "source": [
    "- You can also calculate confusion matrix with cross-validation using the `cross_val_predict` method.  \n",
    "- But then you cannot conveniently use `plot_confusion_matrix`. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "slideshow": {
     "slide_type": "-"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[139296,     21],\n",
       "       [    94,    143]])"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.model_selection import cross_val_predict\n",
    "\n",
    "confusion_matrix(y_train, cross_val_predict(pipe, X_train, y_train))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "<br><br><br><br>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "## Precision, recall, f1 score "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "-"
    }
   },
   "source": [
    "- We have been using `.score` to assess our models, which returns accuracy by default. \n",
    "- Accuracy is misleading when we have class imbalance.\n",
    "- We need other metrics to assess our models."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "- We'll discuss three commonly used metrics which are based on confusion matrix: \n",
    "    - recall\n",
    "    - precision\n",
    "    - f1 score \n",
    "- Note that these metrics will only help us assessing our model.  \n",
    "- Later we'll talk about a few ways to address class imbalance problem. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[[59700     8]\n",
      " [   38    64]]\n"
     ]
    }
   ],
   "source": [
    "from sklearn.metrics import confusion_matrix\n",
    "\n",
    "pipe_lr = make_pipeline(StandardScaler(), LogisticRegression())\n",
    "pipe_lr.fit(X_train, y_train)\n",
    "predictions = pipe_lr.predict(X_valid)\n",
    "TN, FP, FN, TP = confusion_matrix(y_valid, predictions).ravel()\n",
    "print(disp.confusion_matrix)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "### Recall \n",
    "\n",
    "Among all positive examples, how many did you identify?\n",
    "$$ recall = \\frac{TP}{TP+FN} = \\frac{TP}{\\#positives} $$"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 640x480 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "ConfusionMatrixDisplay.from_estimator(\n",
    "    pipe,\n",
    "    X_valid,\n",
    "    y_valid,\n",
    "    display_labels=[\"Non fraud\", \"fraud\"],\n",
    "    values_format=\"d\",\n",
    "    cmap=plt.cm.Blues,\n",
    ");"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "TP = 64.0000, FN = 38.0000\n",
      "Recall: 0.6275\n"
     ]
    }
   ],
   "source": [
    "print(\"TP = %0.4f, FN = %0.4f\" % (TP, FN))\n",
    "recall = TP / (TP + FN)\n",
    "print(\"Recall: %0.4f\" % (recall))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "### Precision \n",
    "\n",
    "Among the positive examples you identified, how many were actually positive?\n",
    "\n",
    "$$ precision = \\frac{TP}{TP+FP}$$"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 640x480 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "ConfusionMatrixDisplay.from_estimator(\n",
    "    pipe,\n",
    "    X_valid,\n",
    "    y_valid,\n",
    "    display_labels=[\"Non fraud\", \"fraud\"],\n",
    "    values_format=\"d\",\n",
    "    cmap=plt.cm.Blues,\n",
    ");"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "TP = 64.0000, FP = 8.0000\n",
      "Precision: 0.8889\n"
     ]
    }
   ],
   "source": [
    "print(\"TP = %0.4f, FP = %0.4f\" % (TP, FP))\n",
    "precision = TP / (TP + FP)\n",
    "print(\"Precision: %0.4f\" % (precision))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "### F1-score\n",
    "\n",
    "- F1-score combines precision and recall to give one score, which could be used in hyperparameter optimization, for instance. \n",
    "- F1-score is a harmonic mean of precision and recall. \n",
    "\n",
    "\n",
    "$$ f1 = 2 \\times \\frac{ precision \\times recall}{precision + recall}$$\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "precision: 0.8889\n",
      "recall: 0.6275\n",
      "f1: 0.7356\n"
     ]
    }
   ],
   "source": [
    "print(\"precision: %0.4f\" % (precision))\n",
    "print(\"recall: %0.4f\" % (recall))\n",
    "f1_score = (2 * precision * recall) / (precision + recall)\n",
    "print(\"f1: %0.4f\" % (f1_score))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "Let's look at all metrics at once on our dataset."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {
    "slideshow": {
     "slide_type": "-"
    }
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>calculation</th>\n",
       "      <th>accuracy</th>\n",
       "      <th>error</th>\n",
       "      <th>precision</th>\n",
       "      <th>recall</th>\n",
       "      <th>f1 score</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>manual</td>\n",
       "      <td>0.999231</td>\n",
       "      <td>0.000769</td>\n",
       "      <td>0.888889</td>\n",
       "      <td>0.627451</td>\n",
       "      <td>0.735632</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  calculation  accuracy     error  precision    recall  f1 score\n",
       "0      manual  0.999231  0.000769   0.888889  0.627451  0.735632"
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "## Calculate evaluation metrics by ourselves\n",
    "data = {\n",
    "    \"calculation\": [],\n",
    "    \"accuracy\": [],\n",
    "    \"error\": [],\n",
    "    \"precision\": [],\n",
    "    \"recall\": [],\n",
    "    \"f1 score\": [],\n",
    "}\n",
    "data[\"calculation\"].append(\"manual\")\n",
    "data[\"accuracy\"].append((TP + TN) / (TN + FP + FN + TP))\n",
    "data[\"error\"].append((FP + FN) / (TN + FP + FN + TP))\n",
    "data[\"precision\"].append(precision)  # TP / (TP + FP)\n",
    "data[\"recall\"].append(recall)  # TP / (TP + FN)\n",
    "data[\"f1 score\"].append(f1_score)  # (2 * precision * recall) / (precision + recall)\n",
    "df = pd.DataFrame(data)\n",
    "df"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "- `scikit-learn` has functions for [these metrics](https://scikit-learn.org/stable/modules/classes.html#module-sklearn.metrics)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>accuracy</th>\n",
       "      <th>error</th>\n",
       "      <th>precision</th>\n",
       "      <th>recall</th>\n",
       "      <th>f1 score</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>calculation</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>manual</th>\n",
       "      <td>0.999231</td>\n",
       "      <td>0.000769</td>\n",
       "      <td>0.888889</td>\n",
       "      <td>0.627451</td>\n",
       "      <td>0.735632</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>sklearn</th>\n",
       "      <td>0.999231</td>\n",
       "      <td>0.000769</td>\n",
       "      <td>0.888889</td>\n",
       "      <td>0.627451</td>\n",
       "      <td>0.735632</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "             accuracy     error  precision    recall  f1 score\n",
       "calculation                                                   \n",
       "manual       0.999231  0.000769   0.888889  0.627451  0.735632\n",
       "sklearn      0.999231  0.000769   0.888889  0.627451  0.735632"
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.metrics import accuracy_score, f1_score, precision_score, recall_score\n",
    "\n",
    "data[\"accuracy\"].append(accuracy_score(y_valid, pipe_lr.predict(X_valid)))\n",
    "data[\"error\"].append(1 - accuracy_score(y_valid, pipe_lr.predict(X_valid)))\n",
    "data[\"precision\"].append(\n",
    "    precision_score(y_valid, pipe_lr.predict(X_valid), zero_division=1)\n",
    ")\n",
    "data[\"recall\"].append(recall_score(y_valid, pipe_lr.predict(X_valid)))\n",
    "data[\"f1 score\"].append(f1_score(y_valid, pipe_lr.predict(X_valid)))\n",
    "data[\"calculation\"].append(\"sklearn\")\n",
    "df = pd.DataFrame(data)\n",
    "df.set_index([\"calculation\"])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The scores match. "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "### Classification report "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "- There is a convenient function called `classification_report` in `sklearn` which gives this info. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([0, 1])"
      ]
     },
     "execution_count": 24,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pipe_lr.classes_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "   non-fraud       1.00      1.00      1.00     59708\n",
      "       fraud       0.89      0.63      0.74       102\n",
      "\n",
      "    accuracy                           1.00     59810\n",
      "   macro avg       0.94      0.81      0.87     59810\n",
      "weighted avg       1.00      1.00      1.00     59810\n",
      "\n"
     ]
    }
   ],
   "source": [
    "from sklearn.metrics import classification_report\n",
    "\n",
    "print(\n",
    "    classification_report(\n",
    "        y_valid, pipe_lr.predict(X_valid), target_names=[\"non-fraud\", \"fraud\"]\n",
    "    )\n",
    ")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "### Macro average\n",
    "\n",
    "- You give equal importance to all classes and average over all classes.  \n",
    "- For instance, in the example above, recall for non-fraud is 1.0 and fraud is 0.63, and so macro average is 0.81. \n",
    "- More relevant in case of multi-class problems.    "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "### Weighted average\n",
    "\n",
    "- Weighted by the number of samples in each class. \n",
    "- Divide by the total number of samples. \n",
    "\n",
    "Which one is relevant when depends upon whether you think each class should have the same weight or each sample should have the same weight. "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Toy example"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "     class 0       0.75      1.00      0.86         3\n",
      "     class 1       1.00      0.50      0.67         2\n",
      "\n",
      "    accuracy                           0.80         5\n",
      "   macro avg       0.88      0.75      0.76         5\n",
      "weighted avg       0.85      0.80      0.78         5\n",
      "\n"
     ]
    }
   ],
   "source": [
    "from sklearn.metrics import classification_report\n",
    "y_true = [0, 1, 0, 1, 0]\n",
    "y_pred = [0, 0, 0, 1, 0]\n",
    "target_names = ['class 0', 'class 1']\n",
    "print(classification_report(y_true, y_pred, target_names=target_names))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "- weighted average is weighted by the proportion of examples in a particular class. So for the toy example above:\n",
    "- weighted_average precision: 3/5 * 0.75 + 2/5 * 1.00 = 0.85\n",
    "- weighted_average recall: 3/5 * 1.00 + 2/5 * 0.5 = 0.80\n",
    "- weighted_average f1-score: 3/5 * 0.86 + 2/5 * 0.67 = 0.78"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "- macro average gives equal weight to both classes. So for the toy example above:\n",
    "- macro average precision: 0.5 * 0.75 + 0.5 * 1.00 =0. 875\n",
    "- macro average recall: 0.5 * 1.00 + 0.5 * 0.5 =0. 75\n",
    "- macro average f1-score: 0.5 * 0.75 + 0.5 * 1.00 =0.765"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "### Interim summary \n",
    "\n",
    "- Accuracy is misleading when you have class imbalance. \n",
    "- A confusion matrix provides a way to break down errors made by our model. \n",
    "- We looked at three metrics based on confusion matrix: \n",
    "    - precision, recall, f1-score. "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "- Note that what you consider \"positive\" (fraud in our case) is important when calculating precision, recall, and f1-score. \n",
    "- If you flip what is considered positive or negative, we'll end up with different TP, FP, TN, FN, and hence different precision, recall, and f1-scores. "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "### Evalution metrics overview  \n",
    "There is a lot of terminology here. "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "![](img/evaluation-metrics.png)\n",
    "<!-- <img src='./img/evaluation-metrics.png' width=\"1000\" height=\"1000\" /> -->"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "### Cross validation with different metrics\n",
    "\n",
    "- We can pass different evaluation metrics with `scoring` argument of `cross_validate`."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>fit_time</th>\n",
       "      <th>score_time</th>\n",
       "      <th>test_accuracy</th>\n",
       "      <th>train_accuracy</th>\n",
       "      <th>test_f1</th>\n",
       "      <th>train_f1</th>\n",
       "      <th>test_recall</th>\n",
       "      <th>train_recall</th>\n",
       "      <th>test_precision</th>\n",
       "      <th>train_precision</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.563688</td>\n",
       "      <td>0.023997</td>\n",
       "      <td>0.999097</td>\n",
       "      <td>0.999360</td>\n",
       "      <td>0.694915</td>\n",
       "      <td>0.782979</td>\n",
       "      <td>0.602941</td>\n",
       "      <td>0.678967</td>\n",
       "      <td>0.820000</td>\n",
       "      <td>0.924623</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.811860</td>\n",
       "      <td>0.025594</td>\n",
       "      <td>0.999223</td>\n",
       "      <td>0.999317</td>\n",
       "      <td>0.739496</td>\n",
       "      <td>0.766595</td>\n",
       "      <td>0.647059</td>\n",
       "      <td>0.660517</td>\n",
       "      <td>0.862745</td>\n",
       "      <td>0.913265</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.486615</td>\n",
       "      <td>0.019543</td>\n",
       "      <td>0.999273</td>\n",
       "      <td>0.999216</td>\n",
       "      <td>0.743363</td>\n",
       "      <td>0.726477</td>\n",
       "      <td>0.617647</td>\n",
       "      <td>0.612546</td>\n",
       "      <td>0.933333</td>\n",
       "      <td>0.892473</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.411075</td>\n",
       "      <td>0.019749</td>\n",
       "      <td>0.999172</td>\n",
       "      <td>0.999254</td>\n",
       "      <td>0.697248</td>\n",
       "      <td>0.742981</td>\n",
