DSM_v_SMOTE_blog_Notebook.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "d9285d13-7c82-415b-baa8-24f0a15cc95a",
   "metadata": {},
   "source": [
    "# Documentation of Experiments to Accompany \"How to Solve the Problem of Imbalanced Datasets: Meet Tonic Data Science Mode\" Blog Post"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "06907021-b6db-4fc6-a42f-5721d1ec8b34",
   "metadata": {},
   "source": [
    "*NB: This notebook was run separately from the one referenced in the blog post - numbers from distribution analysis may differ slightly*"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "db67d294-6598-467d-9dbd-b36e9ecab3cb",
   "metadata": {
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Requirement already satisfied: catboost in /Users/madelyntonic/miniforge3/envs/tenswork/lib/python3.10/site-packages (1.1.1)\n",
      "Requirement already satisfied: graphviz in /Users/madelyntonic/miniforge3/envs/tenswork/lib/python3.10/site-packages (from catboost) (0.20.1)\n",
      "Requirement already satisfied: matplotlib in /Users/madelyntonic/miniforge3/envs/tenswork/lib/python3.10/site-packages (from catboost) (3.7.1)\n",
      "Requirement already satisfied: numpy>=1.16.0 in /Users/madelyntonic/miniforge3/envs/tenswork/lib/python3.10/site-packages (from catboost) (1.24.3)\n",
      "Requirement already satisfied: pandas>=0.24.0 in /Users/madelyntonic/miniforge3/envs/tenswork/lib/python3.10/site-packages (from catboost) (2.0.2)\n",
      "Requirement already satisfied: scipy in /Users/madelyntonic/miniforge3/envs/tenswork/lib/python3.10/site-packages (from catboost) (1.10.1)\n",
      "Requirement already satisfied: plotly in /Users/madelyntonic/miniforge3/envs/tenswork/lib/python3.10/site-packages (from catboost) (5.14.1)\n",
      "Requirement already satisfied: six in /Users/madelyntonic/miniforge3/envs/tenswork/lib/python3.10/site-packages (from catboost) (1.16.0)\n",
      "Requirement already satisfied: python-dateutil>=2.8.2 in /Users/madelyntonic/miniforge3/envs/tenswork/lib/python3.10/site-packages (from pandas>=0.24.0->catboost) (2.8.2)\n",
      "Requirement already satisfied: pytz>=2020.1 in /Users/madelyntonic/miniforge3/envs/tenswork/lib/python3.10/site-packages (from pandas>=0.24.0->catboost) (2023.3)\n",
      "Requirement already satisfied: tzdata>=2022.1 in /Users/madelyntonic/miniforge3/envs/tenswork/lib/python3.10/site-packages (from pandas>=0.24.0->catboost) (2023.3)\n",
      "Requirement already satisfied: contourpy>=1.0.1 in /Users/madelyntonic/miniforge3/envs/tenswork/lib/python3.10/site-packages (from matplotlib->catboost) (1.0.7)\n",
      "Requirement already satisfied: cycler>=0.10 in /Users/madelyntonic/miniforge3/envs/tenswork/lib/python3.10/site-packages (from matplotlib->catboost) (0.11.0)\n",
      "Requirement already satisfied: fonttools>=4.22.0 in /Users/madelyntonic/miniforge3/envs/tenswork/lib/python3.10/site-packages (from matplotlib->catboost) (4.39.3)\n",
      "Requirement already satisfied: kiwisolver>=1.0.1 in /Users/madelyntonic/miniforge3/envs/tenswork/lib/python3.10/site-packages (from matplotlib->catboost) (1.4.4)\n",
      "Requirement already satisfied: packaging>=20.0 in /Users/madelyntonic/miniforge3/envs/tenswork/lib/python3.10/site-packages (from matplotlib->catboost) (23.0)\n",
      "Requirement already satisfied: pillow>=6.2.0 in /Users/madelyntonic/miniforge3/envs/tenswork/lib/python3.10/site-packages (from matplotlib->catboost) (9.5.0)\n",
      "Requirement already satisfied: pyparsing>=2.3.1 in /Users/madelyntonic/miniforge3/envs/tenswork/lib/python3.10/site-packages (from matplotlib->catboost) (3.0.9)\n",
      "Requirement already satisfied: tenacity>=6.2.0 in /Users/madelyntonic/miniforge3/envs/tenswork/lib/python3.10/site-packages (from plotly->catboost) (8.2.2)\n"
     ]
    }
   ],
   "source": [
    "#importing packages\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "import scipy.stats as st\n",
    "\n",
    "from sklearn.ensemble import RandomForestClassifier\n",
    "from sklearn.linear_model import LogisticRegression\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.preprocessing import OneHotEncoder, OrdinalEncoder\n",
    "from sklearn import metrics\n",
    "from sklearn.metrics import roc_auc_score\n",
    "import xgboost as xgb\n",
    "#may need to install catboost\n",
    "!pip install catboost\n",
    "from catboost import CatBoostClassifier\n",
    "\n",
    "from sklearn.metrics import roc_curve, auc\n",
    "from sklearn.metrics import confusion_matrix\n",
    "import seaborn as sns\n",
    "from sklearn.metrics import f1_score\n",
    "\n",
    "from imblearn.over_sampling import SMOTE, SMOTENC\n",
    "\n",
    "pd.set_option('display.max_rows', 100)\n",
    "pd.set_option('display.max_columns', 100)\n",
    "\n",
    "import warnings\n",
    "warnings.filterwarnings(\"ignore\")\n",
    "\n",
    "xgb.verbosity = 0"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "c83adc8e-130d-49b0-9d2f-d337a2306da2",
   "metadata": {},
   "source": [
    "## Data Import and Cleaning"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "c168e10a-0ff6-44a0-b81e-4ad38b1d9037",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Churn value counts:\n",
      "No     5174\n",
      "Yes    1869\n",
      "Name: Churn, dtype: int64\n",
      "\n",
      "Churn percentages:\n",
      "No     73.46\n",
      "Yes    26.54\n",
      "Name: Churn, dtype: float64\n"
     ]
    }
   ],
   "source": [
    "#value counts and percentages of churn from the raw data\n",
    "raw = pd.read_csv('telco.csv', low_memory=False)\n",
    "print('Churn value counts:')\n",
    "print(raw['Churn'].value_counts())\n",
    "print('\\nChurn percentages:')\n",
    "print(round((raw['Churn'].value_counts()/raw.shape[0])*100, 2))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "c67c5fb3-f4e3-4681-b945-f37c1821be21",
   "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>n_values</th>\n",
       "      <th>dtype</th>\n",
       "      <th>null_cnt</th>\n",
       "      <th>null_pct</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>customerID</th>\n",
       "      <td>7043</td>\n",
       "      <td>object</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>gender</th>\n",
       "      <td>2</td>\n",
       "      <td>object</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>SeniorCitizen</th>\n",
       "      <td>2</td>\n",
       "      <td>int64</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Partner</th>\n",
       "      <td>2</td>\n",
       "      <td>object</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Dependents</th>\n",
       "      <td>2</td>\n",
       "      <td>object</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>tenure</th>\n",
       "      <td>73</td>\n",
       "      <td>int64</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>PhoneService</th>\n",
       "      <td>2</td>\n",
       "      <td>object</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>MultipleLines</th>\n",
       "      <td>3</td>\n",
       "      <td>object</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>InternetService</th>\n",
       "      <td>3</td>\n",
       "      <td>object</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>OnlineSecurity</th>\n",
       "      <td>3</td>\n",
       "      <td>object</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>OnlineBackup</th>\n",
       "      <td>3</td>\n",
       "      <td>object</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>DeviceProtection</th>\n",
       "      <td>3</td>\n",
       "      <td>object</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>TechSupport</th>\n",
       "      <td>3</td>\n",
       "      <td>object</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>StreamingTV</th>\n",
       "      <td>3</td>\n",
       "      <td>object</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>StreamingMovies</th>\n",
       "      <td>3</td>\n",
       "      <td>object</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Contract</th>\n",
       "      <td>3</td>\n",
       "      <td>object</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>PaperlessBilling</th>\n",
       "      <td>2</td>\n",
       "      <td>object</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>PaymentMethod</th>\n",
       "      <td>4</td>\n",
       "      <td>object</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>MonthlyCharges</th>\n",
       "      <td>1585</td>\n",
       "      <td>float64</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>TotalCharges</th>\n",
       "      <td>6531</td>\n",
       "      <td>object</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Churn</th>\n",
       "      <td>2</td>\n",
       "      <td>object</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                  n_values    dtype  null_cnt  null_pct\n",
       "customerID            7043   object         0       0.0\n",
       "gender                   2   object         0       0.0\n",