       "      <td>0.558824</td>\n",
       "      <td>0.634686</td>\n",
       "      <td>0.926829</td>\n",
       "      <td>0.895833</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0.417653</td>\n",
       "      <td>0.019334</td>\n",
       "      <td>0.999147</td>\n",
       "      <td>0.999216</td>\n",
       "      <td>0.696429</td>\n",
       "      <td>0.728850</td>\n",
       "      <td>0.582090</td>\n",
       "      <td>0.617647</td>\n",
       "      <td>0.866667</td>\n",
       "      <td>0.888889</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   fit_time  score_time  test_accuracy  train_accuracy   test_f1  train_f1  \\\n",
       "0  0.563688    0.023997       0.999097        0.999360  0.694915  0.782979   \n",
       "1  0.811860    0.025594       0.999223        0.999317  0.739496  0.766595   \n",
       "2  0.486615    0.019543       0.999273        0.999216  0.743363  0.726477   \n",
       "3  0.411075    0.019749       0.999172        0.999254  0.697248  0.742981   \n",
       "4  0.417653    0.019334       0.999147        0.999216  0.696429  0.728850   \n",
       "\n",
       "   test_recall  train_recall  test_precision  train_precision  \n",
       "0     0.602941      0.678967        0.820000         0.924623  \n",
       "1     0.647059      0.660517        0.862745         0.913265  \n",
       "2     0.617647      0.612546        0.933333         0.892473  \n",
       "3     0.558824      0.634686        0.926829         0.895833  \n",
       "4     0.582090      0.617647        0.866667         0.888889  "
      ]
     },
     "execution_count": 27,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "scoring = [\n",
    "    \"accuracy\",\n",
    "    \"f1\",\n",
    "    \"recall\",\n",
    "    \"precision\",\n",
    "]  # scoring can be a string, a list, or a dictionary\n",
    "pipe = make_pipeline(StandardScaler(), LogisticRegression())\n",
    "scores = cross_validate(\n",
    "    pipe, X_train_big, y_train_big, return_train_score=True, scoring=scoring\n",
    ")\n",
    "pd.DataFrame(scores)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "- You can also create [your own scoring function](https://scikit-learn.org/stable/modules/generated/sklearn.metrics.make_scorer.html) and pass it to `cross_validate`. "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<br><br>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "## ❓❓ Questions for you"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "skip"
    }
   },
   "source": [
    "### (iClicker) Exercise 9.1 \n",
    "\n",
    "**iClicker cloud join link: https://join.iclicker.com/3DP5H**\n",
    "\n",
    "**Select all of the following statements which are TRUE.**\n",
    "\n",
    "- (A) In medical diagnosis, false positives are likely to be more damaging than false negatives (assume \"positive\" means the person has a disease, \"negative\" means they don't).\n",
    "- (B) In spam classification, false positives are more damaging than false negatives (assume \"positive\" means the email is spam, \"negative\" means it's not).\n",
    "- (C) If method A gets a higher accuracy than method B, that means its precision is also higher.\n",
    "- (D) If method A gets a higher accuracy than method B, that means its recall is also higher."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<br><br><br><br>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "Method A - higher accuracy but lower precision\n",
    "\n",
    "| Negative | Positive\n",
    "| -------- |:-------------:|\n",
    "| 90      | 5|\n",
    "| 5      | 0|\n",
    "\n",
    "Method B - lower accuracy but higher precision\n",
    "\n",
    "| Negative | Positive\n",
    "| -------- |:-------------:|\n",
    "| 80      | 15|\n",
    "| 0      | 5|\n",
    "\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "<br><br><br><br>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "## Precision-recall curve and ROC curve"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "- Confusion matrix provides a detailed break down of the errors made by the model. \n",
    "- But when creating a confusion matrix, we are using \"hard\" predictions. \n",
    "- Most classifiers in `scikit-learn` provide `predict_proba` method (or `decision_function`) which provides degree of certainty about predictions by the classifier. \n",
    "- Can we explore the degree of uncertainty to understand and improve the model performance? "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "Let's revisit the classification report on our fraud detection example. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "outputs": [],
   "source": [
    "pipe_lr = make_pipeline(StandardScaler(), LogisticRegression())\n",
    "pipe_lr.fit(X_train, y_train);"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {
    "slideshow": {
     "slide_type": "-"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "   non-fraud       1.00      1.00      1.00     59708\n",
      "       fraud       0.89      0.63      0.74       102\n",
      "\n",
      "    accuracy                           1.00     59810\n",
      "   macro avg       0.94      0.81      0.87     59810\n",
      "weighted avg       1.00      1.00      1.00     59810\n",
      "\n"
     ]
    }
   ],
   "source": [
    "y_pred = pipe_lr.predict(X_valid)\n",
    "print(classification_report(y_valid, y_pred, target_names=[\"non-fraud\", \"fraud\"]))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "By default, predictions use the threshold of 0.5. If `predict_proba` > 0.5, predict \"fraud\" else predict \"non-fraud\"."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {
    "slideshow": {
     "slide_type": "-"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "   non-fraud       1.00      1.00      1.00     59708\n",
      "       fraud       0.89      0.63      0.74       102\n",
      "\n",
      "    accuracy                           1.00     59810\n",
      "   macro avg       0.94      0.81      0.87     59810\n",
      "weighted avg       1.00      1.00      1.00     59810\n",
      "\n"
     ]
    }
   ],
   "source": [
    "y_pred = pipe_lr.predict_proba(X_valid)[:, 1] > 0.50\n",
    "print(classification_report(y_valid, y_pred, target_names=[\"non-fraud\", \"fraud\"]))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "- Suppose for your business it is more costly to miss fraudulent transactions and you want to achieve a recall of at least 75% for the \"fraud\" class. \n",
    "- One way to do this is by changing the threshold of `predict_proba`.\n",
    "    - `predict` returns 1 when `predict_proba`'s probabilities are above 0.5 for the \"fraud\" class.\n",
    "\n",
    "**Key idea: what if we threshold the probability at a smaller value so that we identify more examples as \"fraud\" examples?** "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "Let's lower the threshold to 0.1. In other words, predict the examples as \"fraud\" if `predict_proba` > 0.1.  "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [],
   "source": [
    "y_pred_lower_threshold = pipe_lr.predict_proba(X_valid)[:, 1] > 0.1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       1.00      1.00      1.00     59708\n",
      "           1       0.77      0.76      0.77       102\n",
      "\n",
      "    accuracy                           1.00     59810\n",
      "   macro avg       0.89      0.88      0.88     59810\n",
      "weighted avg       1.00      1.00      1.00     59810\n",
      "\n"
     ]
    }
   ],
   "source": [
    "print(classification_report(y_valid, y_pred_lower_threshold))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "### Operating point \n",
    "\n",
    "- Now our recall for \"fraud\" class is >= 0.75. \n",
    "- Setting a requirement on a classifier (e.g., recall of >= 0.75) is called setting the **operating point**. \n",
    "- It's usually driven by business goals and is useful to make performance guarantees to customers. "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "### Precision/Recall tradeoff \n",
    "\n",
    "- But there is a trade-off between precision and recall. \n",
    "- If you identify more things as \"fraud\", recall is going to increase but there are likely to be more false positives. "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "Let's sweep through different thresholds. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {
    "slideshow": {
     "slide_type": "-"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([0. , 0.1, 0.2, 0.3, 0.4, 0.5, 0.6, 0.7, 0.8, 0.9])"
      ]
     },
     "execution_count": 33,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "thresholds = np.arange(0.0, 1.0, 0.1)\n",
    "thresholds"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {
    "slideshow": {
     "slide_type": "skip"
    }
   },
   "outputs": [],
   "source": [
    "def f(threshold):\n",
    "    preds = pipe_lr.predict_proba(X_valid)[:, 1] > threshold    \n",
    "    print(\"Threshold: \", np.round(threshold,4))\n",
    "    print(\"Precision: \", np.round(precision_score(y_valid, preds),4))\n",
    "    print(\"Recall: \", np.round(recall_score(y_valid, preds), 4))    \n",
    "    print(\"f1 score: \", np.round(f1_score(y_valid, preds), 4))        "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {
    "slideshow": {
     "slide_type": "skip"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "50e00dc58c5a4547a433bab4e04f780b",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "interactive(children=(FloatSlider(value=0.5, description='threshold', max=0.9), Output()), _dom_classes=('widg…"
      ]
     },
     "execution_count": 35,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import ipywidgets as widgets\n",
    "from ipywidgets import interact, interactive\n",
    "\n",
    "interactive(\n",
    "    f,\n",
    "    threshold=widgets.FloatSlider(min=0, max=0.9, step=0.1, value=0.5),\n",
    ")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Let's view P and R for different thresholds as a dataframe. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {
    "slideshow": {
     "slide_type": "-"
    }
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>threshold</th>\n",
       "      <th>precision</th>\n",
       "      <th>recall</th>\n",
       "      <th>f1 score</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.001705</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.003405</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.1</td>\n",
       "      <td>0.772277</td>\n",
       "      <td>0.764706</td>\n",
       "      <td>0.768473</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.2</td>\n",
       "      <td>0.793478</td>\n",
       "      <td>0.715686</td>\n",
       "      <td>0.752577</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.3</td>\n",
       "      <td>0.814815</td>\n",
       "      <td>0.647059</td>\n",
       "      <td>0.721311</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0.4</td>\n",
       "      <td>0.876712</td>\n",
       "      <td>0.627451</td>\n",
       "      <td>0.731429</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>0.5</td>\n",
       "      <td>0.888889</td>\n",
       "      <td>0.627451</td>\n",
       "      <td>0.735632</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>0.6</td>\n",
       "      <td>0.897059</td>\n",
       "      <td>0.598039</td>\n",
       "      <td>0.717647</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>0.7</td>\n",
       "      <td>0.893939</td>\n",
       "      <td>0.578431</td>\n",
       "      <td>0.702381</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>0.8</td>\n",
       "      <td>0.903226</td>\n",
       "      <td>0.549020</td>\n",
       "      <td>0.682927</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>0.9</td>\n",
       "      <td>0.892857</td>\n",
       "      <td>0.490196</td>\n",
       "      <td>0.632911</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   threshold  precision    recall  f1 score\n",
       "0        0.0   0.001705  1.000000  0.003405\n",
       "1        0.1   0.772277  0.764706  0.768473\n",
       "2        0.2   0.793478  0.715686  0.752577\n",
       "3        0.3   0.814815  0.647059  0.721311\n",
       "4        0.4   0.876712  0.627451  0.731429\n",
       "5        0.5   0.888889  0.627451  0.735632\n",
       "6        0.6   0.897059  0.598039  0.717647\n",
       "7        0.7   0.893939  0.578431  0.702381\n",
       "8        0.8   0.903226  0.549020  0.682927\n",
       "9        0.9   0.892857  0.490196  0.632911"
      ]
     },
     "execution_count": 36,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pr_dict = {\"threshold\": [], \"precision\": [], \"recall\": [], \"f1 score\": []}\n",
    "for threshold in thresholds:\n",
    "    preds = pipe_lr.predict_proba(X_valid)[:, 1] > threshold\n",
    "    pr_dict[\"threshold\"].append(threshold)\n",
    "    pr_dict[\"precision\"].append(precision_score(y_valid, preds))\n",
    "    pr_dict[\"recall\"].append(recall_score(y_valid, preds))\n",
    "    pr_dict[\"f1 score\"].append(f1_score(y_valid, preds))\n",
    "pd.DataFrame(pr_dict)    "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "### Decreasing the threshold\n",
    "\n",
    "- Decreasing the threshold means a lower bar for predicting fraud. \n",
    "    - You are willing to risk more false positives in exchange of more true positives. \n",
    "    - recall would either stay the same or go up and precision is likely to go down\n",
    "    - occasionally, precision may increase if all the new examples after decreasing the threshold are TPs. "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "### Increasing the threshold\n",
    "\n",
    "- Increasing the threshold means a higher bar for predicting fraud. \n",
    "    - recall would go down or stay the same but precision is likely to go up \n",
    "    - occasionally, precision may go down as the denominator for precision is TP+FP.    "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "### Precision-recall curve\n",
    "\n",
    "Often, when developing a model, it's not always clear what the operating point will be and to understand the the model better, it's informative to look at all possible thresholds and corresponding trade-offs of precision and recall in a plot.  \n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "from sklearn.metrics import precision_recall_curve\n",
    "\n",
    "precision, recall, thresholds = precision_recall_curve(\n",
    "    y_valid, pipe_lr.predict_proba(X_valid)[:, 1]\n",
    ")\n",
    "plt.plot(precision, recall, label=\"logistic regression: PR curve\")\n",
    "plt.xlabel(\"Precision\")\n",
    "plt.ylabel(\"Recall\")\n",
    "plt.plot(\n",
    "    precision_score(y_valid, pipe_lr.predict(X_valid)),\n",
    "    recall_score(y_valid, pipe_lr.predict(X_valid)),\n",
    "    \"or\",\n",
    "    markersize=10,\n",
    "    label=\"threshold 0.5\",\n",
    ")\n",
    "plt.legend(loc=\"best\");"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "- Each point in the curve corresponds to a possible threshold of the `predict_proba` output. \n",
    "- We can achieve a recall of 0.8 at a precision of 0.4. \n",
    "- The red dot marks the point corresponding to the threshold 0.5.\n",
    "- The top-right would be a perfect classifier (precision = recall = 1)."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "- The threshold is not shown here, but it's going from 0 (upper-left) to 1 (lower right).\n",
    "- At a threshold of 0 (upper left), we are classifying everything  as \"fraud\".\n",
    "- Raising the threshold increases the precision but at the expense of lowering the recall. \n",
    "- At the extreme right, where the threshold is 1, we get into the situation where all the examples classified as \"fraud\" are actually \"fraud\"; we have no false positives. \n",
    "- Here we have a high precision but lower recall. \n",
    "- Usually the goal is to keep recall high as precision goes up. "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "### A few comments on PR curve\n",
    "\n",
    "- Different classifiers might work well in different parts of the curve, i.e., at different operating points.   \n",
    "- We can compare PR curves of different classifiers to understand these differences. "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "### AP score \n",
    "\n",
    "- Often it's useful to have one number summarizing the PR plot (e.g., in hyperparameter optimization)\n",
    "- One way to do this is by computing the area under the PR curve. \n",
    "- This is called **average precision** (AP score)\n",
    "- AP score has a value between 0 (worst) and 1 (best). "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Average precision of logistic regression: 0.757\n"
     ]
    }
   ],
   "source": [
    "from sklearn.metrics import average_precision_score\n",
    "\n",
    "ap_lr = average_precision_score(y_valid, pipe_lr.predict_proba(X_valid)[:, 1])\n",
    "print(\"Average precision of logistic regression: {:.3f}\".format(ap_lr))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "You can also use the following handy function of `sklearn` to get the PR curve and the corresponding AP score. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "from sklearn.metrics import PrecisionRecallDisplay\n",
    "\n",
    "PrecisionRecallDisplay.from_estimator(pipe_lr, X_valid, y_valid);"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "### AP vs. F1-score\n",
    "\n",
    "It is very important to note this distinction:\n",
    "\n",
    "- F1 score is for a given threshold and measures the quality of `predict`.\n",
    "- AP score is a summary across thresholds and measures the quality of `predict_proba`.\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "```{important}\n",
    "Remember to pick the desired threshold based on the results on the validation set and **not** on the test set.\n",
    "```"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "### A few comments on PR curve\n",
    "\n",
    "- Different classifiers might work well in different parts of the curve, i.e., at different operating points.   \n",
    "- We can compare PR curves of different classifiers to understand these differences. \n",
    "- Let's create PR curves for SVC and Logistic Regression. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<style>#sk-container-id-1 {color: black;background-color: white;}#sk-container-id-1 pre{padding: 0;}#sk-container-id-1 div.sk-toggleable {background-color: white;}#sk-container-id-1 label.sk-toggleable__label {cursor: pointer;display: block;width: 100%;margin-bottom: 0;padding: 0.3em;box-sizing: border-box;text-align: center;}#sk-container-id-1 label.sk-toggleable__label-arrow:before {content: \"▸\";float: left;margin-right: 0.25em;color: #696969;}#sk-container-id-1 label.sk-toggleable__label-arrow:hover:before {color: black;}#sk-container-id-1 div.sk-estimator:hover label.sk-toggleable__label-arrow:before {color: black;}#sk-container-id-1 div.sk-toggleable__content {max-height: 0;max-width: 0;overflow: hidden;text-align: left;background-color: #f0f8ff;}#sk-container-id-1 div.sk-toggleable__content pre {margin: 0.2em;color: black;border-radius: 0.25em;background-color: #f0f8ff;}#sk-container-id-1 input.sk-toggleable__control:checked~div.sk-toggleable__content {max-height: 200px;max-width: 100%;overflow: auto;}#sk-container-id-1 input.sk-toggleable__control:checked~label.sk-toggleable__label-arrow:before {content: \"▾\";}#sk-container-id-1 div.sk-estimator input.sk-toggleable__control:checked~label.sk-toggleable__label {background-color: #d4ebff;}#sk-container-id-1 div.sk-label input.sk-toggleable__control:checked~label.sk-toggleable__label {background-color: #d4ebff;}#sk-container-id-1 input.sk-hidden--visually {border: 0;clip: rect(1px 1px 1px 1px);clip: rect(1px, 1px, 1px, 1px);height: 1px;margin: -1px;overflow: hidden;padding: 0;position: absolute;width: 1px;}#sk-container-id-1 div.sk-estimator {font-family: monospace;background-color: #f0f8ff;border: 1px dotted black;border-radius: 0.25em;box-sizing: border-box;margin-bottom: 0.5em;}#sk-container-id-1 div.sk-estimator:hover {background-color: #d4ebff;}#sk-container-id-1 div.sk-parallel-item::after {content: \"\";width: 100%;border-bottom: 1px solid gray;flex-grow: 1;}#sk-container-id-1 div.sk-label:hover label.sk-toggleable__label {background-color: #d4ebff;}#sk-container-id-1 div.sk-serial::before {content: \"\";position: absolute;border-left: 1px solid gray;box-sizing: border-box;top: 0;bottom: 0;left: 50%;z-index: 0;}#sk-container-id-1 div.sk-serial {display: flex;flex-direction: column;align-items: center;background-color: white;padding-right: 0.2em;padding-left: 0.2em;position: relative;}#sk-container-id-1 div.sk-item {position: relative;z-index: 1;}#sk-container-id-1 div.sk-parallel {display: flex;align-items: stretch;justify-content: center;background-color: white;position: relative;}#sk-container-id-1 div.sk-item::before, #sk-container-id-1 div.sk-parallel-item::before {content: \"\";position: absolute;border-left: 1px solid gray;box-sizing: border-box;top: 0;bottom: 0;left: 50%;z-index: -1;}#sk-container-id-1 div.sk-parallel-item {display: flex;flex-direction: column;z-index: 1;position: relative;background-color: white;}#sk-container-id-1 div.sk-parallel-item:first-child::after {align-self: flex-end;width: 50%;}#sk-container-id-1 div.sk-parallel-item:last-child::after {align-self: flex-start;width: 50%;}#sk-container-id-1 div.sk-parallel-item:only-child::after {width: 0;}#sk-container-id-1 div.sk-dashed-wrapped {border: 1px dashed gray;margin: 0 0.4em 0.5em 0.4em;box-sizing: border-box;padding-bottom: 0.4em;background-color: white;}#sk-container-id-1 div.sk-label label {font-family: monospace;font-weight: bold;display: inline-block;line-height: 1.2em;}#sk-container-id-1 div.sk-label-container {text-align: center;}#sk-container-id-1 div.sk-container {/* jupyter's `normalize.less` sets `[hidden] { display: none; }` but bootstrap.min.css set `[hidden] { display: none !important; }` so we also need the `!important` here to be able to override the default hidden behavior on the sphinx rendered scikit-learn.org. See: https://github.com/scikit-learn/scikit-learn/issues/21755 */display: inline-block !important;position: relative;}#sk-container-id-1 div.sk-text-repr-fallback {display: none;}</style><div id=\"sk-container-id-1\" class=\"sk-top-container\"><div class=\"sk-text-repr-fallback\"><pre>Pipeline(steps=[(&#x27;standardscaler&#x27;, StandardScaler()), (&#x27;svc&#x27;, SVC())])</pre><b>In a Jupyter environment, please rerun this cell to show the HTML representation or trust the notebook. <br />On GitHub, the HTML representation is unable to render, please try loading this page with nbviewer.org.</b></div><div class=\"sk-container\" hidden><div class=\"sk-item sk-dashed-wrapped\"><div class=\"sk-label-container\"><div class=\"sk-label sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-1\" type=\"checkbox\" ><label for=\"sk-estimator-id-1\" class=\"sk-toggleable__label sk-toggleable__label-arrow\">Pipeline</label><div class=\"sk-toggleable__content\"><pre>Pipeline(steps=[(&#x27;standardscaler&#x27;, StandardScaler()), (&#x27;svc&#x27;, SVC())])</pre></div></div></div><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-estimator sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-2\" type=\"checkbox\" ><label for=\"sk-estimator-id-2\" class=\"sk-toggleable__label sk-toggleable__label-arrow\">StandardScaler</label><div class=\"sk-toggleable__content\"><pre>StandardScaler()</pre></div></div></div><div class=\"sk-item\"><div class=\"sk-estimator sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-3\" type=\"checkbox\" ><label for=\"sk-estimator-id-3\" class=\"sk-toggleable__label sk-toggleable__label-arrow\">SVC</label><div class=\"sk-toggleable__content\"><pre>SVC()</pre></div></div></div></div></div></div></div>"
      ],
      "text/plain": [
       "Pipeline(steps=[('standardscaler', StandardScaler()), ('svc', SVC())])"
      ]
     },
     "execution_count": 40,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pipe_svc = make_pipeline(StandardScaler(), SVC())\n",
    "\n",
    "pipe_svc.fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<style>#sk-container-id-2 {color: black;background-color: white;}#sk-container-id-2 pre{padding: 0;}#sk-container-id-2 div.sk-toggleable {background-color: white;}#sk-container-id-2 label.sk-toggleable__label {cursor: pointer;display: block;width: 100%;margin-bottom: 0;padding: 0.3em;box-sizing: border-box;text-align: center;}#sk-container-id-2 label.sk-toggleable__label-arrow:before {content: \"▸\";float: left;margin-right: 0.25em;color: #696969;}#sk-container-id-2 label.sk-toggleable__label-arrow:hover:before {color: black;}#sk-container-id-2 div.sk-estimator:hover label.sk-toggleable__label-arrow:before {color: black;}#sk-container-id-2 div.sk-toggleable__content {max-height: 0;max-width: 0;overflow: hidden;text-align: left;background-color: #f0f8ff;}#sk-container-id-2 div.sk-toggleable__content pre {margin: 0.2em;color: black;border-radius: 0.25em;background-color: #f0f8ff;}#sk-container-id-2 input.sk-toggleable__control:checked~div.sk-toggleable__content {max-height: 200px;max-width: 100%;overflow: auto;}#sk-container-id-2 input.sk-toggleable__control:checked~label.sk-toggleable__label-arrow:before {content: \"▾\";}#sk-container-id-2 div.sk-estimator input.sk-toggleable__control:checked~label.sk-toggleable__label {background-color: #d4ebff;}#sk-container-id-2 div.sk-label input.sk-toggleable__control:checked~label.sk-toggleable__label {background-color: #d4ebff;}#sk-container-id-2 input.sk-hidden--visually {border: 0;clip: rect(1px 1px 1px 1px);clip: rect(1px, 1px, 1px, 1px);height: 1px;margin: -1px;overflow: hidden;padding: 0;position: absolute;width: 1px;}#sk-container-id-2 div.sk-estimator {font-family: monospace;background-color: #f0f8ff;border: 1px dotted black;border-radius: 0.25em;box-sizing: border-box;margin-bottom: 0.5em;}#sk-container-id-2 div.sk-estimator:hover {background-color: #d4ebff;}#sk-container-id-2 div.sk-parallel-item::after {content: \"\";width: 100%;border-bottom: 1px solid gray;flex-grow: 1;}#sk-container-id-2 div.sk-label:hover label.sk-toggleable__label {background-color: #d4ebff;}#sk-container-id-2 div.sk-serial::before {content: \"\";position: absolute;border-left: 1px solid gray;box-sizing: border-box;top: 0;bottom: 0;left: 50%;z-index: 0;}#sk-container-id-2 div.sk-serial {display: flex;flex-direction: column;align-items: center;background-color: white;padding-right: 0.2em;padding-left: 0.2em;position: relative;}#sk-container-id-2 div.sk-item {position: relative;z-index: 1;}#sk-container-id-2 div.sk-parallel {display: flex;align-items: stretch;justify-content: center;background-color: white;position: relative;}#sk-container-id-2 div.sk-item::before, #sk-container-id-2 div.sk-parallel-item::before {content: \"\";position: absolute;border-left: 1px solid gray;box-sizing: border-box;top: 0;bottom: 0;left: 50%;z-index: -1;}#sk-container-id-2 div.sk-parallel-item {display: flex;flex-direction: column;z-index: 1;position: relative;background-color: white;}#sk-container-id-2 div.sk-parallel-item:first-child::after {align-self: flex-end;width: 50%;}#sk-container-id-2 div.sk-parallel-item:last-child::after {align-self: flex-start;width: 50%;}#sk-container-id-2 div.sk-parallel-item:only-child::after {width: 0;}#sk-container-id-2 div.sk-dashed-wrapped {border: 1px dashed gray;margin: 0 0.4em 0.5em 0.4em;box-sizing: border-box;padding-bottom: 0.4em;background-color: white;}#sk-container-id-2 div.sk-label label {font-family: monospace;font-weight: bold;display: inline-block;line-height: 1.2em;}#sk-container-id-2 div.sk-label-container {text-align: center;}#sk-container-id-2 div.sk-container {/* jupyter's `normalize.less` sets `[hidden] { display: none; }` but bootstrap.min.css set `[hidden] { display: none !important; }` so we also need the `!important` here to be able to override the default hidden behavior on the sphinx rendered scikit-learn.org. See: https://github.com/scikit-learn/scikit-learn/issues/21755 */display: inline-block !important;position: relative;}#sk-container-id-2 div.sk-text-repr-fallback {display: none;}</style><div id=\"sk-container-id-2\" class=\"sk-top-container\"><div class=\"sk-text-repr-fallback\"><pre>Pipeline(steps=[(&#x27;standardscaler&#x27;, StandardScaler()),\n",