       "SeniorCitizen            2    int64         0       0.0\n",
       "Partner                  2   object         0       0.0\n",
       "Dependents               2   object         0       0.0\n",
       "tenure                  73    int64         0       0.0\n",
       "PhoneService             2   object         0       0.0\n",
       "MultipleLines            3   object         0       0.0\n",
       "InternetService          3   object         0       0.0\n",
       "OnlineSecurity           3   object         0       0.0\n",
       "OnlineBackup             3   object         0       0.0\n",
       "DeviceProtection         3   object         0       0.0\n",
       "TechSupport              3   object         0       0.0\n",
       "StreamingTV              3   object         0       0.0\n",
       "StreamingMovies          3   object         0       0.0\n",
       "Contract                 3   object         0       0.0\n",
       "PaperlessBilling         2   object         0       0.0\n",
       "PaymentMethod            4   object         0       0.0\n",
       "MonthlyCharges        1585  float64         0       0.0\n",
       "TotalCharges          6531   object         0       0.0\n",
       "Churn                    2   object         0       0.0"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#checking data for cleanliness\n",
    "n_values = raw.nunique()\n",
    "dtype = raw.dtypes\n",
    "null_cnt = raw.isna().sum()\n",
    "null_pct = raw.isna().sum()/raw.shape[0]\n",
    "info = pd.concat([n_values, dtype, null_cnt, null_pct], axis=1)\n",
    "info.columns = ['n_values', 'dtype', 'null_cnt', 'null_pct']\n",
    "info"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "d675e053-740c-4a03-a038-b4e7a6efe91b",
   "metadata": {},
   "outputs": [],
   "source": [
    "#cleaning and splitting data\n",
    "raw_cleaned = raw.copy()\n",
    "\n",
    "def clean_total_charges(x):\n",
    "    try:\n",
    "        return float(x)\n",
    "    except:\n",
    "        return np.nan\n",
    "\n",
    "raw_cleaned['TotalCharges'] = raw['TotalCharges'].apply(clean_total_charges)\n",
    "\n",
    "#dropping rows with NaNs\n",
    "data_cleaned = raw_cleaned.dropna().reset_index(drop=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "1124ff54-5ac5-492e-9082-e264ff3ad603",
   "metadata": {},
   "source": [
    "## Modeling prep"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "65e5d838-6fe4-405d-a266-e3aabd6ae57f",
   "metadata": {},
   "outputs": [],
   "source": [
    "#splitting and saving data \n",
    "random_state = 0\n",
    "data_train, data_test = train_test_split(data_cleaned, test_size=0.25, random_state=random_state)\n",
    " \n",
    "data_train.to_csv('telco_train.csv', index=False)\n",
    "data_test.to_csv('telco_test.csv', index=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "795b31cc-b044-4aa5-bec9-e86154385ee7",
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "#reading test/train data previously split in data\n",
    "data_train = pd.read_csv('telco_train.csv')\n",
    "data_test = pd.read_csv('telco_test.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "97e11038-8fe1-45bb-90a8-1a6233e239d8",
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "#defining inputs and outputs\n",
    "target = 'Churn'\n",
    "modeling_columns = [x for x in data_train.columns if x != 'customerID']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "ebd5989a-f9a5-4dc8-a857-dcbdfd88f2c0",
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "#defining columns to encode\n",
    "categorical_columns = [x for x in modeling_columns if data_train[x].dtype == 'object']\n",
    "input_columns = [x for x in modeling_columns if x != target]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "fdfe4157-03be-42ff-ae3e-204ad87bfc19",
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "#defining the features of the data\n",
    "X = data_train[input_columns].copy()\n",
    "\n",
    "#instatiating encoder\n",
    "feature_dimensions_X = []\n",
    "for col in X.columns:\n",
    "    if col in categorical_columns:\n",
    "        feature_dimensions_X.append(X[col].nunique())\n",
    "    else:\n",
    "        feature_dimensions_X.append(1)\n",
    "\n",
    "column_to_one_hot_encoder = {}\n",
    "for col in X.columns:\n",
    "    if col in categorical_columns:\n",
    "        #instatiating the encoder\n",
    "        encoder = OneHotEncoder(sparse=False)\n",
    "        #training the encoder\n",
    "        encoder.fit(X[[col]])\n",
    "        column_to_one_hot_encoder[col] = encoder"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "0201fff9-803b-4bed-9877-70d7ac8011fc",
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "#defining one hot encoder function\n",
    "def one_hot_encode(data, column_to_one_hot_encoder):\n",
    "    categorical_columns = list(column_to_one_hot_encoder.keys())\n",
    "    oh_encoded_list = []\n",
    "    for col in data.columns:\n",
    "        if col in categorical_columns:\n",
    "            encoder = column_to_one_hot_encoder[col]\n",
    "            #implementing the encoder\n",
    "            col_ohe = encoder.transform(data[[col]])\n",
    "            ohe_column_names = [col + '_' + x for x in list(encoder.categories_[0])]\n",
    "            col_ohe_df = pd.DataFrame(col_ohe, columns=ohe_column_names)\n",
    "            oh_encoded_list.append(col_ohe_df)\n",
    "        else:\n",
    "            oh_encoded_list.append(data[[col]].reset_index(drop=True))\n",
    "    \n",
    "    data_oh_encoded = pd.concat(oh_encoded_list, axis=1)\n",
    "    return data_oh_encoded"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "2ccbf2c9-8057-4af6-84e6-919c4729a1ff",
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "#defining and transforming test and train datasets using the above function\n",
    "#defining test and train dataframes\n",
    "X_train = data_train[input_columns].copy()\n",
    "X_test = data_test[input_columns].copy()\n",
    "\n",
    "y_train = data_train[target].copy()\n",
    "y_test = data_test[target].copy()\n",
    "\n",
    "#encoding churn column for modeling\n",
    "y_train_encoded = y_train.apply(lambda x : int(x == 'Yes'))\n",
    "y_test_encoded = y_test.apply(lambda x : int(x == 'Yes'))\n",
    "\n",
    "#onehot encoding inputs\n",
    "X_train_oh_encoded = one_hot_encode(X_train, column_to_one_hot_encoder)\n",
    "X_test_oh_encoded = one_hot_encode(X_test, column_to_one_hot_encoder)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "0387e0ef-b5b6-432e-8f16-162aa02ee96f",
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "#defining models\n",
    "log_reg = LogisticRegression(random_state=7001)\n",
    "xgb_class = xgb.XGBClassifier(use_label_encoder=False, eval_metric='logloss', random_state=7001)\n",
    "catboost = CatBoostClassifier(silent=True, random_state=7001)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "113f17f6-bb3f-4ca0-9465-9767d6ddef81",
   "metadata": {},
   "source": [
    "## Modeling on Imbalanced Dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "71fa4e02-b9b7-4bcd-a47e-6ab56680de8f",
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "#running models and defining predictions - will use these later to evaluate individual models\n",
    "\n",
    "#log reg\n",
    "#model\n",
    "unaug_log_reg = log_reg.fit(X_train_oh_encoded, y_train_encoded)\n",
    "#y proba predictions\n",
    "unaug_log_reg_pred_proba = unaug_log_reg.predict_proba(X_test_oh_encoded)[:,1]\n",
    "#y predictions \n",
    "unaug_log_reg_pred = unaug_log_reg.predict(X_test_oh_encoded)\n",
    "\n",
    "#xgb\n",
    "#model\n",
    "unaug_xgb = xgb_class.fit(X_train_oh_encoded, y_train_encoded)\n",
    "#y proba predictions\n",
    "unaug_xgb_pred_proba = unaug_xgb.predict_proba(X_test_oh_encoded)[:,1]\n",
    "#y predictions \n",
    "unaug_xgb_pred = unaug_xgb.predict(X_test_oh_encoded)\n",
    "\n",
    "#catb\n",
    "#dropping churn from categorical columns\n",
    "#categorical_columns.remove('Churn')\n",
    "#model\n",
    "unaug_catboost = catboost.fit(X_train, y_train_encoded, cat_features=[col for col in categorical_columns if col!=target])\n",
    "#y proba predictions\n",
    "unaug_catboost_pred_proba = unaug_catboost.predict_proba(X_test)[:,1]\n",
    "#y predictions\n",
    "unaug_catboost_pred = unaug_catboost.predict(X_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "7e8557d3-a40d-4088-8f57-97860b8d0549",
   "metadata": {
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Baseline ROC AUC scores modeling with imbalanced data: \n",
      "\n",
      "Logistic Regression Baseline ROC AUC: 0.8358537380314905\n",
      "XGBoost Baseline ROC AUC: 0.816745077361237\n",
      "CatBoost Baseline ROC AUC: 0.8404286906352546\n",
      "\n",
      " Baseline F1 scores modeling with imbalanced data: \n",
      "\n",
      "Logistic F1: 0.5854858548585485\n",
      "XGBoost F1: 0.5247524752475248\n",
      "CatBoost F1: 0.5717884130982368\n"