       "                (&#x27;logisticregression&#x27;, LogisticRegression(max_iter=1000))])</pre><b>In a Jupyter environment, please rerun this cell to show the HTML representation or trust the notebook. <br />On GitHub, the HTML representation is unable to render, please try loading this page with nbviewer.org.</b></div><div class=\"sk-container\" hidden><div class=\"sk-item sk-dashed-wrapped\"><div class=\"sk-label-container\"><div class=\"sk-label sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-4\" type=\"checkbox\" ><label for=\"sk-estimator-id-4\" class=\"sk-toggleable__label sk-toggleable__label-arrow\">Pipeline</label><div class=\"sk-toggleable__content\"><pre>Pipeline(steps=[(&#x27;standardscaler&#x27;, StandardScaler()),\n",
       "                (&#x27;logisticregression&#x27;, LogisticRegression(max_iter=1000))])</pre></div></div></div><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-estimator sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-5\" type=\"checkbox\" ><label for=\"sk-estimator-id-5\" class=\"sk-toggleable__label sk-toggleable__label-arrow\">StandardScaler</label><div class=\"sk-toggleable__content\"><pre>StandardScaler()</pre></div></div></div><div class=\"sk-item\"><div class=\"sk-estimator sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-6\" type=\"checkbox\" ><label for=\"sk-estimator-id-6\" class=\"sk-toggleable__label sk-toggleable__label-arrow\">LogisticRegression</label><div class=\"sk-toggleable__content\"><pre>LogisticRegression(max_iter=1000)</pre></div></div></div></div></div></div></div>"
      ],
      "text/plain": [
       "Pipeline(steps=[('standardscaler', StandardScaler()),\n",
       "                ('logisticregression', LogisticRegression(max_iter=1000))])"
      ]
     },
     "execution_count": 41,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pipe_lr = make_pipeline(StandardScaler(), LogisticRegression(max_iter=1000))\n",
    "pipe_lr.fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "How to get precision and recall for different thresholds? \n",
    "- Use the function `precision_recall_curve`"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {},
   "outputs": [],
   "source": [
    "precision_lr, recall_lr, thresholds_lr = precision_recall_curve(\n",
    "    y_valid, pipe_lr.predict_proba(X_valid)[:, 1]\n",
    ")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "How many thresholds? \n",
    "- It uses `n_thresholds` where `n_thresholds` is the number of unique `predict_proba` scores in our dataset. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "59049"
      ]
     },
     "execution_count": 43,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(np.unique(pipe_lr.predict_proba(X_valid)[:, 1]))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "- For each threshold, precision and recall is calculated.  \n",
    "- The last precision and recall values are 1. and 0. respectively and do not have a corresponding threshold. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "((59049,), (59050,), (59050,))"
      ]
     },
     "execution_count": 44,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "thresholds_lr.shape, precision_lr.shape, recall_lr.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "outputs": [],
   "source": [
    "precision_lr, recall_lr, thresholds_lr = precision_recall_curve(\n",
    "    y_valid, pipe_lr.predict_proba(X_valid)[:, 1]\n",
    ")\n",
    "precision_svc, recall_svc, thresholds_svc = precision_recall_curve(\n",
    "    y_valid, pipe_svc.decision_function(X_valid)\n",
    ")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "For logistic regression, what's the index of the threshold that is closest to the default threshold of 0.5? \n",
    "- We are subtracting 0.5 from the thresholds so that \n",
    "    - the numbers close to 0 become -0.5\n",
    "    - the numbers close to 1 become 0.5    \n",
    "    - the numbers close to 0.5 become 0\n",
    "- After this transformation, we are interested in the threshold index where the number is close to 0. So we take  absolute values and argmin.       "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {},
   "outputs": [],
   "source": [
    "close_default_lr = np.argmin(np.abs(thresholds_lr - 0.5))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "SVC doesn't have `predict_proba`. Instead it has something called `decision_function`. The index of the threshold that is closest to 0 of decision function is the default threshold in SVC. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {},
   "outputs": [],
   "source": [
    "close_zero_svm = np.argmin(np.abs(thresholds_svc))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(precision_svc, recall_svc, label=\"svc\")\n",
    "plt.plot(precision_lr, recall_lr, label=\"logistic regression\")\n",
    "plt.plot(\n",
    "    precision_svc[close_zero_svm],\n",
    "    recall_svc[close_zero_svm],\n",
    "    \"o\",\n",
    "    markersize=10,\n",
    "    label=\"default threshold svc\",\n",
    "    c=\"b\",\n",
    ")\n",
    "plt.plot(\n",
    "    precision_lr[close_default_lr],\n",
    "    recall_lr[close_default_lr],\n",
    "    \"*\",\n",
    "    markersize=10,\n",
    "    label=\"default threshold logistic regression\",\n",
    "    c=\"r\",\n",
    ")\n",
    "\n",
    "plt.xlabel(\"Precision\")\n",
    "plt.ylabel(\"Recall\")\n",
    "plt.legend(loc=\"best\");"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "outputs": [],
   "source": [
    "svc_preds = pipe_svc.predict(X_valid)\n",
    "lr_preds = pipe_lr.predict(X_valid)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {
    "slideshow": {
     "slide_type": "-"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "f1_score of logistic regression: 0.736\n",
      "f1_score of svc: 0.726\n"
     ]
    }
   ],
   "source": [
    "print(\"f1_score of logistic regression: {:.3f}\".format(f1_score(y_valid, lr_preds)))\n",
    "print(\"f1_score of svc: {:.3f}\".format(f1_score(y_valid, svc_preds)))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "outputs": [],
   "source": [
    "ap_lr = average_precision_score(y_valid, pipe_lr.predict_proba(X_valid)[:, 1])\n",
    "ap_svc = average_precision_score(y_valid, pipe_svc.decision_function(X_valid))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {
    "slideshow": {
     "slide_type": "-"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Average precision of logistic regression: 0.757\n",
      "Average precision of SVC: 0.790\n"
     ]
    }
   ],
   "source": [
    "print(\"Average precision of logistic regression: {:.3f}\".format(ap_lr))\n",
    "print(\"Average precision of SVC: {:.3f}\".format(ap_svc))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "- Comparing the precision-recall curves provide us a detail insight compared to f1 score.\n",
    "- For example, F1 scores for SVC and logistic regressions are pretty similar. In fact, f1 score of logistic regression is a tiny bit better. \n",
    "- But when we look at the PR curve, we see that SVC is doing better than logistic regression for most of the other thresholds. "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<br><br>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "### Receiver Operating Characteristic (ROC) curve \n",
    "\n",
    "- Another commonly used tool to analyze the behavior of classifiers at different thresholds.  \n",
    "- Similar to PR curve, it considers all possible thresholds for a given classifier given by `predict_proba` but instead of precision and recall it plots false positive rate (FPR) and true positive rate (TPR or recall).\n",
    "$$ TPR = \\frac{TP}{TP + FN}, FPR  = \\frac{FP}{FP + TN}$$\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "from sklearn.metrics import roc_curve\n",
    "\n",
    "fpr, tpr, thresholds = roc_curve(y_valid, pipe_lr.predict_proba(X_valid)[:, 1])\n",
    "plt.plot(fpr, tpr, label=\"ROC Curve\")\n",
    "plt.xlabel(\"FPR\")\n",
    "plt.ylabel(\"TPR (recall)\")\n",
    "\n",
    "default_threshold = np.argmin(np.abs(thresholds - 0.5))\n",
    "\n",
    "plt.plot(\n",
    "    fpr[default_threshold],\n",
    "    tpr[default_threshold],\n",
    "    \"or\",\n",
    "    markersize=10,\n",
    "    label=\"threshold 0.5\",\n",
    ")\n",
    "plt.legend(loc=\"best\");"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "- The ideal curve is close to the top left\n",
    "    - Ideally, you want a classifier with high recall while keeping low false positive rate.  \n",
    "- The red dot corresponds to the threshold of 0.5, which is used by predict.\n",
    "- We see that compared to the default threshold, we can achieve a better recall of around 0.8 without increasing FPR. "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "Let's compare ROC curve of different classifiers. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "metadata": {
    "slideshow": {
     "slide_type": "-"
    }
   },
   "outputs": [],
   "source": [
    "fpr_lr, tpr_lr, thresholds_lr = roc_curve(y_valid, pipe_lr.predict_proba(X_valid)[:, 1])\n",
    "\n",
    "fpr_svc, tpr_svc, thresholds_svc = roc_curve(\n",
    "    y_valid, pipe_svc.decision_function(X_valid)\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "metadata": {},
   "outputs": [],
   "source": [
    "close_default_lr = np.argmin(np.abs(thresholds_lr - 0.5))\n",
    "close_zero_svm = np.argmin(np.abs(thresholds_svc))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(fpr_svc, tpr_svc, label=\"svc\")\n",
    "plt.plot(fpr_lr, tpr_lr, label=\"logistic regression\")\n",
    "plt.plot(\n",
    "    fpr_svc[close_zero_svm],\n",
    "    tpr_svc[close_zero_svm],\n",
    "    \"o\",\n",
    "    markersize=10,\n",
    "    label=\"default threshold svc\",\n",
    "    c=\"b\",\n",
    ")\n",
    "plt.plot(\n",
    "    fpr_lr[close_default_lr],\n",
    "    tpr_lr[close_default_lr],\n",
    "    \"*\",\n",
    "    markersize=10,\n",
    "    label=\"default threshold logistic regression\",\n",
    "    c=\"r\",\n",
    ")\n",
    "\n",
    "plt.xlabel(\"False positive rate\")\n",
    "plt.ylabel(\"True positive rate (Recall)\")\n",
    "plt.legend(loc=\"best\");"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "### Area under the curve (AUC)\n",
    "\n",
    "- AUC provides a single meaningful number for the model performance. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "AUC for LR: 0.970\n",
      "AUC for SVC: 0.938\n"
     ]
    }
   ],
   "source": [
    "from sklearn.metrics import roc_auc_score\n",
    "\n",
    "roc_lr = roc_auc_score(y_valid, pipe_lr.predict_proba(X_valid)[:, 1])\n",
    "roc_svc = roc_auc_score(y_valid, pipe_svc.decision_function(X_valid))\n",
    "print(\"AUC for LR: {:.3f}\".format(roc_lr))\n",
    "print(\"AUC for SVC: {:.3f}\".format(roc_svc))\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "- AUC of 0.5 means random chance. \n",
    "- AUC can be interpreted as evaluating the **ranking** of positive examples.\n",
    "- What's the probability that a randomly picked positive point has a higher score according to the classifier than a randomly picked point from the negative class. \n",
    "- AUC of 1.0 means all positive points have a higher score than all negative points. "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "skip"
    }
   },
   "source": [
    "```{important}\n",
    "For classification problems with imbalanced classes, using AP score or AUC is often much more meaningful than using accuracy. \n",
    "```"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Similar to `PrecisionRecallCurveDisplay`, there is a `RocCurveDisplay` function in sklearn. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "from sklearn.metrics import RocCurveDisplay\n",
    "\n",
    "RocCurveDisplay.from_estimator(pipe_lr, X_valid, y_valid);"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "### Let's look at all the scores at once"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "fit_time                  0.438035\n",
       "score_time                0.035626\n",
       "test_accuracy             0.999182\n",
       "test_f1                   0.714290\n",
       "test_recall               0.601712\n",
       "test_precision            0.881915\n",
       "test_roc_auc              0.967228\n",
       "test_average_precision    0.744074\n",
       "dtype: float64"
      ]
     },
     "execution_count": 59,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "scoring = [\"accuracy\", \"f1\", \"recall\", \"precision\", \"roc_auc\", \"average_precision\"]\n",
    "pipe = make_pipeline(StandardScaler(), LogisticRegression())\n",
    "scores = cross_validate(pipe, X_train_big, y_train_big, scoring=scoring)\n",
    "pd.DataFrame(scores).mean()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "skip"
    }
   },
   "source": [
    "```{seealso}\n",
    "Check out [these visualization](https://github.com/dariyasydykova/open_projects/tree/master/ROC_animation) on ROC and AUC.  \n",
    "```"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "skip"
    }
   },
   "source": [
    "<br><br><br><br>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "## Dealing with class imbalance"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "### Class imbalance in training sets\n",
    "\n",
    "- This typically refers to having many more examples of one class than another in one's training set.\n",
    "- Real world data is often imbalanced. \n",
    "    - Our Credit Card Fraud dataset is imbalanced.\n",
    "    - Ad clicking data is usually drastically imbalanced. (Only around ~0.01% ads are clicked.)\n",
    "    - Spam classification datasets are also usually imbalanced."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "### Addressing class imbalance\n",
    "A very important question to ask yourself: \"Why do I have a class imbalance?\"\n",
    "\n",
    "- Is it because one class is much more rare than the other?\n",
    "    - If it's just because one is more rare than the other, you need to ask whether you care about one type of error more than the other.    \n",
    "- Is it because of my data collection methods?\n",
    "    - If it's the data collection, then that means _your test and training data come from different distributions_!\n",
    "  \n",
    "In some cases, it may be fine to just ignore the class imbalance."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "### Which type of error is more important? \n",
    "\n",
    "- False positives (FPs) and false negatives (FNs) have quite different real-world consequences. \n",
    "- In PR curve and ROC curve, we saw how changing the prediction threshold can change FPs and FNs. \n",
    "- We can then pick the threshold that's appropriate for our problem. \n",
    "- Example: if we want high recall, we may use a lower threshold (e.g., a threshold of 0.1). We'll then catch more fraudulent transactions. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "   non-fraud       1.00      1.00      1.00     59708\n",
      "       fraud       0.89      0.63      0.74       102\n",
      "\n",
      "    accuracy                           1.00     59810\n",
      "   macro avg       0.94      0.81      0.87     59810\n",
      "weighted avg       1.00      1.00      1.00     59810\n",
      "\n"
     ]
    }
   ],
   "source": [
    "pipe_lr = make_pipeline(StandardScaler(), LogisticRegression())\n",
    "pipe_lr.fit(X_train, y_train)\n",
    "y_pred = pipe_lr.predict(X_valid)\n",
    "print(classification_report(y_valid, y_pred, target_names=[\"non-fraud\", \"fraud\"]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "metadata": {
    "slideshow": {
     "slide_type": "-"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "   non-fraud       1.00      1.00      1.00     59708\n",
      "       fraud       0.77      0.76      0.77       102\n",
      "\n",
      "    accuracy                           1.00     59810\n",
      "   macro avg       0.89      0.88      0.88     59810\n",
      "weighted avg       1.00      1.00      1.00     59810\n",
      "\n"
     ]
    }
   ],
   "source": [
    "y_pred = pipe_lr.predict_proba(X_valid)[:, 1] > 0.10\n",
    "print(classification_report(y_valid, y_pred, target_names=[\"non-fraud\", \"fraud\"]))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "### Handling imbalance\n",
    "\n",
    "Can we change the model itself rather than changing the threshold so that it takes into account the errors that are important to us?\n",
    "\n",
    "There are two common approaches for this: \n",
    "- **Changing the data (optional)** (not covered in this course)\n",
    "   - Undersampling\n",
    "   - Oversampling \n",
    "       - Random oversampling\n",
    "       - SMOTE \n",
    "- **Changing the training procedure** \n",
    "    - `class_weight`"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "### Changing the training procedure \n",
    "\n",
    "- All `sklearn` classifiers have a parameter called `class_weight`.\n",
    "- This allows you to specify that one class is more important than another.\n",
    "- For example, maybe a false negative is 10x more problematic than a false positive. "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "### Example: `class_weight` parameter of `sklearn LogisticRegression` \n",
    "> class sklearn.linear_model.LogisticRegression(penalty='l2', dual=False, tol=0.0001, C=1.0, fit_intercept=True, intercept_scaling=1, **class_weight=None**, random_state=None, solver='lbfgs', max_iter=100, multi_class='auto', verbose=0, warm_start=False, n_jobs=None, l1_ratio=None)\n",
    "\n",
    "> class_weight: dict or 'balanced', default=None\n",
    "\n",
    "> Weights associated with classes in the form {class_label: weight}. If not given, all classes are supposed to have weight one. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "metadata": {
    "slideshow": {
     "slide_type": "skip"
    }
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/kvarada/opt/miniconda3/envs/cpsc330/lib/python3.10/site-packages/IPython/core/display.py:419: UserWarning: Consider using IPython.display.IFrame instead\n",
      "  warnings.warn(\"Consider using IPython.display.IFrame instead\")\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<iframe src=https://scikit-learn.org/stable/modules/generated/sklearn.linear_model.LogisticRegression.html width=1000 height=650></iframe>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "execution_count": 62,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "url = \"https://scikit-learn.org/stable/modules/generated/sklearn.linear_model.LogisticRegression.html\"\n",
    "HTML(\"<iframe src=%s width=1000 height=650></iframe>\" % url)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 640x480 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "ConfusionMatrixDisplay.from_estimator(\n",
    "    pipe_lr,\n",
    "    X_valid,\n",
    "    y_valid,\n",
    "    display_labels=[\"Non fraud\", \"fraud\"],\n",
    "    values_format=\"d\",\n",
    "    cmap=plt.cm.Blues,\n",
    ");"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 64,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([0, 1])"
      ]
     },
     "execution_count": 64,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pipe_lr.named_steps[\"logisticregression\"].classes_"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "Let's set \"fraud\" class a weight of 10."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 65,
   "metadata": {
    "slideshow": {
     "slide_type": "-"
    }
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAqkAAAHACAYAAACBNfHIAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjUuMywgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/NK7nSAAAACXBIWXMAAA9hAAAPYQGoP6dpAAB68klEQVR4nO3dd3hUZdrH8e8kIT2ZhBICBBJQQxEUBQsC0glSNIggIgq4VFd2LeiqKxAQRSn6roKgdBVBiigivakgIAKC0gUCoYUYYCaVtHn/yGY2QyY9IcP4+3Cd65qc57nPeWYI5M7TjsFisVgQEREREXEgLhXdABERERGR6ylJFRERERGHoyRVRERERByOklQRERERcThKUkVERETE4ShJFRERERGHoyRVRERERByOW0U3QKQgWVlZnD9/Hj8/PwwGQ0U3R0REislisZCQkEDNmjVxcSm/vrHU1FTS0tJKfR13d3c8PT3LoEVSWkpSxaGdP3+e2rVrV3QzRESklGJiYggJCSmXa6empuLlVwUykkt9reDgYE6dOqVE1QEoSRWH5ufnB4B7owEYXN0ruDUi5ePM1ikV3QSRcpNgNnNr3drW/8/LQ1paGmQk49FoAJTmZ0VmGhcPLSAtLU1JqgNQkioOLWeI3+DqriRVnJa/v39FN0Gk3N2QKVtunqX6WWExaKmOI1GSKiIiIs7BAJQmGdbSB4eiJFVEREScg8El+yhNvDgM/W2IiIiIiMNRT6qIiIg4B4OhlMP9Gu93JEpSRURExDlouN+p6G9DRERERByOelJFRETEOWi436koSRUREREnUcrhfg0wOxT9bYiIiIiIw1FPqoiIiDgHDfc7FSWpIiIi4hy0ut+p6G9DRERERByOelJFRETEOWi436koSRURERHnoOF+p6IkVURERJyDelKdin5lEBERERGHo55UERERcQ4a7ncqSlJFRETEORgMpUxSNdzvSPQrg4iIiIg4HPWkioiIiHNwMWQfpYkXh6EkVURERJyD5qQ6Ff1tiIiIiIjDUU+qiIiIOAftk+pUlKSKiIiIc9Bwv1PR34aIiIiIOBz1pIqIiIhz0HC/U1GSKiIiIs5Bw/1ORUmqiIiIOAf1pDoV/cogIiIiIg5HPakiIiLiHDTc71SUpIqIiIhz0HC/U9GvDCIiIiLicNSTKiIiIk6ilMP96rtzKEpSRURExDlouN+p6FcGEREREXE46kkVERER52AwlHJ1v3pSHYmSVBEREXEO2oLKqehvQ0REREQcjnpSRURExDlo4ZRTUZIqIiIizkHD/U5FSaqIiIg4B/WkOhX9yiAiIiIiDkc9qSIiIuIcNNzvVJSkioiIiHPQcL9T0a8MIiIiIuJw1JMqIiIiTsFgMGBQT6rTUJIqIiIiTkFJqnPRcL+IiIiIOBz1pIqIiIhzMPz3KE28OAz1pIqIiIhTyBnuL81REvPnzy/StTdu3JjvNU6cOMGwYcOoW7cunp6eVKtWjYiICJYvX16kNuzdu5f+/fsTEhKCh4cHNWrUoGfPnmzevLlI8Vu2bKFnz57UqFEDDw8PQkJC6N+/P3v37i1S/FdffUVERARBQUF4enpSt25dhg0bxh9//FGkeHuUpIqIiIiUARcXF6pXr57v4eHhYTdu9erV3HHHHXzyySdER0fj4eHB5cuXWb9+PY899hjPPPMMFosl3/vOnj2b++67j4ULF3Lu3Dm8vLyIjY3l66+/pkOHDkRFRRXY7qioKNq3b8/XX39NbGwsXl5enDt3joULF3Lfffcxe/bsfGMtFgvPPPMMvXr1Yv369cTHx+Ph4UF0dDSffPIJd955J6tXry7S53c9JakiIiLiFCqqJzVH7dq1uXjxYr5H69at88ScOnWKPn36kJycTMuWLTl69CgmkwmTycSYMWMAmDdvHpMnT7Z7zx07djB8+HAyMjKIjIwkJiaGq1evEhcXx7BhwwAYN24cS5YssRu/ZMkSxo0bB8CwYcOIi4vj6tWrxMTEEBkZSUZGBsOHD2fHjh124ydPnsy8efMAGDt2rLXtR44c4YEHHiA5OZk+ffpw6tSp4n2YKEkVERERJ1HRSWpJjBkzhqSkJIKDg1m1ahXh4eEA+Pr6Mm7cOIYOHQrAW2+9xZUrV/LEv/LKK2RmZtKkSROWLFlCSEgIAFWqVGHmzJlEREQA8K9//YvMzEyb2MzMTF555RUAunTpwsyZM6lSpQoAISEhfPnllzRu3NimXm5XrlxhwoQJQHaCGxUVha+vLwD169dn1apVBAcHk5SUZE24i0NJqoiIiDiFmy1JTUpKss45HTFiBAEBAXnqvPbaawCYzWa+/vprm7KTJ0+ybds2AEaNGkWlSpXyjY+OjuaHH36wKfv+++85ffq0Tb3c3N3dGTVqFADbtm3L0xu6YsUKEhIS8o0PDAxk+PDhACxfvpykpKQ8dQqiJFVERESkAmzbto2UlBQAHnroIbt1wsLCaNiwIQDr16+3KduwYYP1dZcuXezGt2rVCj8/vwLj/fz8aNmypd343O3KL75Ro0aEhoYWGJ+SkmJNqItKSaqIiIg4B0MZHKUQFxdHs2bN8PX1xcvLi3r16tG/f3+2bt1qt/7vv/9ufd24ceN8r5tTdvDgQbvxQUFBBAUF2Y11dXWlQYMGBcY3bNgQV1dXu/FBQUFUq1atwPiitN1efGGUpIqIiIhTKKvhfrPZbHNcu3atSPdPTk5m7969uLu7k5WVxalTp1i4cCHt2rXjmWeeISMjw6b++fPngexhcS8vr3yvW6tWLZv618fnlDtivLe3t3Uaw/XxhVGSKiIiIpJL7dq1MRqN1mPixIkF1q9ZsyZjx45l//79pKamcvnyZZKTk9m+fTsdO3YEslfov/DCCzZxOfM5vb29C7x+TnlOfWeJL4yeOCUiIiJOwWCgdIuf/hsaExODv7+/9XR++5vm6Ny5M507d7Y55+rqygMPPMC6det49NFH+eabb/joo4/4xz/+wW233VbyNv6FqCdVREREnIKBUg73/zdL9ff3tzkKS1IL4uLiwpQpUwDIysri22+/tZblLGhKTk4u8Bo55Tn1nSW+MEpSRURERMrRrbfeStWqVYHsbaNy1KxZE8jebzRnlb89586ds6l/fXxOuSPGJycnc/XqVbvxhVGSKiIiIk7hZtsnNffK99wr/a+XU3b77bfbjb906RJxcXF2YzMzMzly5EiB8YcPH86z0X+O3NfOL74obbcXXxglqSIiIuIcKngLqvycOHGCP//8E4C6detaz7dq1cq6qn/t2rV2Y0+fPs3hw4cB8sx77dSpk/V1fvHbt2+3LljKLz4hIYGffvrJbnzu6+YXf/jwYc6cOVNgvJeXF61atbJbJz9KUkVERERKyGKxFFr+8ssvA9nzU7t3724t8/HxoVevXgDMmDEDk8mUJ/7dd98FsudzRkZG2pTVq1fPmvhNnTqV9PT0PPHvvPMOAKGhoTz44IM2ZW3atLFuwp9TL7f09HSmTp0KZCfUuRNsgJ49e+Ln54fFYrEbf/XqVWbOnAlAr1698PHxyVOnIEpSRURExDmUdqi/BMP9p0+f5t577+Xjjz/m5MmT1qQ1KyuLnTt38tBDD7FixQog+/n29evXt4kfP348Pj4+XLhwgR49enD8+HEg+5Gp48ePtyZ5b7zxBoGBgXnu/+677+Lq6sr+/fvp27evdX7o5cuXefbZZ1mzZg0AkyZNyrNhv6urK5MmTQJg9erVPPvss1y+fBnInmfat29fDhw4YFMvt8DAQN544w0AZs6cyfjx462PPj127Bg9evTgwoUL+Pj4MH78+OJ+tBgshf0KIFKBzGYzRqMRjyZDMLi6V3RzRMrFld3TKroJIuXGbDZTvYoRk8lks61TWd/DaDRSud9cXNwL3rOzIFlpyVz+4plitTU6Otqmh9HDwwM/Pz8SEhJsHgIwaNAgPvnkE9zc8u7+uXr1anr37m1dBW80GklMTLTOEx00aBBz5szJd87s7NmzGTFihPVhAQEBAZhMJmvCPHbsWKKiovJ9D1FRUYwbNw7IntdrNBqti53c3NyYMWMGgwcPthtrsVj429/+xrx584DsxNfX19faK+zt7c3SpUvp2rVrvvfPj3pSRURExClUxMKp6tWr8+GHH9KvXz8aNWqEv78/V69epVKlSjRo0IBnnnmGbdu2MXfuXLsJKkDXrl05cOAAQ4YMISwsjNTUVAIDA+nUqRPLli1j7ty5BbZt8ODB7Nq1i379+lGrVi2Sk5MJCgoiMjKSTZs2FZigQnaSumnTJiIjIwkKCiI5OZlatWrRr18/du7cmW+CCtmf+dy5c1m2bBmdOnUiMDCQ1NRUQkNDGTJkCPv37y9RggrqSRUHp55U+StQT6o4sxvZk1rlyXml7kmNXzioXNsqRacnTomIiIhzKO0K/Ru7A5UUQkmqiIiIOIXS7nV6o/dJlYJpTqqIiIiIOBz1pIqIiIhTUE+qc1GSKiIiIk5BSapz0XC/iIiIiDgc9aSKiIiIU1BPqnNRkioiIiLOQVtQORUN94uIiIiIw1FPqoiIiDgFDfc7FyWpIiIi4hSUpDoXJakiIiLiFJSkOhclqSIOZvrY/vTrfn+BdYJbPs+1tAybc14elRj+RDt6drqberWrkZVl4fjpWBZ/9zNzlv1AVpalwGs+cNct/O2xB7nvznpUDfQlMfkaMRcu89O+Pxj7wddkZGbliendpTn9etxPk/AQ/H29SE65xpGTF/hqw17mLd9GekZmnphODzTi7tvDaNqwNnc2qEONakYAbu/2BucvXS3k0xHJKz0jk5/2/sHGHYfYvuc4J2LiSE65RmWjD3ffHsrAR1sR0apxnrjAe54r0vVnRD1F3273lXWzRaQQSlIdQHJyMhMnTmTZsmVER0eTmpoKwL59+2jatGnFNq4MzZ8/n0GDBhEaGkp0dHRFN8fh7fz1BCfPxtkty7wuYQzw92bljH/QJDwEc2IKu/afJCsri+aN6zLp5d50ad2Yvi/MtJs0Arz9Yi9GPNGOtPQM9vwezY59J6gc4EP9usE82689b89cRUZKmk3MnLcG8WjnZmRmZvHzgZNciDNRrbIf991Zj/vuvIVenZvzyLMfkHot3SZu1oSBGP28S/HJiNjavuc4PZ+bBkD1Kv7c37QePp4eHDl1gbU//s7aH39nQM+WvP9aX5uesicKSDzPxl7hx1+OYTAYeODu28r9PUgZ0ep+p+KwSWpUVBTjxo0DwMvLiz/++IOaNWvarRsdHU3dunUB2LJlC23btr1RzSwTjz/+OKtWrQKy32v16tUBqFSpUkU2SyrYp9/8xKJVu4pU971X+9IkPIRDf5ynz/MfcS72KgDVKvvxxdRhtL+/If8a0pUJM77NE/vq0K6MeKIdu/afYMgb84m5eMWm/K5GdUi9rte2W9s7eLRzM0wJyXQf9h9+P37OWhZaqwrfffw8995Rl+F92/J/CzbYxH67ZT8nY+LYfySG/Udi+GPDO0V6jyL5cXEx8HD7pgzr25YH7rrVpuyr9XsYOmYBC1Zs5/4769n0iH4U9VS+13zpnS/58ZdjtL23PnVqVC63tkvZ0nC/c7kptqBKSUmxJqzO5siRI9YE9csvvyQ5OZmLFy9y8eJFbr/99gpundwMgqsaeaRDUwBembzUmqACxF1O4Pm3vgBgRL92+Hp72MTeGhrEi4MiiI030+f5mXkSVIB9h87k6bl9sHl9AJau/cUmQQU4fS6euct+BOCeJnXzXG/kmwt5f/56Nu88TPzVxOK9WRE7HrynPgveHZwnQQV4tHMz+nXPTkwXf/dzka6Xei2d5et/AaD/wy3KrqEiUiw3RZIKMHfuXI4dO1bRzShzv/32GwBVqlShT58+FdwauRnd1agOLi4uXEtL56d9f+QpP/jHeeIuJ+Dt6U6nlra/+DzTqzWV3Fz59OvtmBNTinzPa2nphVcCJaHiEJqEhwBwLjbvL2H2rNz8K6aEFAKN3nRre0d5Nk3KWE5PamkOcRwOO9yfo3bt2gQGBnLgwAFef/11li1bVtFNKlPJyckA+Pr6VnBLxNG0bhZOo1tr4uvtyRVTEnsORrNh+yHS0m2H3n28sntHTQkpWCz2F0fFX02kWmU/mjaozYoNe63n29/fEICf9p7A39eLRzvdze231SIry8LhE+dZuflXLpuS8lxvw0+HGPlURx6LaM6CFdvzDPc/81hrsrKy+PTrn0r9OYiU1smY7Lnd1av6F6n+wpU7AOjT5V483DXt6mZioJTD/ZqU6lAcvifVxcWFiRMnArB8+XJ+/rlowzXXy8zMZO7cubRv356qVavi4eFBrVq16N27N1u3bs03rm3bthgMBqKiorBYLMyaNYv77rsPf39//Pz8aNGiBZ9//nmx2xMVFYXBYGDgwIEAnD592uY3uZzz8L/fDLdu3cqlS5d48cUXCQ8Px9vb2+YfY3JyMosWLeLpp5+madOmVKtWDQ8PD2rWrElkZCRr1qwptD0FzefdunVrob9p7ty5k8jISKpWrYqXlxf169fn3//+N4mJ6lErrie638dzT3ZgYM+WvDCwM59PHsqv34yjQ4uGNvXiriQAEFTFHx8v9zzXMRgM1P7vnLo6NatYz1dyc+W20CAgO7HcvWw077/+BIN7P8jQx9vw/utP8Os343i0U7M81/zxl2NMmbOWAH9vtn72L777+HlmTxjINx+N5Oelo0lLz+DJUZ/wy+/RZfVxiJRI7J9mvvjv3O6H2zcttP6Z8/H8uOc4AP0f0VC/SEVy+CQVoGvXrrRp0waAV199tdjxJpOJjh078re//Y0tW7Zw9epVvL29uXDhAsuWLaNdu3a8/PLLBV4jMzOTnj17MnToUPbu3YvBYCAxMZGdO3fy1FNPMXbs2GK1ydfXl+rVq+Pvn/2bvYuLC9WrV7ceRqMxT8wff/zBHXfcwfvvv09MTAxubrYd4UuWLKFfv3589tlnHDhwgPT0dNzc3Lhw4QLffPMNXbt2ZdSoUcVqZ3HMnTuXli1b8s033xAfH4+HhwfR0dG8/fbb3HPPPVy5UrShtr+634+d419TltLi8QnUbvMSt3V+lZ5/n8au/SeoUc3IF1OH0TLXauM9v0eTlHINgKciH8hzvb7d7rX2tvr5eFnPBxq9cXHJ/i9g0su9uRRvptuw/6N2m5e457HxLPx2J34+nnw8/mlaNL0lz3XfmrmKoaPnk5yaxgN330qviOY8eE99DAYD3/98lMMnL5Tp5yJSXBkZmQwbswBzYgqNbq3JwEdbFRqz8NudWCwW7mpYh8a31boBrZSypOF+53JTJKkA77yTvQJ4y5YtrF27tlixf/vb39i6dSvu7u588MEHmM1mrly5wvnz53nmmWcAmDJlCjNnzsz3GtOnT2fr1q3Mnz8fs9mMyWQiJiaGHj16ADBhwgSOHz9e5DaNGjWKixcv8p///AfIntaQs2Aq9/ncXnjhBQICAti0aRNJSUmYzWaOHj1qLQ8MDGTUqFFs27aNxMRErl69SlJSEufPn2fcuHFUqlSJqVOnsnLlyiK3s6j27t3LsGHDyMrKom3bthw+fJirV6+SmJjIokWLuHjxIuPHjy/z+zqjGYu28MmX33Pk5EUSk6/x55VEtv58hC6D3+e7rftxr+TGxJd6WesnJl9j+sLNAIx59mGGPt6G6lX8qRroy1OPtODdUb2tUwQslv8tgMo9rJV6LZ3Iv0/jp71/kJh8jT9OX+K58Z+zYftB3Nxc+dfQrjZtdHN14cPRT/LJmwNZ/f0BHuj7FjVbvUCzR8cxa8n3PB35AJvnv0LjcP2Ql4rz4juL+X73USobfVjwzt9wr1TwDLesrCy+WLUTgCe1YOrmZCiDQxzGTZOk3n///fTs2ROA1157Ld+5d9fbtWsXy5cvB+DDDz9k5MiReHtn79EYHBzMnDlz6NUr+wf+6NGjrXuUXu/KlSusWLGCAQMG4OWV3RsVEhLC0qVLqVmzJllZWSxZsqRU77EwLi4ubNy4kfbt21t7wMLDw63ljzzyCJMnT6Zly5bW9whQo0YNxowZw9tvvw3ABx98UOZte+ONN8jIyCA8PJzVq1fToEEDIHsbrb59+7J48WKuXr1a6HWuXbuG2Wy2OeR/3vlkNZC9EKRW9QDr+XdnrWbush/x8nTn3VG9ObL2bY6vf4cP3niSA0djWLgy+wfvFXOyNSYx+X/f699u2W93kdOc/67Sb9H0Fiq5uVrP//PpTvR/uAXrt/3O8LGfcvjEBVKupXMyJo5/v/8VC1b8ROUAH9556bEyff8iRfXqlGV89s0OAvy9WTH9OW4NrV5ozNafj3L24hW8PCrxWJfmN6CVIlKQmyZJBXj77bdxdXXl119/ZdGiRUWK+fLLL4HshHLw4MF267z55psA/Pnnn2zYsMFunZYtW9KuXbs85z08PIiIiADgwIEDRWpTST311FOEhISUOL5bt24A7Nixg8xM+5u6l8TVq1dZt24dAC+//LI1ic8tIiKCFi0K75mYOHEiRqPRetSuXbvM2ukMjp66aH1dMyjQ+jory8JL735Jy75vM+Gjb5n/1TamLdzEEy/OpMfwD/Dz8QTg0B/nrTFJKWnEXc6ezxp97k+798s5717JjSoB/1vc98R/t/RZvn6P3bhl63YDcP+dtxTaeyVS1t54/ys+/nIrRj8vvvrw79xRv2j/j3z+3wVTPdo3xeib9/8xcXwa7ncuN1WS2qBBAwYNGgRk93qmpxe+Dc4vv2TvddeuXTtr7+P1GjZsSK1atWzqX++++/J/MknOQwYuX75caHtKo2XLloXWiY2NZezYsbRo0YIqVarg5uZm/YfXqFEjIHuBVVnOD927dy9ZWdnDyO3bt8+3XkFlOV577TVMJpP1iImJKbN2OoPKRh/r69w9oTkOnTjP1HnreGHiYkb/3wrW/vg7FouF++6sB8CWXUds6u8/cgaAKgE+ea51/fmcea8AIcHZCXJCkv2RB3Ni9nlXVxeMfvphLzfOmA++ZvoXm/H39eKrD5/jrkahRYq7Ykpi9ffZHQ1PacHUTUtJqnO5qZJUyF6F7uXlxcmTJwucQ5rj0qVLANYkND85PZQ59a/n5+eXb2zOAqaiJM2lERQUVGD5jh07aNCgAePHj2fnzp1cvnwZLy8vgoKCqF69OlWrVrXWTUrKu61QSeX+zAr6nIvSC+zh4YG/v7/NIf/zaOfslfbmxBSOR8cWKaZnp7upXaMyPx84yf4jtkn/1xv3Admb89v7z7ntfdnTNo5FX7RJSC/EmQBodnuY3Xs2bxxmbaf2SpUbJerDb/jws434+3qxYtpz3H170RJUgKVrd3MtLYO6IVVtFibKzcVgKP0hjuOmS1Jr1arFyJEjgezFSn+lrY1cXV3zLcvIyOCJJ57g6tWrNG3alNWrV2M2m0lISCA2NpaLFy+yc+dOa/2izumVG6txeC0eerAJrq62/zQNBgP9H27B6GezF+p9/OVWMnI9BSq4qtFmjmqOiFaN+b/XnyD1WjovTlycp3zJmt2cjImj0a01eX1YN5tEtVWz23i2X3bv9ydffm8Tt3JTdnI7ol+7PD/QG4fX4vXh3YHsJDgrS99rUv4mzPiW/3y6AaNf8RNUgM9X/m/BlHrTRBzDTTlZ7NVXX2XWrFlcunSJqVOnMmDAgHzrBgUFcfToUc6ePVvgNXPKC+utdFQ7duzg9OnTuLq6smrVKrs9mhcvXrQTmS2nNzi/hWOQvZWXPbk/s3PnzlGvXj279c6dO2f3vPxPnRpVWDhlKFdMSew/GkPc5QSMvl40vKWmda/TZWt/4d1Ztnve3n17KJ9NGszvx89x+nw86RmZ3H5rLerXDSYhKZX+L3/CwVzzUXOkZ2Ty9CuzWDnjn4z6Wxce7dyM346dpUa1AJrdHoqrqwtfrNppXUCVY/KctbRsdhvNbg9j1cf/ZM/BaE6fj6dWUADNbg/Dzc2Vg8fPETXtmzz3HPW3LnRumfeRv19MHWbdheDA0RhGvVu+CxHFeaz+/gBT52bPi68XUo3ZS3+ApXnrVQnw4c3nH81z/sDRGH47dhZXVxfrI1Tl5pTdG1qKzfz1+4lDuSmT1MDAQF599VX+9a9/MXXqVLp27Zpv3ebNm/Pjjz+yZcsWsrKy7M5LPXLkiDWBuueee8qt3eUpZ+5mtWrV8h1y37hxY77xgYGBNtexZ9euXXbP33333bi4uJCVlcWWLVvyTVI3b96c77Ul2+/Hz/HRF5u5q2Edbgutzn131MNgMBB3OYGvN+7li293suGnQ3niDp84z+LVP3NPk7q0vbcBrq4unL14mekLNzN94Sbr8Lw9B/84zwN93+KlZyLo3PJ2urRuTEpqOtv3HmfBip/4akPexVGJydd4aPD7DOrVisgOd9GgXg3urF+b5NQ09h46w7dbfmX20h9IvZZ3CkzdkKrc06RunvN3Nvjf4pZraRl5ykXyczXXrhX7Dp9h3+EzduvVrlHZbpKas2Cq/f0NqVEtoFzaKDdIaYfslaQ6lJsySQUYOXIkH374IWfPnrWuzrenb9++vP/++5w7d47Zs2czdOjQPHXGjBkDQNWqVenYsWO5tbk85Wz+HxsbS2xsLNWr2263cvbs2QK3nrrzzjsBOH/+PLt27cqzUOzSpUvMmjXLbmxAQACdO3dm7dq1TJkyhSeffBJPT0+bOhs3buSnn/SIzMKcOR/Pv9//qthxp87+yd/HFf/JZzli4828Mnkpr0y20/2Uj/SMTD758vs8UwEK8/dxn5eqrSLX69fjfvr1uL/E8ZNe7sOkl/uUYYtEpCzcdHNSc3h5eREVFQXAt99+m2+9e++917oP6siRI5k2bRrJydm/dV+8eJEhQ4awdGn2D+Y333wzT3J1s2jVqhU+Pj5YLBb69OnDsWPHgOwnZa1bt876eNf8PPDAA4SGZs/hGjBgAL/88gsWi4WsrCy2bt1K27ZtrSv47XnzzTdxdXXlyJEjdOvWzfqQgYyMDJYsWUKfPn0ICAgouzcsIiJyHa3udy43bZIKMHDgQOum8QWZM2cObdq0IS0tjZEjR2I0GqlcuTI1a9Zk9uzZQPYToIYPH17eTS43RqORKVOmAPDDDz9Qv359/Pz88PX1pUuXLphMJubNm5dvvIuLCx9//DGVKlXi6NGj3HPPPfj6+uLj40O7du3IyMhg+vTp+cY3b96cjz76CIPBwObNm2nQoAEBAQH4+vry+OOPU716dWuPtYiISHnQ6n7nclMnqa6urtanKBXEaDSyadMm5syZQ9u2bfHz8yMxMZHg4GB69erFli1bmDx58g1ocfkaPnw43333HW3btsXX15eMjAzrbgj79++nSZMmBcZHRETw448/0r17dwIDA8nMzKR27dq8+uqr7Nmzh+Dg4ALjhw4dyvbt2+nRoweVK1fm2rVrhIaG8tprr/Hzzz9b572KiIiIFMZg0V5E4sDMZjNGoxGPJkMwuLpXdHNEysWV3dMqugki5cZsNlO9ihGTyVRue1/n/KwIf/ErXD3sP5ykKDKvJXHsvUfLta1SdDftwikRERGR3Eo7ZK/hfsdyUw/3i4iIiIhzUk+qiIiIOIXSrtDX6n7HoiRVREREnIKG+52LklQRERFxCupJdS6akyoiIiIiDkc9qSIiIuIU1JPqXJSkioiIiFPQnFTnouF+EREREXE46kkVERERp2CglMP9qCvVkagnVURERJxCznB/aY6y8s4771jnyBaWOCckJBAVFUWTJk3w9fXFaDRyzz33MHXqVNLS0gq9V2xsLC+99BL169fHy8uLypUr07p1a2bPno3FYik0/sSJEwwbNoy6devi6elJtWrViIiIYPny5UV6r3v37qV///6EhITg4eFBjRo16NmzJ5s3by5SfH4MlqK0XqSC5DyP2aPJEAyu7hXdHJFycWX3tIpugki5MZvNVK9ixGQy4e/vX273MBqN3PHaSlw9fUp8nczUJA5MfLjUbT169ChNmzYlNTXVei6/dOv06dO0bduW6OhoALy9vcnMzOTatWsA3HXXXWzatInAwEC78Xv27CEiIoL4+HgAfH19SU1NJSMjA4CIiAhWrlyJu7v9n6GrV6+md+/eJCcnA+Dv709iYiJZWVkADBo0iDlz5uSbaM+ePZsRI0ZY72c0GjGbzdb3O3bsWKKiouzGFkY9qSIiIuIUcvdclvQoraysLJ555hlSU1Np0aJFgXUzMjLo0aMH0dHR1KhRgw0bNpCUlERycjKLFy/Gz8+Pffv20b9/f7vxJpOJ7t27Ex8fT4MGDdi9ezcJCQkkJSUxbdo0KlWqxLp163j++eftxp86dYo+ffqQnJxMy5YtOXr0KCaTCZPJxJgxYwCYN28ekydPthu/Y8cOhg8fTkZGBpGRkcTExHD16lXi4uIYNmwYAOPGjWPJkiVF/PRsKUkVERERp+AIw/0ffvghP/30E08++SSdO3cusO6CBQv47bffAFi+fDkdO3YEwMXFhccff5yPP/4YyO7t3LRpU574KVOmcPHiRby8vFi9ejXNmzcHwN3dnb///e+MGzcOgE8++YRjx47liR8zZgxJSUkEBwezatUqwsPDgeze2HHjxjF06FAA3nrrLa5cuZIn/pVXXiEzM5MmTZqwZMkSQkJCAKhSpQozZ84kIiICgH/9619kZmYW8snlpSRVREREpAycOnWKf//731SpUoX333+/0PoLFiwAoF27dnZ7Xfv27UvdunUB+PTTT/OU55zLXS+3kSNH4uvrS2ZmJgsXLrQpS0pKss45HTFiBAEBAXniX3vtNSB7OsXXX39tU3by5Em2bdsGwKhRo6hUqVK+8dHR0fzwww95ygujJFVEREScQkUP9w8ZMoSkpCTee+89qlWrVmDd5ORktm/fDsBDDz2U7/vp0qULAOvXr7cpO3r0KGfOnCkw3tfXl9atW9uN37ZtGykpKQXGh4WF0bBhQ7vxGzZssL7OaeP1WrVqhZ+fn934olCSKiIiIk6hIof7Z82axaZNm+jYsSNPP/10ofUPHz5sXZzUuHHjfOvllF28eJHLly9bz//+++956hQUf+jQIZvzxY0/ePCg3figoCCCgoLsxrq6utKgQQO78UWhJFVEREScQln1pJrNZpsjZ6V9fs6dO8fLL7+Ml5eXdR5pYc6fP299XatWrXzr5S7LHVPceLPZTGJiYp74wMBAvLy8Co3Pfb/cXxd074Lii0JJqoiIiEgutWvXxmg0Wo+JEycWWH/YsGGYTCaioqKoV69eke6RkJBgfe3t7Z1vvdxluWPKKr6g2NzluWPLIr4o9MQpERERcQ6lXaH/39iYmBibfVI9PDzyDfn888/57rvvaNq0KS+++GIpbi7XU5IqIiIiTqG0i59yYv39/Yu0mX9sbCzPP/88rq6uzJo1Cze3oqdVOQuKAOtG+vbkLssdc318fu0tLL6ge+cuzx1bFvFFoeF+ERERkRJ49dVXiY+PZ+jQoTRo0IDExESbI/cjTa8/V7NmTWvZuXPn8r1H7rLcMcWN9/f3x9fXN0/8lStXrKv8C4rPfb/cXxd074Lii0JJqoiIiDiFG726/9SpUwDMmDEDPz+/PEfuuaw551555RUAGjZsiItLdhqWe6X99XLKgoODqVy5svV87hX5RYlv1KiRzfnixt9+++124y9dukRcXJzd2MzMTI4cOWI3viiUpIqIiIhTqOh9UovD29ubli1bArB27Vq7dSwWC+vWrQPI8/Sq8PBw6tSpU2B8UlISP/74o934Vq1aWVf15xd/+vRpDh8+bDe+U6dO1tf5xW/fvt26YKqwp2/ZoyRVREREpAS2bt2KxWLJ9xg7dqy1bs65//u//7OeGzBgAABbtmxh165dea6/dOlSTp48CZBn71WDwWA9t3jxYqKjo/PET58+ncTERFxdXXnyySdtynx8fOjVqxeQ3RNsMpnyxL/77rtAdi9wZGSkTVm9evVo1aoVAFOnTiU9PT1P/DvvvANAaGgoDz74YJ7ywihJFREREadQkZv5l8SAAQNo0qQJFouFXr16sWnTJgCysrJYunQpQ4YMAbKfCNWhQ4c88aNGjSI4OJjk5GS6devGnj17AEhLS2PGjBmMHj0agKFDhxIeHp4nfvz48fj4+HDhwgV69OjB8ePHgewe2PHjxzNz5kwA3njjDQIDA/PEv/vuu7i6urJ//3769u1rnX96+fJlnn32WdasWQPApEmTcHV1Lfbno9X9IiIi4hTKanX/jeLm5sbKlStp164d0dHRdOzYEW9vb7KyskhNTQXgrrvuYuHChXbjjUYjq1atIiIigkOHDtG8eXP8/PxITU219mx27tyZ999/32583bp1WbJkCb179+bHH38kPDwco9FIYmIimZmZAAwaNIiXX37ZbvwDDzzAzJkzGTFiBF999RVfffUVAQEBmEwmLBYLAGPHjqVPnz4l+nzUkyoiIiJSQcLCwjhw4ABjxoyhcePGGAwGKlWqRLNmzZgyZQo7d+6024uZo1mzZhw8eJAXXniB2267jfT0dHx8fGjVqhWzZs1izZo1Be7z2rVrVw4cOMCQIUMICwsjNTWVwMBAOnXqxLJly5g7d26ByfvgwYPZtWsX/fr1o1atWiQnJxMUFERkZCSbNm0iKiqqxJ+NwZKT6oo4ILPZjNFoxKPJEAyu7hXdHJFycWX3tIpugki5MZvNVK9ixGQyFWnv0ZLew2g00uKtdbh5+pT4OhmpSez4d0S5tlWKTsP9IiIi4hRKO6/0Rs9JlYIVKUn99NNPS3WT61ekiYiIiJS1m21OqhSsSEnqwIEDS/UXpyRVRERERIqjSEnq008/rd8uRERExKFpuN+5FClJnT9/fjk3Q0RERKR0NNzvXLQFlYiIiIg4nDJZ3Z+WlkZ8fDweHh5Urly5LC4pIiIiUiwGSjncX2YtkbJQqp7Uzz//nHvvvRcfHx9CQkIYNWqUtWzFihX069ePU6dOlbqRIiIiIoVxMRhKfYjjKHGSOnjwYAYMGMAvv/yCl5cX1z8TIDw8nMWLF7N8+fJSN1JERERE/lpKlKQuXLiQuXPn0rhxY3bv3o3JZMpT5/bbbyckJIQ1a9aUupEiIiIihclZ3V+aQxxHieakfvLJJ/j6+rJq1Spq166db70mTZpw+PDhEjdOREREpKi0ut+5lChJ3b9/P/fdd1+BCSpA5cqViY2NLVHDRERERIrDxZB9lCZeHEeJhvuvXbuG0WgstF5cXByurq4luYWIiIiI/IWVqCe1Vq1ahQ7jWywWDh06RN26dUvUMBEREZFiMZRyyF49qQ6lRD2pHTp04MiRI3zzzTf51vnss884e/YsnTp1KnHjRERERIpKC6ecS4mS1FGjRuHh4UG/fv34v//7P86fP28tu3z5MjNnzuTZZ5/Fx8eHf/zjH2XWWBERERH5ayhRknrbbbexYMECsrKyeOmll6hduzYGg4EFCxZQrVo1/v73v5ORkcH8+fOpU6dOWbdZREREJA9DGfwRx1Hizfx79+7N7t276d27N35+flgsFiwWC56envTo0YMdO3bQq1evsmyriIiISL5yVveX5hDHUaKFUzkaN27M4sWLsVgsxMfHk5WVRdWqVXFxKdXTVkVERETkL65USWoOg8FA1apVy+JSIiIiIiWizfydS6mT1LS0NPbu3cvZs2exWCyEhITQrFkz3N3dy6J9IiIiIkVS2hX6ylEdS4mT1LS0NMaNG8dHH32E2Wy2KfPz82PEiBFERUXh4eFR6kaKiIiIyF9LiZLU1NRUOnfuzPbt27FYLFSpUoWwsDAAoqOjiY+PZ9KkSWzbto0NGzbg6elZlm0WERERycPFYMClFN2hpYmVsleiFU7vvvsu27Zt47bbbuPbb78lLi6O3bt3s3v3buLi4li1ahXh4eH89NNPTJo0qazbLCIiIpKHNvN3LiVKUr/44gt8fX3ZvHkz3bp1y1PetWtXNm3ahLe3NwsXLix1I0VEREQKk7NwqjSHOI4SJalnzpyhXbt21KxZM986NWvWpH379pw5c6bEjRMRERGRv6YSzUkNDAzEy8ur0Hqenp4EBgaW5BYiIiIixaLV/c6lRElqx44d2bhxI9euXct39X5qaio//vgj7du3L1UDRURERIpCC6ecS4mG+ydMmEB6ejr9+vXj0qVLecr//PNP+vfvT3p6Om+99VapGykiIiIify1F6kkdP358nnPdunXj008/Zd26dXTu3Jm6desCcOrUKdavX09KSgpPP/00n376KaNHjy7bVouIiIhcx/DfozTx4jiKlKRGRUVhMBiwWCzWczkr4JKTk/n666/txi1YsACDwaAkVURERMqdHovqXIqUpI4dO7a82yEiIiIiYqUkVURERJyCiyH7KE28OI4Sre4XERERcTQa7ncuJVrdLyIiIiJSnkrVk3rt2jW2bNnC0aNHMZvNNgurcmjhlIiIiNwo6gx1HiVOUlesWMGwYcOIj4/Pt47FYlGSKiIiIjeEhvudS4mS1F9++YXHH38cgL59+3Lw4EF+++03Xn31VY4fP86GDRswm8387W9/IyQkpEwbLCIiImKPFk45lxIlqVOmTCEzM5MVK1bw8MMPM2jQIH777Tfr06Xi4uJ4+umnWbNmDfv27SvTBouIiIiI8yvRwqnt27fTqFEjHn74Ybvl1apVY/HixSQlJTFu3LhSNVBERESkKHKG+0tziOMoUZIaFxdHgwYNrF+7uWV3yKamplrPGY1G2rRpw+rVq0vZRBEREZHCGcrgEMdRoiTVz8+PjIwM69dGoxGA8+fP29SrVKkSFy9eLEXzREREROSvqERJakhICDExMdavc3pVt2zZYj2Xnp7Ozp07qV69eimbKCIiIlI4F4Oh1Ic4jhIlqa1ateLgwYOYTCYAunXrhpubGy+++CLTp0/n22+/5dFHH+X8+fO0bt26TBssIiIiYo/BUPqjJPbu3cu4ceN4+OGHadCgAVWqVKFSpUpUqVKFli1b8tZbb3H58uUCrxEbG8tLL71E/fr18fLyonLlyrRu3ZrZs2fb3Yf+eidOnGDYsGHUrVsXT09PqlWrRkREBMuXLy/ye+jfvz8hISF4eHhQo0YNevbsyebNm4sUv2XLFnr27EmNGjXw8PAgJCSE/v37s3fv3iLF22OwFOWdX2fTpk0MHz6cqVOnWhdPjRs3jnHjxlknHVssFgIDA9m3bx916tQpcQPlr81sNmM0GvFoMgSDq3tFN0ekXFzZPa2imyBSbsxmM9WrGDGZTPj7+5fbPYxGI0/P24G7t2+Jr5OWnMing1oUu63PPfcc06dPt37t6elJpUqVSEhIsJ6rWrUqK1eupEWLFnni9+zZQ0REhHXveV9fX1JTU61TKyMiIli5ciXu7vZ/Dq5evZrevXuTnJwMgL+/P4mJiWRlZQEwaNAg5syZk+/CsNmzZzNixAjr/YxGo81DmsaOHUtUVFS+7z8qKsq6UN5gMODv72/tyHRzc2PGjBkMHjw43/j8lKgntUOHDhw/ftxmdf/YsWNZtmwZffr0oWPHjvzjH/9gz549SlBFRETkhqio1f333nsvkydPZseOHVy5coWUlBTMZjMJCQksWLCAatWq8eeffxIZGWlN3nKYTCa6d+9OfHw8DRo0YPfu3SQkJJCUlMS0adOoVKkS69at4/nnn7d771OnTtGnTx+Sk5Np2bIlR48exWQyYTKZGDNmDADz5s1j8uTJduN37NjB8OHDycjIIDIykpiYGK5evUpcXBzDhg0DsjsilyxZYjd+yZIl1gR12LBhxMXFcfXqVWJiYoiMjCQjI4Phw4ezY8eOYn+uJepJFblR1JMqfwXqSRVndiN7Ugcu2FnqntT5A+4v87auX7+eiIgIAD7//HOefPJJa9no0aOZMGECXl5eHDx4kLp169rETpw4kddffx1XV1cOHTpEeHi4TflTTz3F559/TnBwMIcPHyYgIMCmfNiwYXzyySf4+/sTHR1NYGCgTXnr1q3Ztm0bTZo0Yc+ePVSqVMmmvEuXLqxbt46wsDD++OMPXF1drWWZmZnccsstnD59mi5durBmzRqb2LS0NJo1a8bvv/9Oq1at+PHHH4v1uZWoJ1VEREREiub++++3vj579qxN2aeffgpkP8Hz+gQVYOTIkfj6+pKZmcnChQttypKSkqxzTkeMGJEnQQV47bXXgOxE/uuvv7YpO3nyJNu2bQNg1KhReRLU3PHR0dH88MMPNmXff/89p0+ftqmXm7u7O6NGjQJg27ZtnDp1Kk+dgihJFREREafgqKv7c/cg3nLLLdbXR48e5cyZMwA89NBDdmN9fX2ti9DXr19vU7Zt2zZSUlIKjA8LC6Nhw4Z24zds2GB93aVLF7vxrVq1ws/Pr8B4Pz8/WrZsaTc+d7uujy9MkR6L2r59+2JdNDeDwcCmTZtKHC8iIiJSFKVZoZ8TX1auXbvGhQsXWLVqlXVu6K233kqPHj2sdX7//Xfr68aNG+d7rcaNG7NmzRoOHTpkc7448YcPH+bgwYN244OCgggKCrIb6+rqap0rm198w4YNbaYB5BYUFES1atWIi4vLE1+YIiWpW7duLdZFc9MjxkRERORGKO2jTXNizWazzXkPDw88PDyKdA1PT0+uXbuW53zLli354osvbK6T+yFItWrVyveaOWVms5nExER8fX1t4gMDA/Hy8io0/vqHLuV8XdC9c8p3795dqvi4uLg88YUpUpKae5N+EREREWdWu3Ztm68L24Ipt+DgYFJTU0lMTCQpKQmAdu3aMWnSpDw7HuXeosrb2zvfa+YuS0hIsCapOfEFxeYuz30/R4gvTJGS1DZt2hTroiJl7czWKeW2KlRERJyDC6VbbJMTGxMTY/Mzp6i9qJC9wCjHpUuX+Oyzz3jrrbe49957eeONNxg/fnwpWvjXooVTIiIi4hTKap9Uf39/m6M4SWpuQUFBvPTSS6xduxaDwcCbb77JqlWrrOU5C5IA60b89uQuyx2T87qg2NzluWMdIb4wSlJFREREytG9995Lq1atAPjkk0+s52vWrGl9fe7cuXzjc8r8/f2tQ/2543MeIFBYfO775f66oHuXZ3xhlKSKiIiIUzAYwKUUR3mu9c5ZXPTHH39Yz+VekZ97pf71csoaNWpkc7648bfffrvd+EuXLhEXF2c3NjMzkyNHjhQYf/jwYTIzM+3G57729fGFUZIqIiIiTqE0CWrOUV5OnjwJ2A55h4eHWxdTrV271m5cUlKSdZ/Vzp0725S1atXKuqo/v/jTp09z+PBhu/GdOnWyvs4vfvv27dYFT/nFJyQk8NNPP9mNz33d6+MLoyRVREREpIQyMzMp7AnzmzZt4ueffwagbdu21vMGg4Gnn34agMWLF9ssusoxffp0EhMTcXV1tXmcKoCPjw+9evUCYMaMGZhMpjzx7777LpCdHEdGRtqU1atXzzoNYerUqaSnp+eJf+eddwAIDQ3lwQcftClr06YNoaGhNvVyS09PZ+rUqUB2Qm3viVoFUZIqIiIiTqGsFk4VR0xMDHfddRcff/wxJ0+etElYY2JieOedd3jkkUewWCxUrlyZF154wSZ+1KhRBAcHk5ycTLdu3dizZw+Q/dz7GTNmMHr0aACGDh1KeHh4nvuPHz8eHx8fLly4QI8ePTh+/DiQ3QM7fvx4Zs6cCcAbb7xBYGBgnvh3330XV1dX9u/fT9++fa3zRy9fvsyzzz7LmjVrAJg0aVKeDftdXV2ZNGkSAKtXr+bZZ5/l8uXLQPY81L59+3LgwAGbesVhsBSW/otUILPZjNFoJDbepC2oRERuQmazmepVjJhM5ff/eM7PipFf/oKHt2/hAfm4lpzIh483L1Zbo6OjbXoI3d3d8ff3JyUlxbpPKkDdunVZvnw5d911V55r7Nmzh4iICOLj44HsXs/U1FRrz2bnzp1ZuXJlvrsMrF69mt69e1tX0RuNRhITE63zRAcNGsScOXPyTcJnz57NiBEjyMjIACAgIACTyWRNuAvbJzYqKopx48YB2b8oGI1Grl69CoCbmxszZsxg8ODB+cbnRz2pIiIiIiVUs2ZNli5dyt///neaN29O1apVMZvNZGVlUadOHXr06MHs2bM5ePCg3QQVoFmzZhw8eJAXXniB2267jfT0dHx8fGjVqhWzZs1izZo1BW6D1bVrVw4cOMCQIUMICwsjNTWVwMBAOnXqxLJly5g7d26BvcSDBw9m165d9OvXj1q1apGcnExQUBCRkZFs2rSp0AcZREVFsWnTJiIjIwkKCiI5OZlatWrRr18/du7cWaIEFcqgJ9VkMrF7927i4uIIDQ3lgQceKM3lRGyoJ1VE5OZ2I3tS/7Gk9D2pH/QpXk+qlJ8S96QmJCQwePBggoKCiIiIoH///syePdtaPnv2bGrWrMmuXbvKpKEiIiIiBXExGEp9iOMoUZKakpJC27ZtmTt3LoGBgTz00EN5VrZ1796d2NhYvv7667Jop4iIiEiBXMrgEMdRor+P9957j3379vHEE09w4sQJm0d85QgODqZhw4Zs2bKl1I0UERERkb+WEiWpX375JcHBwcyZMwcfH59864WHh3P27NkSN05ERESkqAyG0h/iONxKEnTixAk6deqEp6dngfW8vb35888/S9QwERERkeJwoXTzSl1QlupIStST6urqavepBNc7e/ZsgT2tIiIiIiL2lKgn9ZZbbmH//v1kZGTg5mb/EomJiRw4cIBGjRqVqoEiIiIiRVHaIXsN9zuWEvWkPvzww1y4cIEJEybkW2fChAmYTCZ69uxZ4saJiIiIFJWLofSHOI4SJakvvPACtWrV4s033yQyMpIvvvgCgNjYWL766iv69u3L5MmTCQsLY/jw4WXaYBERERFxfiUa7g8ICGDt2rU8/PDDrFy5km+//RaDwcDatWtZu3YtFouF0NBQvv32W81JFRERkRvCYKBUC6c03O9YSpSkAjRq1Ijff/+d+fPns3r1ak6ePElWVha1a9fmoYceYujQoXh7e5dlW0VERETypTmpzqXESSqAp6cnw4cP15C+iIiIiJSpUiWpIiIiIo6itIuftHDKsShJFREREadg+O+f0sSL4yhRklqvXr0i1zUYDJw4caIktxEREREpMvWkOpcSJanR0dGF1jEYDFgsFgyahSwiIiIixVSiJPXUqVN2z2dlZXH69GlWrVrFhx9+yGuvvcagQYNK1UARERGRolBPqnMpUZIaGhqab1ndunVp27Yt9913H0888QRt2rQpsL6IiIhIWTAYDKUawdXor2Mp0ROniqJ37940bNiQiRMnltctRERERMRJlVuSCtCwYUN2795dnrcQERERAf433F+aQxxHuW5Bde7cOdLS0srzFiIiIiKAnjjlbMqtJ/Xzzz9nx44dNGrUqLxuISIiIiJOqkQ9qc8880y+ZQkJCRw5coRDhw5hMBj45z//WeLGiYiIiBSVi8GASym6Q0sTK2WvREnq/PnzC63j7+/PuHHj6N+/f0luISIiIlIs2oLKuZQoSZ03b16+Ze7u7tSqVYt7770XT0/PEjdMRERERP66SpSkDhgwoKzbISIiIlI6pVw4hXpSHUqJFk4988wzvPLKK2XdFhEREZESc8FQ6kMcR4mS1M8//zzfR6OKiIiIVIScLahKc4jjKFGSGhwcrEeHiYiIiEi5KVGS2qlTJ7Zv3056enpZt0dERESkRPTEKedSoiQ1KiqKa9euMWTIEBISEsq6TSIiIiLFlrNPamkOcRwl3oKqS5cufPrpp3z33Xd07NiRsLAwvLy88tQ1GAyMHj261A0VERERkb8Og8VisRRWqX379nTp0sW6ot/FxQWDwUBBoTnlBoOBzMzMsmux/KWYzWaMRiOx8Sb8/f0rujkiIlJMZrOZ6lWMmEzl9/94zs+K/2z6DS8fvxJfJyUpgX92aFKubZWiK1JP6tatWwkLC7N+PWbMGC2cEhEREYfiQikfi6otqBxKiYb7o6KiyrgZIiIiIiL/U6IkVURERMTRlHavUw0SOxYlqSIiIuIUXCjhtkW54sVx6O9DRERERBxOkZPUBQsW4OrqWuzDzU2dtSIiIlL+DAZDqQ9xHEXOIIuwU5WIiIhIhTH89yhNvDiOIiepXbp04V//+ld5tkVERESkxEr71Cg9ccqxFDlJDQ4Opk2bNuXZFhERERERQKv7RURExImoL9R5aHW/iIiIOIWcfVJLc5REfHw88+bNo3///jRq1AgfHx88PDwICQkhMjKSFStWFHqNhIQEoqKiaNKkCb6+vhiNRu655x6mTp1KWlpaofGxsbG89NJL1K9fHy8vLypXrkzr1q2ZPXt2kdYVnThxgmHDhlG3bl08PT2pVq0aERERLF++vEifwd69e+nfvz8hISF4eHhQo0YNevbsyebNm4sUb4/BUoSWu7i4MHDgQObOnVviG4mURM7zmGPj9RxlEZGbkdlspnoVIyZT+f0/nvOzYtb3h/D29SvxdZITExjSplGx21qpUiUyMjKsX3t6euLq6kpSUpL13EMPPcSyZcvw9vbOE3/69Gnatm1LdHQ0AN7e3mRmZnLt2jUA7rrrLjZt2kRgYKDd++/Zs4eIiAji4+MB8PX1JTU11dqmiIgIVq5cibu7u9341atX07t3b5KTkwHw9/cnMTGRrKwsAAYNGsScOXPy3f1g9uzZjBgxwno/o9GI2Wy2Jsdjx44t0dNK1ZMqIiIiTqGitqDKyMjg3nvv5aOPPuLEiROkpKSQmJjIqVOn+Nvf/gbAmjVrGDZsmN3YHj16EB0dTY0aNdiwYQNJSUkkJyezePFi/Pz82LdvH/3797d7b5PJRPfu3YmPj6dBgwbs3r2bhIQEkpKSmDZtGpUqVWLdunU8//zzduNPnTpFnz59SE5OpmXLlhw9ehSTyYTJZGLMmDEAzJs3j8mTJ9uN37FjB8OHDycjI4PIyEhiYmK4evUqcXFx1vc7btw4lixZUtyPtWg9qSIVRT2pIiI3txvZkzr3h8Ol7kl95sGGxW7rli1baNeuXb7lw4cP5+OPPwbgzJkz1K5d21o2Z84cBg8eDMBPP/1EixYtbGIXLVpEv379ANi4cSMdOnSwKR89ejQTJkzAy8uLgwcPUrduXZvyiRMn8vrrr+Pq6sqhQ4cIDw+3KX/qqaf4/PPPCQ4O5vDhwwQEBNiUDxs2jE8++QR/f3+io6Pz9Oa2bt2abdu20aRJE/bs2UOlSpVsyrt06cK6desICwvjjz/+wNXVNd/P6XrqSRUREREphYISVMDamwrwyy+/2JQtWLDAeo3rE1SAvn37WhPPTz/9NE95zrnc9XIbOXIkvr6+ZGZmsnDhQpuypKQk65zTESNG5ElQAV577TUg+xeBr7/+2qbs5MmTbNu2DYBRo0blSVBzx0dHR/PDDz/kKS+IklQRERFxCo76xClPT0/r68zMTOvr5ORktm/fDmTPWc3vPXXp0gWA9evX25QdPXqUM2fOFBjv6+tL69at7cZv27aNlJSUAuPDwsJo2LCh3fgNGzZYX+e08XqtWrXCz8/PbnxhlKSKiIiIUzCUwVEetm7dan3dpEkT6+vDhw9bFyc1btw43/icsosXL3L58mXr+d9//z1PnYLiDx06ZHO+uPEHDx60Gx8UFERQUJDdWFdXVxo0aGA3vjBKUkVERERyMZvNNkfOKvuSuHr1KhMnTgSy52/Wr1/fWnb+/Hnr61q1auV7jdxluWOKG282m0lMTMwTHxgYiJeXV6Hxue+X++uC7l1QfGGUpIqIiIhTKKvh/tq1a2M0Gq1HTpJZXFlZWTz11FNcuHABT09Ppk2bZlOekJBgfW1vayp7Zbljyiq+oNjc5bljyyK+MHrilIiIiDgFF0rX+5YTGxMTY7O638PDo0TX++c//8mqVasAmD59OnfccUcpWvfXoyRVREREnEJpFz/lxPr7+5d6u6xRo0ZZe07ff/99nnnmmTx1chYUAdaN9O3JXZY75vr4/NpcWHxB985dnju2LOILo+F+ERERkTL0yiuvMHXqVACmTJmS70b6NWvWtL4+d+5cvtfLXZY7prjx/v7++Pr65om/cuWKdZV/QfG575f764LuXVB8YZSkioiIiFNwhNX9L7/8svXpTJMmTeKll17Kt27Dhg1xcclOxXKvtL9eTllwcDCVK1e2ns+9Ir8o8Y0aNbI5X9z422+/3W78pUuXiIuLsxubmZnJkSNH7MYXRkmqiIiIOAWDofRHaYwaNYopU6YA2Qnqyy+/XGB9b29vWrZsCcDatWvt1rFYLKxbtw6Azp0725SFh4dTp06dAuOTkpL48ccf7ca3atXKuqo/v/jTp09z+PBhu/GdOnWyvs4vfvv27dYFU9fHF0ZJqoiIiEgpjRo1ymaIv7AENceAAQOA7Eer7tq1K0/50qVLOXnyJABPP/20TZnBYLCeW7x4MdHR0Xnip0+fTmJiIq6urjz55JM2ZT4+PvTq1QuAGTNmYDKZ8sS/++67QPZ80sjISJuyevXq0apVKwCmTp1Kenp6nvh33nkHgNDQUB588ME85QVRkioiIiJOwQVDqY+SyD0H9b333itwiP96AwYMoEmTJlgsFnr16sWmTZuA7O2rli5dypAhQ4DsJ0J16NAhT/yoUaMIDg4mOTmZbt26sWfPHgDS0tKYMWMGo0ePBmDo0KGEh4fniR8/fjw+Pj5cuHCBHj16cPz4cSC7B3b8+PHMnDkTgDfeeIPAwMA88e+++y6urq7s37+fvn37WuefXr58mWeffZY1a9YA2T3Lrq6uRf5cAAwWi8VSrAiRG8hsNmM0GomNN5V6paWIiNx4ZrOZ6lWMmEzl9/94zs+KL3ccx9u3eCvIc0tOTODxFrcVq61nzpwhNDQUABcXF6pVq1Zg/VGjRjFq1Cibc9HR0bRr187aE+rt7U1WVhapqakA3HXXXWzatMlukgiwZ88eIiIiiI+PB7J7PVNTU609m507d2blypX5bqW1evVqevfubV2FbzQaSUxMtD7CddCgQcyZMyffnRNmz57NiBEjyMjIACAgIACTyUROijl27FiioqIK/FzsUU+qiIiISAnlPNY053VsbGyBR+4nPuUICwvjwIEDjBkzhsaNG2MwGKhUqRLNmjVjypQp7Ny5M98EFaBZs2YcPHiQF154gdtuu4309HR8fHxo1aoVs2bNYs2aNQXu9dq1a1cOHDjAkCFDCAsLIzU1lcDAQDp16sSyZcuYO3dugVt7DR48mF27dtGvXz9q1apFcnIyQUFBREZGsmnTphIlqKCeVHFw6kkVEbm53cie1CU7/ih1T2qfFreWa1ul6LSZv4iIiDiF0q7QL+3qfilbGu4XEREREYejnlQRERFxCoZSrNDPiRfHoSRVREREnIKG+52LklQRERFxCkpSnYvmpIqIiIiIw1FPqoiIiDgFw3//lCZeHIeSVBEREXEKLobsozTx4jg03C8iIiIiDkc9qSIiIuIUNNzvXJSkioiIiFPQ6n7noiRVxMmlZ2Ty094/2LjjENv3HOdETBzJKdeobPTh7ttDGfhoKyJaNc43fuuuI0z/YjN7D54mOTWN2sGV6dH+Tl4YGIGvt8cNfCciBYu5eJkPPt3I1l1HOBt7BYvFQvWqRh646xae7deeJuEh1rpZWVns/i2aTTsO8cMvxzgWfZGExFT8fb24o34IT3S/n95dmmNQ1iJSYQwWi8VS0Y34q8vMzOQ///kPn332GceOHSM5ORmAFStWEBkZWbGNKyNbt26lXbt2ABTnW85sNmM0GomNN+Hv719ezXNqW3cdoedz0wCoXsWfOxvWxsfTgyOnLnD4xAUABvRsyfuv9c3zA/mjLzbz7/e/wmAw0KLpLQRV8WPHvhPExpu5LbQ6a2a/QJUA3xv+nkSu98vv0Tz63DQSklKpGRTAnQ1q4+riwm/HznL6fDxuri7MmjCQyI53A3AyJo5mj44DINDozV0N62D08+b0uXj2HjoNQESrxnw6aTDuldSfUxpms5nqVYyYTOX3/3jOz4pVv5zCx7fk90hKNNO9ed1ybasUnf7lOYDnn3+eadOykwh3d3eqV68OgKenZ0U2S5yEi4uBh9s3ZVjftjxw1602ZV+t38PQMQtYsGI7999Zj77d7rOWHTgawxv/twJXVxcWTR1Gp5a3A5Ccmka/Fz/m+91HeXHiYha8O/iGvh8Re55/exEJSakM6NmSya/0oZKbK5DdYzrx49VMmbuW599eRJfWTfD0qITBAA82D2fkUx1pd18DXF3/t454+57jPP7CDNZt+53/m7+BV4Y8VFFvS4pJq/udi1b3V7CEhAQ+/vhjACZNmkRqaioXL17k4sWLdOnSpYJbJ87gwXvqs+DdwXkSVIBHOzejX/fsxHTxdz/blL0/fz0Wi4Unu99vTVABvD3d+WD0k7i4GFi5+VeORV8s3zcgUojLVxM5ePwcAG+M6G5NUAFcXFx4dWhXvDwqYUpIsX6/1g2pxjcz/kHHBxrZJKgALZvdxvMDOgOwePWuG/QuROR6SlIr2JEjR0hPTwdgxIgRmv8kN1zOPL1zsVes59LSM1i/7SAAj3VpniemTo3K3HdnPQBWbdl/A1opkj9390pFrlvZWLTpKU3q5/y7uFqSJkkFMZTBH3EcSlIrWM78UwBfX83tkxvvZEwcANWr/m/+1R+nL5GcmgZA04Z17Mbd9d/zvx07W84tFCmYr7cHLe66BYAJM1aRnpFpLcvKyuKdT1aTci2djg80IiQ4sEjXPHkm778LcXw5q/tLc4jjUJJaQebPn4/BYKBt27bWcwaDwXrknA8LC8NgMDB//nwSExMZM2YMTZo0wc/PD4PBQHR0NADp6emsXLmSoUOH0rx5c2rUqIG7uztBQUFERESwaNGifBcs5bQlLCws3/ZGR0db25Zzz+sdOXKEJ598kuDgYDw9PalXrx4jR44kNja2BJ+Q3Aixf5r5YlX2cObD7Ztaz58+Hw+A0c8LPx/7c6NrVc/+YX/6XHz5NlKkCP7zej/CalVlwYrtNH1kLE+O+oSnX5nF3T3H8cFnG3m8673MeWtQka6VnJrGx19uBeDhdk3Lr9FS5gxlcIjj0MKpCuLl5UX16tVJS0vjypXsYdacBVMAlStXtqkfHx9Ps2bNOHbsGO7u7nh7e9uUb9++nUceecT6tb+/P56ensTFxbF+/XrWr1/PihUrWLx4MS4uZf+7ydq1a4mMjOTatWtAdq/whQsXmDZtGsuXL+ett94q83tK6WRkZDJszALMiSk0urUmAx9tZS1LTE4FwMcr/y2mcsoSklLLt6EiRXBbWHXWz32J4WMXsHnnEc5fumota1A3mFbNbsPf16tI1xr1zpecPh9PjWpGXhzUuZxaLCKFUU9qBXn88ce5ePEiX331lfVczoKp688DREVFYTabWbFiBYmJiVy5coWYmBiCgoIA8Pb2ZtiwYWzYsAGTyYTJZMJsNhMfH89//vMf/P39Wbp0qXUXgbJ09uxZHn/8ca5du8Ydd9zBrl27SEhIICkpiTVr1uDq6sqLL75YpGtdu3YNs9lsc0j5ePGdxXy/+yiVjT4seOdv2mZHbmo795+g5RNvc/jEBWZNGMjRtW9zatMkFr03jPTMTEa+uZCRby4s9DqTZ69h0Xe78PSoxNyJz1BZW6zdVFww4GIoxaG+VIeiJPUmkZKSwurVq4mMjKRSpexFAiEhIdYe1XvvvZeZM2fSsWNHm73dKleuzD/+8Q/mzJkDwAcffFDmbXv77bcxm81UqVKFDRs2cO+99wLZq2q7dOnCmjVrSEpKKtK1Jk6ciNFotB61a9cu8/YKvDplGZ99s4MAf29WTH+OW0Or25T7emcP8SelXMv3Gjll+U0HELlRTAnJPPXyLP68ksink4bwWERzgqr4E+DvTZfWTVj2n7/j7enO5yt38OMvx/K9zvSFm3j74+/wcHfjs0lDuP/OW27gu5CyoOF+56Ik9SbRpUsX7rrrrhLHd+vWDYATJ05w8WLZbRlksVj48ssvARg+fLi1Zze3xo0b89hjjxXpeq+99pq1J9hkMhETE1NmbZVsb7z/FR9/uRWjnxdfffh37qif9xeBOjWzp5uYElLyHc7P2Q2gTs0q5ddYkSJYt+0gf15JJKxWFZo3DstTHhZSlWb/Pb/15yN2r/HJl1t54/9W4F7JjQXvDqbjA43KscUiUhRKUm8SLVu2LLROQkICkydPpk2bNgQFBeHu7m5d7JR7DuvZs2W3GvvUqVNcvnwZgPbt2+dbr6Cy3Dw8PPD397c5pOyM+eBrpn+xGX9fL7768DnuahRqt95todXx9nQH4NfDZ+zW2fff83fUD7FbLnKjnL2Y/X9QQb36/r7ZZVfMyXnKZi35nn9NWfbfBPVvBT4mWByculKdiiah3STs9VDmduzYMTp06GCTgHp7exMQEGBdKJWzyr6oQ+9FcenSJevrWrVq5VsvJESJTEWL+vAbPvxsI/6+XqyY9hx3324/QQVwr+RG51a38/XGfSxb+wutm4fblJ+5cJmfD5wCoHu7O8u13SKFqRkUAMDx6FhMiSkYr1sglZ6Ryf4j2aMyodf1/M9d/iOvTF5qTVC7tG5yQ9os5aO0e51qn1THop7Um4Srq2uB5YMGDeLs2bOEhYWxdOlS4uPjSUpK4tKlS1y8eJFz585Z6+a3FZU4rwkzvuU/n27A6Fd4gprj+QGdMRgMLFy1k40/HbKeT05N4x9vLiQzM4uH2zclPCy4PJsuUqiODzTCx8udlGvpPP/WFyQm/28udVp6Bq+/t5yzF69Qyc2VR9r/b9rUghXbGfXuEiWoIg5KPalOICYmhp9++gmARYsWcf/99+epU9A8VDe37G+D1NT8txIymUx2z+fu4T137hz169e3Wy93kiw31urvDzB17joA6oVUY/bSH2Bp3npVAnx48/lHrV/f2aA2E57vyb/f/4o+z8+g5d23Ui3Qjx2//sHFP83cFlqd917re6Pehki+qgb6MfXVvjw3/nO+3riPbXuOc1ejUCq5ufLr4TOcv3QVFxcD74x6jLCQqgD8dvQsL0xcjMViIaxWFVZu+pWVm361e/2Pop66ge9GSqW0G/KrI9WhKEl1ArkXF+W3uGrjxo35xgcGZm/KfunSJa5du4aHR969MXftsv/86rp161K5cmUuX77Mli1b8p17unnz5nzvL+Xraq45ePsOn7HOJb1e7RqVbZJUgGf7tafRLTWZtnAzew9Fk5ySRkhwIC8M7MwLAztrZb84jMe73kujW2syc9EWftp3gh92H8ViyX5iVO8u9zCsbxua3R5mrW9KTLaOKh2LjuVYdP4PHVGSevMo7bRS5aiORUmqEzAajdbX+/fvt24BlSMhIYEJEybkG3/nndlzCi0WCytWrKBvX9vesZSUFN5//327sQaDgT59+jBz5kxmzpzJP//5T6pWrWpT59ChQyxbtqxY70nKTr8e99OvR97e9aJqe18D2t7XoAxbJFI+moSHMH1s0RLKVs3CubK77PeNFpGyozmpTqBhw4bUqZP9HPVnnnmGPXv2WMt27NhB27ZtrU+1sickJIRWrbKfNvTiiy+yceNGMjOzn329Z88eOnbsaLNA6nqvvfYafn5+/Pnnn3Tq1IlffvkFyE56169fz0MPPZTnCVkiIiJlTqv7nYqSVCfg4uLC9OnTcXNz4+DBgzRv3hwfHx98fHx44IEHOHr0qHUv0/x8+OGH+Pv7c+HCBTp16oSvry++vr40b96cEydO8Nlnn+UbW6dOHRYtWoSHhwe//vor99xzD/7+/vj4+BAREUF6ejrvvfdeWb9tERERG4Yy+COOQ0mqk+jevTs//PAD3bp1IyAggIyMDKpWrcqgQYPYs2cPHTp0KDC+adOm7Nq1i759+xIUFERWVhZVq1bl73//O7/++iuNGhW8sXW3bt3Yu3evNT4tLY3q1avz3HPPsW/fPurWrVuWb1dERCQPg6H0hzgOg0X7EYkDM5vNGI1GYuNN2thfROQmZDabqV7FiMlUfv+P5/ys2HogBl+/kt8jMcFM2ztql2tbpei0cEpEREScglb3OxclqSIiIuIclKU6Fc1JFRERERGHo55UERERcQqlXaGv1f2ORUmqiIiIOIXSrtDX6n7HouF+EREREXE46kkVERERp6B1U85FSaqIiIg4B2WpTkXD/SIiIiLicNSTKiIiIk5Bq/udi5JUERERcQpa3e9cNNwvIiIiTsFQBkdJJCcns2bNGiZMmMCjjz5KaGgoBoMBg8FAVFRUka4RGxvLSy+9RP369fHy8qJy5cq0bt2a2bNnY7FYCo0/ceIEw4YNo27dunh6elKtWjUiIiJYvnx5ke6/d+9e+vfvT0hICB4eHtSoUYOePXuyefPmIsVv2bKFnj17UqNGDTw8PAgJCaF///7s3bu3SPH2GCxFeeciFcRsNmM0GomNN+Hv71/RzRERkWIym81Ur2LEZCq//8dzflbsOHQOX7+S3yMxwUyLRrWK3datW7fSrl07u2Vjx44tNFHds2cPERERxMfHA+Dr60tqaioZGRkAREREsHLlStzd3e3Gr169mt69e5OcnAyAv78/iYmJZGVlATBo0CDmzJmDIZ+u4tmzZzNixAjr/YxGI2az2ZocF/YeoqKiGDduHAAGgwF/f39MJhMAbm5uzJgxg8GDBxf4GdijnlQRERFxDhXVlQoEBgbSoUMHXn75ZRYtWkRwcHCR4kwmE927dyc+Pp4GDRqwe/duEhISSEpKYtq0aVSqVIl169bx/PPP240/deoUffr0ITk5mZYtW3L06FFMJhMmk4kxY8YAMG/ePCZPnmw3fseOHQwfPpyMjAwiIyOJiYnh6tWrxMXFMWzYMADGjRvHkiVL7MYvWbLEmqAOGzaMuLg4rl69SkxMDJGRkWRkZDB8+HB27NhRpM8jN/WkikNTT6qIyM3tRvak7jx8vtQ9qfc3rFnstmZmZuLq6mpzLiwsjNOnTxfaCzl69GgmTJiAl5cXBw8epG7dujblEydO5PXXX8fV1ZVDhw4RHh5uU/7UU0/x+eefExwczOHDhwkICLApHzZsGJ988gn+/v5ER0cTGBhoU966dWu2bdtGkyZN2LNnD5UqVbIp79KlC+vWrSMsLIw//vjD5n1mZmZyyy23cPr0abp06cKaNWtsYtPS0mjWrBm///47rVq14scff8z3c7BHPakiIiIipXB9glocn376KQB9+/bNk6ACjBw5El9fXzIzM1m4cKFNWVJSknXO6YgRI/IkqACvvfYakJ3If/311zZlJ0+eZNu2bQCMGjUqT4KaOz46OpoffvjBpuz777/n9OnTNvVyc3d3Z9SoUQBs27aNU6dO5alTECWpIiIi4hRyVveX5riRjh49ypkzZwB46KGH7Nbx9fWldevWAKxfv96mbNu2baSkpBQYHxYWRsOGDe3Gb9iwwfq6S5cuduNbtWqFn59fgfF+fn60bNnSbnzudl0fXxglqSIiIuIUKnBKaon8/vvv1teNGzfOt15O2aFDh0oVf/DgQbvxQUFBBAUF2Y11dXWlQYMGBcY3bNgw397koKAgqlWrZje+MEpSRURERHIxm802x7Vr18rlPufPn7e+rlWrVr71csrMZjOJiYl54gMDA/Hy8io0Pvf9cn9d0L3LM74wSlJFRETEOZRRV2rt2rUxGo3WY+LEieXS3ISEBOtrb2/vfOvlLssdk/O6oNjc5bljHSG+MHrilIiIiDiFsnosakxMjM3qfg8Pj1K3TYpPSaqIiIhILv7+/jdk28OcBUmQ/dSq/O6Zs0n/9TE5r3OXFxSfO9YR4guj4X4RERFxCjfb6v6aNWtaX587dy7fejll/v7++Pr65om/cuWKdZV/QfG575f764LuXZ7xhVGSKiIiIk7hZlvdn3tFfu6V+tfLKWvUqFGp4m+//Xa78ZcuXSIuLs5ubGZmJkeOHCkw/vDhw2RmZtqNz33t6+MLoyRVREREnMNNlqWGh4dTp04dANauXWu3TlJSkvVJTZ07d7Ypa9WqlXVVf37xp0+f5vDhw3bjO3XqZH2dX/z27dutC57yi09ISOCnn36yG5/7utfHF0ZJqoiIiEgFMBgMPP300wAsXryY6OjoPHWmT59OYmIirq6uPPnkkzZlPj4+9OrVC4AZM2ZgMpnyxL/77rtA9nzQyMhIm7J69erRqlUrAKZOnUp6enqe+HfeeQeA0NBQHnzwQZuyNm3aEBoaalMvt/T0dKZOnQpkJ9T2nqhVECWpIiIi4hQMZfCnpK5cucKff/5pPbKysoDsRUO5z+fe5xSyH0caHBxMcnIy3bp1Y8+ePUD2c+9nzJjB6NGjARg6dCjh4eF57jt+/Hh8fHy4cOECPXr04Pjx40B2D+z48eOZOXMmAG+88QaBgYF54t99911cXV3Zv38/ffv2tc4fvXz5Ms8++yxr1qwBYNKkSXk27Hd1dWXSpEkArF69mmeffZbLly8D2fNQ+/bty4EDB2zqFYfBYrFYih0lcoOYzWaMRiOx8aYbstJSRETKltlspnoVIyZT+f0/nvOzYu8fF/HzK/k9EhLM3H1rcInaGhYWZn2OfUEGDBjA/Pnzbc7t2bOHiIgI4uPjgexez9TUVGvPZufOnVm5cmW+W2GtXr2a3r17W1fRG41GEhMTrfNEBw0axJw5czDkszJs9uzZjBgxgoyMDAACAgIwmUzkpIhjx44lKioq3/cUFRXFuHHjgOzeYaPRyNWrVwFwc3NjxowZDB48uJBPJi/1pIqIiIhUoGbNmnHw4EFeeOEFbrvtNtLT0/Hx8aFVq1bMmjWLNWvWFLhXa9euXTlw4ABDhgwhLCyM1NRUAgMD6dSpE8uWLWPu3Ln5JqgAgwcPZteuXfTr149atWqRnJxMUFAQkZGRbNq0qcAEFbKT1E2bNhEZGUlQUBDJycnUqlWLfv36sXPnzhIlqKCeVHFw6kkVEbm53cie1H1l0JN6Vwl7UqXsaTN/ERERcQ6lXaF/o/egkgJpuF9EREREHI56UkVERMQplHaFfmlipewpSRURERGnUNpHm97ox6JKwTTcLyIiIiIORz2pIiIi4hS0bsq5KEkVERER56As1akoSRURERGnoIVTzkVzUkVERETE4agnVURERJyCgVKu7i+zlkhZUJIqIiIiTkFTUp2LhvtFRERExOGoJ1VEREScgjbzdy5KUkVERMRJaMDfmWi4X0REREQcjnpSRURExClouN+5KEkVERERp6DBfuei4X4RERERcTjqSRURERGnoOF+56IkVURERJyC4b9/ShMvjkNJqoiIiDgHTUp1KpqTKiIiIiIORz2pIiIi4hTUkepclKSKiIiIU9DCKeei4X4RERERcTjqSRURERGnoNX9zkVJqoiIiDgHTUp1KhruFxERERGHo55UERERcQrqSHUuSlJFRETEKWh1v3PRcL+IiIiIOBz1pIqIiIiTKN3qfg34OxYlqSIiIuIUNNzvXDTcLyIiIiIOR0mqiIiIiDgcDfeLiIiIU9Bwv3NRkioiIiJOQY9FdS4a7hcRERERh6OeVBEREXEKGu53LkpSRURExCnosajORcP9IiIiIuJw1JMqIiIizkFdqU5FSaqIiIg4Ba3udy4a7hcRERERh6OeVBEREXEKWt3vXJSkioiIiFPQlFTnoiRVREREnIOyVKeiOakiIiIi4nDUkyoiIiJOQav7nYuSVBEREXEKWjjlXJSkikOzWCwAJJjNFdwSEREpiZz/v3P+Py9P5lL+rChtvJQtJani0BISEgC4tW7tCm6JiIiURkJCAkajsVyu7e7uTnBwMLeVwc+K4OBg3N3dy6BVUloGy4341UakhLKysjh//jx+fn4YNA5zQ5jNZmrXrk1MTAz+/v4V3RyRMqfv8RvLYrGQkJBAzZo1cXEpv/XaqamppKWllfo67u7ueHp6lkGLpLTUkyoOzcXFhZCQkIpuxl+Sv7+/foCLU9P3+I1TXj2ouXl6eiq5dDLagkpEREREHI6SVBERERFxOEpSRcSGh4cHY8eOxcPDo6KbIlIu9D0ucnPQwikRERERcTjqSRURERERh6MkVUREREQcjpJUEREREXE4SlJFxEZycjKjR4+mYcOGeHl5YTAYMBgM/PrrrxXdtDI1f/58DAYDYWFhFd2Uv7TMzEzee+897rrrLnx8fKzfb19//XVFN63MbN261fq+RKTolKSKFEFUVJT1h4y3tzfnz5/Pt250dLS17tatW29cI8vI448/zoQJEzhy5AgGg4Hq1atTvXp1KlWqVNFNEyf0/PPP89JLL/Hrr7+SkZFh/X7TpuwioiRVpJhSUlIYN25cRTejXBw5coRVq1YB8OWXX5KcnMzFixe5ePEit99+ewW3TpxNQkICH3/8MQCTJk0iNTXV+v3WpUuXCm6diFQ0JakiJTB37lyOHTtW0c0oc7/99hsAVapUoU+fPhXcGnF2R44cIT09HYARI0ZoOFxEbChJFSmG2rVrc8cdd5CRkcHrr79e0c0pc8nJyQD4+vpWcEvkryDn+w30PScieSlJFSkGFxcXJk6cCMDy5cv5+eefS3SdzMxM5s6dS/v27alatSoeHh7UqlWL3r17FziPtW3bthgMBqKiorBYLMyaNYv77rsPf39//Pz8aNGiBZ9//nmx25Mz53bgwIEAnD592jqvNvd5wGa+7aVLl3jxxRcJDw/H29vbpicsOTmZRYsW8fTTT9O0aVOqVauGh4cHNWvWJDIykjVr1hTanrZt2+ZbpyiLUXbu3ElkZCRVq1bFy8uL+vXr8+9//5vExMQifzZS9nIWreX++839/ZZzPiwsDIPBwPz580lMTGTMmDE0adIEPz8/DAYD0dHRAKSnp7Ny5UqGDh1K8+bNqVGjBu7u7gQFBREREcGiRYvI77k1RVlAl3ueec49r3fkyBGefPJJgoOD8fT0pF69eowcOZLY2NgSfEIiAuBW0Q0Qudl07dqVNm3a8P333/Pqq6+yefPmYsWbTCYiIyOtyairqyt+fn5cuHCBZcuWsWzZMkaNGsXkyZPzvUZmZiY9e/bkm2++wc3NDW9vbxISEti5cyc7d+7k+PHjxZo36+vrS/Xq1UlJScFsNuPi4kK1atWs5UajMU/MH3/8Qd++fYmNjcXT0zPPwqolS5YwaNAgIDsB8ff3x83NjQsXLvDNN9/wzTff8NJLLzFlypQit7M45s6dy5AhQ8jKyrK+h+joaN5++22++uorhg4dWi73lcJ5eXlRvXp10tLSuHLlCgDVq1e3lleuXNmmfnx8PM2aNePYsWO4u7vj7e1tU759+3YeeeQR69f+/v54enoSFxfH+vXrWb9+PStWrGDx4sW4uJR938zatWuJjIzk2rVrQPa/pwsXLjBt2jSWL1/OW2+9Veb3FPlLsIhIocaOHWsBLKGhoRaLxWLZsWOHBbAAljVr1tjUPXXqlLVsy5Ytea7Vq1cvC2Bxd3e3fPDBB5akpCSLxWKxXLhwwfLMM89YY2fMmJEntk2bNhbAEhgYaDEajZb58+dbkpOTLRaLxRITE2Pp0aOHBbC4uLhYjh07Vuz3OW/ePJv3aU9O+3x9fS3169e3bNq0yZKZmWmxWCyWo0ePWut9/fXXllGjRlm2bdtmfY8Wi8Vy/vx5y7hx4yyVKlWyAJZvvvkmzz1yPu82bdrk244tW7ZY23K9PXv2WNzc3CyApW3btpbDhw9bLBaLJS0tzbJo0SJLQECAJSAgoND3KuWroL9Di8ViCQ0NtX6vBQcHW1asWGFJS0uzWCzZ3+8531e7du2yDBs2zLJhwwaLyWSyxsfHx1v+85//WPz9/S2A5T//+U+eexTlez73v+lTp07ZlMXExFivf8cdd1h27dplsVgslszMTMuaNWssISEh1u81/cgVKR79ixEpguuTVIvFYunZs6cFsDRt2tSSlZVlPV9Qkrpz505r2ccff2z3XjlJbNWqVS0pKSk2ZTlJKmDZvHlzntjU1FRLzZo1LYBlwoQJxX6fxUlS/f39LTExMcW+R47JkydbAEuHDh3ylJU2SX3ooYcsgCU8PNyaxOe2du1aa6yS1IpT1CTV1dXVsnfv3hLfZ+nSpRbAcsstt+QpK22SOmLECAtgqVKliiU2NjZP7G+//Wb9hUxJqkjxaE6qSAm9/fbbuLq68uuvv7Jo0aIixXz55ZcAhISEMHjwYLt13nzzTQD+/PNPNmzYYLdOy5YtadeuXZ7zHh4eREREAHDgwIEitamknnrqKUJCQkoc361bNwB27NhBZmZmWTWLq1evsm7dOgBefvllvLy88tSJiIigRYsWZXZPKV9dunThrrvuKnF8zvfaiRMnuHjxYlk1C4vFYv03PXz4cIKCgvLUady4MY899liZ3VPkr0RJqkgJNWjQwDrncvTo0datdAryyy+/ANCuXbt858Y1bNiQWrVq2dS/3n333ZfvPWrWrAnA5cuXC21PabRs2bLQOrGxsYwdO5YWLVpQpUoV3NzcrAtQGjVqBGQvsMqZl1gW9u7da52H2r59+3zrFVQmjqUo32sJCQlMnjyZNm3aEBQUhLu7u80DOHKcPXu2zNp16tQp678zfa+JlD0tnBIphaioKBYuXMjJkyeZOXMmI0eOLLD+pUuXAKxJaH5CQkI4d+6ctf71/Pz88o11c8v+Z12UpLk07PUa5bZjxw66du3K1atXred8fX2tuwBkZmby559/ApCUlETVqlXLpF25P7OCPufS9ALLjVXY99qxY8fo0KGDTQLq7e1NQECA9ZfBnFX2SUlJZdYufa+JlC/1pIqUQq1atayJ6YQJE/5SWxu5urrmW5aRkcETTzzB1atXadq0KatXr8ZsNpOQkEBsbCwXL15k586d1vqWfLYHEoGCv9cABg0axNmzZwkLC2Pp0qXEx8eTlJTEpUuXuHjxIufOnbPW1feayM1DSapIKb366qsEBgZy6dIlpk6dWmDdnB6hwoYcc8oL60FyVDt27OD06dO4urqyatUqHnrooTy9vwXNDczpDU5NTc23jslksns+92eWOzm5XkFlcvOIiYnhp59+AmDRokU89thjebaw0veayM1JSapIKQUGBvLqq68CMHXqVOLi4vKt27x5cwC2bNlinTd5vSNHjlh/qN1zzz1l3NobIyYmBoBq1arlOwy6cePGfOMDAwNtrmPPrl277J6/++67rUO8W7ZsyTe+uPvbimPK/T2S3+KqonyvXbp0ybrP6fXy+16rW7euNSHW95pI2VOSKlIGRo4cSUhICAkJCdbV+fb07dsXyO5ZmT17tt06Y8aMAaBq1ap07Nix7Bt7A+Rs/h8bG2v3iTtnz57lgw8+yDf+zjvvBOD8+fN2E4RLly4xa9Ysu7EBAQF07twZgClTptjtIdu4caO1901ubrkfNLF///485QkJCUyYMCHf+JzvNYvFwooVK/KUp6Sk8P7779uNNRgM9OnTB4CZM2da51jndujQIZYtW1bwmxARu5SkipQBLy8voqKiAPj222/zrXfvvffSq1cvIDuxnTZtmvX55RcvXmTIkCEsXboUyN6KytPTs3wbXk5atWqFj48PFouFPn36cOzYMSD7SVnr1q2zPt41Pw888AChoaEADBgwgF9++QWLxUJWVhZbt26lbdu2+fZEQ/Zn5+rqypEjR+jWrRtHjx4FsufKLlmyhD59+hAQEFB2b1gqTMOGDalTpw4AzzzzDHv27LGW7dixg7Zt2xa4e0RISAitWrUC4MUXX2Tjxo3WLdH27NlDx44d813ACPDaa6/h5+fHn3/+SadOnaw7clgsFtavX89DDz2U5wlZIlI0SlJFysjAgQNp0KBBofXmzJlDmzZtSEtLY+TIkRiNRipXrkzNmjWtvaujRo1i+PDh5d3kcmM0Gq2PO/3hhx+oX78+fn5++Pr60qVLF0wmE/Pmzcs33sXFhY8//phKlSpx9OhR7rnnHnx9ffHx8aFdu3ZkZGQwffr0fOObN2/ORx99hMFgYPPmzTRo0ICAgAB8fX15/PHHqV69urXHWm5uLi4uTJ8+HTc3Nw4ePEjz5s3x8fHBx8eHBx54gKNHj1r3Ms3Phx9+iL+/PxcuXKBTp074+vri6+tL8+bNOXHiBJ999lm+sXXq1GHRokV4eHjw66+/cs899+Dv74+Pjw8RERGkp6fz3nvvlfXbFvlLUJIqUkZcXV15++23C61nNBrZtGkTc+bMoW3btvj5+ZGYmEhwcDC9evViy5YtTJ48+Qa0uHwNHz6c7777jrZt2+Lr60tGRoZ1N4T9+/fTpEmTAuMjIiL48ccf6d69O4GBgWRmZlK7dm1effVV9uzZQ3BwcIHxQ4cOZfv27fTo0YPKlStz7do1QkNDee211/j555+tcxHl5te9e3d++OEHunXrRkBAABkZGVStWpVBgwaxZ88eOnToUGB806ZN2bVrF3379iUoKIisrCyqVq3K3//+d3799Vfrnr756datG3v37rXGp6WlUb16dZ577jn27dtH3bp1y/LtivxlGCzaj0NEREREHIx6UkVERETE4ShJFRERERGHoyRVRERERByOklQRERERcThKUkVERETE4ShJFRERERGHoyRVRERERByOklQRERERcThKUkVERETE4ShJFRERERGHoyRVREolLCwMg8Fgc3h4eFCnTh0ef/xxfvzxx4puolVUVBQGg4GoqCib8/Pnz8dgMDBw4MAKaVdZyO+9FWTr1q0YDAbatm1bYW0ojYEDB2IwGJg/f/4NuZ+I3FhKUkWkTLRs2ZIBAwYwYMAAHnroIbKysliyZAlt2rThvffeq+jm3TA5SXt0dHRFN0VE5KbmVtENEBHnMHjwYJueyNTUVIYNG8ann37KK6+8Qvfu3QkPD6+4BhagZ8+e3H///RiNxopuioiI/Jd6UkWkXHh6ejJ9+nR8fHzIzMzkq6++qugm5ctoNNKgQQNq1KhR0U0REZH/UpIqIuXG19eX+vXrA9gMf+fMXQWYN28eLVq0wGg05hkmP3/+PC+++CINGzbE29sbPz8/7rnnHqZNm0ZGRobde6akpBAVFcVtt92Gh4cHNWrUYMCAAZw5cybfdhY2J/XcuXO8/PLLNGnSBD8/P3x8fAgPD2fgwIH89NNPNtc4ffo0AHXr1rWZp7t161aba96o91ZSGzduZOTIkTRt2pSqVavi4eFBSEgIjz/+OLt37y40/vTp0zz99NPUqFEDT09PwsPDiYqKIiUlJd+YY8eOMWzYMG655RY8PT0xGo08+OCDfP7552X51kTkJqHhfhEpV2azGQAPD488ZSNHjuSjjz7igQceoFu3bpw8edKavP7www9ERkZy5coVwsLC6NSpE9euXePnn39m5MiRfPvtt6xatYpKlSpZr5ecnEyHDh3YuXMnPj4+dO7cGS8vL9atW8d3331Ht27dit3+TZs28dhjj3H16lWCgoLo0KED7u7uREdH88UXXwDwwAMPcOuttzJgwACWLVtGUlISvXr1wtfX13qd4OBg62tHeW8FGT58ODExMdx+++20bNkSNzc3jhw5wpIlS/jqq69YvHgxvXr1sht76tQpmjVrhpubGw8++CApKSls2bKFcePGsXHjRjZu3Iinp6dNzNKlS3n66adJTU2lQYMGdO3aFZPJxK5du3jqqafYvHkzc+fOLdP3KCIOziIiUgqhoaEWwDJv3rw8Zfv377e4uLhYAMvcuXOt5wELYPH397fs2LEjT9yFCxcsVapUsRgMBstHH31kyczMtJb9+eeflvbt21sAy7hx42ziRo0aZQEsDRo0sJw7d856PikpyfLII49Y7zt27FibuHnz5lkAy4ABA2zOnzlzxmI0Gi2A5dVXX7Vcu3bNpjw2Ntby448/2v08Tp06Ze/juuHvrSBbtmyxAJY2bdrkKVuxYoXl8uXLds+7ublZqlSpYklOTrYpGzt2rLUdjzzyiE15TEyMJTw83PpZ5nbgwAGLh4eHxdPT07J8+XKbsujoaEuTJk0sgGXBggU2ZQMGDMj3e09Ebn5KUkWkVOwlqVevXrV89913lltuucUCWGrWrGlJTEy0luckMuPHj7d7zX/9618WwPLcc8/ZLT979qylUqVKlmrVqlmysrIsFovFkpycbPHz87MAljVr1uSJuXDhgsXT07NYSerzzz9vASw9evQowieRrbAk9Ua/t4IUlKQW5IknnrAAlu+++87mfE6S6uXlZblw4UKeuG+//db6y0lKSor1/OOPP24BLFOmTLF7v59//tkCWJo1a2ZzXkmqiHPTnFQRKRODBg2yzr8MCAigW7dunDhxgltuuYXVq1fj4+OTJ+axxx6ze63vvvsOgMcff9xuea1atbjtttuIi4vj+PHjAOzdu5eEhASqVq1Kly5d8sQEBwfTuXPnYr2ntWvXAjB06NBixRXEUd5bUZw/f55Zs2bx0ksvWXdvGDhwIAcPHgTg6NGjduM6d+5sM70hR/fu3alSpQpms5m9e/cCkJWVxZo1a4D8P5PmzZvj6+vLvn37SE1NLYu3JiI3Ac1JFZEy0bJlS2699VYA3N3dCQoK4v7776dLly64udn/ryYsLMzu+ZMnTwLQunXrQu8bFxdHeHg4Z8+eLfCakL2YqThyFkE1aNCgWHEFcZT3Vphx48bx1ltvkZ6enm+dnPnGxWlLWFgY8fHx1vcUHx9vvU7t2rULbVd8fDy1atUqtJ6I3PyUpIpImbh+n9Si8PLysns+KysLyO5ptdcDm1uVKlWKdc+KdjO8t6+++oqoqCh8fX2ZNm0a7du3p2bNmnh5eWEwGHj99deZOHEiFoulxPfIic35PAAGDBhQaJy9BXgi4pyUpIqIw6lduzbHjx/nX//6F82bNy9STE7vWkFPeiruU6Dq1KnD0aNHOXLkiLWXuLQc5b0VZMmSJQC89dZbdqc65ExDyM+pU6fyLctpZ0hICABVq1bFy8uLlJQUpkyZQtWqVUvYahFxNpqTKiIO56GHHgL+lywVRbNmzfD19eXPP/9k/fr1ecpjY2Ptni9IzvzPWbNmFTnG3d0dIN+9Th3lvRXk8uXLAISGhuYpu3TpEhs2bCgwfv369Vy6dCnP+dWrVxMfH4+fnx/NmjUDwNXVlU6dOgHF+0xExPkpSRURh/Pyyy8TEBDAe++9x9SpU0lLS8tT59SpUzabvHt5eVl7/V544QUuXLhgLUtJSWHEiBEFbiRvz4svvoifnx8rV67kjTfeyDM/89KlS2zbts3mXE4PYc7iIkd9bwVp2LAhAJ988olN+0wmEwMGDMBkMhUYb69N58+f56WXXgKy92DNvU/q2LFjcXd35+WXX2bBggU2UwBy/P777w791DIRKXtKUkXE4YSEhPDNN98QGBjIqFGjqF27Nh06dKB///706NGDW2+9lXr16jFt2jSbuPHjx3Pvvfdy6NAhwsPDefjhh+nTpw/16tXjhx9+4Omnny5WO+rUqcOyZcvw8/Pjrbfeonbt2vTs2ZM+ffpw3333ERISwuzZs21icja479+/P7169WLw4MEMHjzYuhLeUd5bQZ5//nkCAgJYvXo19erV47HHHuORRx4hNDSU/fv388wzzxQY//TTT7N161bq1atHnz59ePjhhwkPD+fIkSO0aNGCcePG2dS/++67rUn5wIEDCQ0NJSIigv79+9O1a1dq165NkyZN1NMq8hejJFVEHNKDDz7IwYMHGT16NCEhIezevZulS5fy66+/Ur16dcaOHZtnGN7Hx4ctW7YwevRoqlevzrp16/jhhx/o0KEDv/zyS4lWwHfu3Jnff/+df/7znwQEBLB27VrWrFnD1atXeeqppxg+fLhN/REjRjBx4kRCQ0NZvXo1c+bMYc6cOTa9n47y3vJTt25d9u3bx5NPPomrqyurVq1i//79PPHEE+zbt6/QVfh169bll19+oV27dvzwww+sW7eOGjVqMGbMGDZu3Gh3wVzv3r05ePAgL7zwAgEBAWzfvp3ly5dz6NAhbr31Vt555x3eeuutMnuPIuL4DJbSLM8UERERESkH6kkVEREREYejJFVEREREHI6SVBERERFxOEpSRURERMThKEkVEREREYejJFVEREREHI6SVBERERFxOEpSRURERMThKEkVEREREYejJFVEREREHI6SVBERERFxOEpSRURERMTh/D8ivSKtRVlWPQAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 640x480 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "pipe_lr_weight = make_pipeline(\n",
    "    StandardScaler(), LogisticRegression(max_iter=500, class_weight={0:1, 1: 10})\n",
    ")\n",
    "pipe_lr_weight.fit(X_train, y_train)\n",
    "ConfusionMatrixDisplay.from_estimator(\n",
    "    pipe_lr_weight,\n",
    "    X_valid,\n",
    "    y_valid,\n",
    "    display_labels=[\"Non fraud\", \"fraud\"],\n",
    "    values_format=\"d\",\n",
    "    cmap=plt.cm.Blues,\n",
    ");"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "- Notice we've reduced false negatives and predicted more Fraud this time.\n",
    "- This was equivalent to saying give 10x more \"importance\" to fraud class. \n",
    "- Note that as a consequence we are also increasing false positives.    "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "### `class_weight=\"balanced\"`\n",
    "- A useful setting is `class_weight=\"balanced\"`.\n",
    "- This sets the weights so that the classes are \"equal\".\n",
    "\n",
    "> class_weight: dict, ‘balanced’ or None\n",
    "If ‘balanced’, class weights will be given by n_samples / (n_classes * np.bincount(y)). If a dictionary is given, keys are classes and values are corresponding class weights. If None is given, the class weights will be uniform.\n",
    "\n",
    "> sklearn.utils.class_weight.compute_class_weight(class_weight, classes, y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 66,
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 640x480 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "pipe_lr_balanced = make_pipeline(\n",
    "    StandardScaler(), LogisticRegression(max_iter=500, class_weight=\"balanced\")\n",
    ")\n",
    "pipe_lr_balanced.fit(X_train, y_train)\n",
    "ConfusionMatrixDisplay.from_estimator(\n",
    "    pipe_lr_balanced,\n",
    "    X_valid,\n",
    "    y_valid,\n",
    "    display_labels=[\"Non fraud\", \"fraud\"],\n",
    "    values_format=\"d\",\n",
    "    cmap=plt.cm.Blues,\n",
    ");"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We have reduced false negatives but we have many more false positives now ..."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "### Are we doing better with `class_weight=\"balanced\"`?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 67,
   "metadata": {},
   "outputs": [],
   "source": [
    "comp_dict = {}\n",
    "pipe_lr = make_pipeline(StandardScaler(), LogisticRegression(max_iter=500))\n",
    "scoring = [\"accuracy\", \"f1\", \"recall\", \"precision\", \"roc_auc\", \"average_precision\"]\n",
    "orig_scores = cross_validate(pipe_lr, X_train_big, y_train_big, scoring=scoring)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 68,
   "metadata": {
    "slideshow": {
     "slide_type": "-"
    }
   },
   "outputs": [],
   "source": [
    "pipe_lr_balanced = make_pipeline(\n",
    "    StandardScaler(), LogisticRegression(max_iter=500, class_weight=\"balanced\")\n",
    ")\n",
    "scoring = [\"accuracy\", \"f1\", \"recall\", \"precision\", \"roc_auc\", \"average_precision\"]\n",
    "bal_scores = cross_validate(pipe_lr_balanced, X_train_big, y_train_big, scoring=scoring)\n",
    "comp_dict = {\n",
    "    \"Original\": pd.DataFrame(orig_scores).mean().tolist(),\n",
    "    \"class_weight='balanced'\": pd.DataFrame(bal_scores).mean().tolist(),\n",
    "}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 69,
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Original</th>\n",
       "      <th>class_weight='balanced'</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>fit_time</th>\n",
       "      <td>0.481765</td>\n",
       "      <td>0.420509</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>score_time</th>\n",
       "      <td>0.036980</td>\n",
       "      <td>0.037592</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>test_accuracy</th>\n",
       "      <td>0.999182</td>\n",
       "      <td>0.973210</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>test_f1</th>\n",
       "      <td>0.714290</td>\n",
       "      <td>0.101817</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>test_recall</th>\n",
       "      <td>0.601712</td>\n",
       "      <td>0.891001</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>test_precision</th>\n",
       "      <td>0.881915</td>\n",
       "      <td>0.054009</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>test_roc_auc</th>\n",
       "      <td>0.967228</td>\n",
       "      <td>0.970505</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>test_average_precision</th>\n",
       "      <td>0.744074</td>\n",
       "      <td>0.730368</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                        Original  class_weight='balanced'\n",
       "fit_time                0.481765                 0.420509\n",
       "score_time              0.036980                 0.037592\n",
       "test_accuracy           0.999182                 0.973210\n",
       "test_f1                 0.714290                 0.101817\n",
       "test_recall             0.601712                 0.891001\n",
       "test_precision          0.881915                 0.054009\n",
       "test_roc_auc            0.967228                 0.970505\n",
       "test_average_precision  0.744074                 0.730368"
      ]
     },
     "execution_count": 69,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pd.DataFrame(comp_dict, index=bal_scores.keys())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "- Recall is much better but precision has dropped a lot; we have many false positives. \n",
    "- You could also optimize `class_weight` using hyperparameter optimization for your specific problem. "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "- Changing the class weight will **generally reduce accuracy**.\n",
    "- The original model was trying to maximize accuracy.\n",
    "- Now you're telling it to do something different.\n",
    "- But that can be fine, accuracy isn't the only metric that matters."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "### Stratified Splits\n",
    "\n",
    "- A similar idea of \"balancing\" classes can be applied to data splits.\n",
    "- We have the same option in `train_test_split` with the `stratify` argument. \n",
    "- By default it splits the data so that if we have 10% negative examples in total, then each split will have 10% negative examples."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "- If you are carrying out cross validation using `cross_validate`, by default it uses [`StratifiedKFold`](https://scikit-learn.org/stable/modules/generated/sklearn.model_selection.StratifiedKFold.html). From the documentation: \n",
    "\n",
    "> This cross-validation object is a variation of KFold that returns stratified folds. The folds are made by preserving the percentage of samples for each class.\n",
    "\n",
    "- In other words, if we have 10% negative examples in total, then each fold will have 10% negative examples."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "### Is stratifying a good idea? \n",
    "\n",
    "  - Well, it's no longer a random sample, which is probably theoretically bad, but not that big of a deal.\n",
    "  - If you have many examples, it shouldn't matter as much.\n",
    "  - It can be especially useful in multi-class, say if you have one class with very few cases.\n",
    "  - In general, these are difficult questions."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "<br><br><br><br>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "## What did we learn today? \n",
    "\n",
    "- A number of possible ways to evaluate machine learning models \n",
    "    - Choose the evaluation metric that makes most sense in your context or which is most common in your discipline  \n",
    "- Two kinds of binary classification problems \n",
    "    - Distinguishing between two classes (e.g., dogs vs. cats)\n",
    "    - Spotting a class (e.g., spot fraud transaction, spot spam)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "- Precision, recall, f1-score are useful when dealing with spotting problems. \n",
    "- The thing that we are interested in spotting is considered \"positive\".   \n",
    "- Do you need to deal with class imbalance in the given problem? \n",
    "- Methods to deal with class imbalance \n",
    "    - Changing the training procedure \n",
    "        - `class_weight`"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "### Relevant papers and resources \n",
    "\n",
    "- [The Relationship Between Precision-Recall and ROC Curves](https://www.biostat.wisc.edu/~page/rocpr.pdf)\n",
    "- [Article claiming that PR curve are better than ROC for imbalanced datasets](https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0118432)\n",
    "- [Precision-Recall-Gain Curves: PR Analysis Done Right](https://papers.nips.cc/paper/2015/file/33e8075e9970de0cfea955afd4644bb2-Paper.pdf)\n",
    "- [ROC animation](https://github.com/dariyasydykova/open_projects/tree/master/ROC_animation)\n",
    "- [Generalization in Adaptive Data Analysis and Holdout Reuse](https://arxiv.org/pdf/1506.02629.pdf)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "![](img/eva-seeyou.png)"
   ]
  }
 ],
 "metadata": {
  "celltoolbar": "Slideshow",
  "kernelspec": {
   "display_name": "Python [conda env:cpsc330]",
   "language": "python",
   "name": "conda-env-cpsc330-py"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.10.6"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}