     ]
    }
   ],
   "source": [
    "#scoring models \n",
    "print('Baseline ROC AUC scores modeling with imbalanced data: \\n')\n",
    "print(f'Logistic Regression Baseline ROC AUC: {(roc_auc_score(y_test_encoded, unaug_log_reg_pred_proba))}')\n",
    "print(f'XGBoost Baseline ROC AUC: {(roc_auc_score(y_test_encoded, unaug_xgb_pred_proba))}')\n",
    "print(f'CatBoost Baseline ROC AUC: {(roc_auc_score(y_test_encoded, unaug_catboost_pred_proba))}')\n",
    "\n",
    "print('\\n Baseline F1 scores modeling with imbalanced data: \\n')\n",
    "print(f\"Logistic F1: {f1_score(y_test_encoded, unaug_log_reg_pred)}\")\n",
    "print(f\"XGBoost F1: {f1_score(y_test_encoded, unaug_xgb_pred)}\")\n",
    "print(f\"CatBoost F1: {f1_score(y_test_encoded, unaug_catboost_pred)}\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "id": "2e6042c0-6b82-4982-ab2b-a38064c8cd41",
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "#defining function to get a confusion matrix heatmap \n",
    "def cf_matrix_plot(y_test, y_pred, axes):\n",
    "    \n",
    "    cf_matrix = confusion_matrix(y_test, y_pred, normalize='true')\n",
    "    cf_heatmap = sns.heatmap(cf_matrix, annot=True, cmap=\"BuPu\", fmt='g', ax=axes)\n",
    "    \n",
    "    return cf_heatmap"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "id": "21e2566b-6cee-43fe-a32e-133690d1d0a4",
   "metadata": {
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1300x400 with 6 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "#subplot of confusion matricies\n",
    "unaug_fig, axes= plt.subplots(1,3, figsize=(13,4))\n",
    "axes = axes.flatten()\n",
    "unaug_fig.suptitle('Modeling Confusion Matrices - Unaugmented Data \\n')\n",
    "\n",
    "cf_matrix_plot(y_test_encoded, unaug_log_reg_pred, axes[0])\n",
    "axes[0].set_title(f\"Logistic Regression - F1: {round(f1_score(y_test_encoded, unaug_log_reg_pred),3)}\")\n",
    "axes[0].set_ylabel('True values')\n",
    "axes[0].set_xlabel('Predicted values')\n",
    "\n",
    "cf_matrix_plot(y_test_encoded, unaug_xgb_pred, axes[1])\n",
    "axes[1].set_title(f\"XGBoost - F1: {round(f1_score(y_test_encoded, unaug_xgb_pred), 3)}\")\n",
    "axes[1].set_ylabel('True values')\n",
    "axes[1].set_xlabel('Predicted values')\n",
    "\n",
    "cf_matrix_plot(y_test_encoded, unaug_catboost_pred, axes[2])\n",
    "axes[2].set_title(f\"CatBoost - F1: {round(f1_score(y_test_encoded, unaug_catboost_pred), 3)}\")\n",
    "axes[2].set_ylabel('True values')\n",
    "axes[2].set_xlabel('Predicted values')\n",
    "\n",
    "unaug_fig.tight_layout()\n",
    "plt.savefig('unaug_confusion_matrices_accomp.jpeg')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "cbb270ee-2f5b-4f8c-949d-bc41fa72a0a3",
   "metadata": {},
   "source": [
    "## Modeling with Tonic Data Science Mode (DSM) - synthesized data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "id": "f1e5e86e-c3a6-438d-83f1-c63fbffd40d7",
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "#install the tonic api \n",
    "#!pip install tonic-api\n",
    "\n",
    "#connecting to DSM\n",
    "from tonic_api.api import TonicApi\n",
    "\n",
    "api_token = \"<APITOKEN\"\n",
    "workspace_id = \"<WORKSPACEID>\"\n",
    "job_id = \"<JOBID>\"\n",
    "\n",
    "tonic = TonicApi(\"https://djinn.tonic.ai\", api_token)\n",
    "\n",
    "#connecting to the workspace and model for this specific dataset\n",
    "workspace = tonic.get_workspace(workspace_id)\n",
    "model = workspace.get_trained_model_by_training_job_id(job_id)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "id": "2e9c4ad6-bdf2-4a7e-85a8-58941e3f6f91",
   "metadata": {
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2431 synthetic data points are needed to balance the dataset 50/50.\n"
     ]
    }
   ],
   "source": [
    "#calculating the number of samples needed to balance the data\n",
    "n_samples = data_train.shape[0]\n",
    "n_churn = data_train[data_train['Churn'] == 'Yes'].shape[0]\n",
    "n_samples_needed = int((0.5 * n_samples - n_churn) / (1 - 0.5)) + 1 \n",
    "\n",
    "print(f'{n_samples_needed} synthetic data points are needed to balance the dataset 50/50.')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "id": "e08e8860-6dc2-4f0e-b18a-563789e6b2bc",
   "metadata": {
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1 iterations complete.\n",
      "2 iterations complete.\n",
      "3 iterations complete.\n",
      "4 iterations complete.\n",
      "5 iterations complete.\n",
      "6 iterations complete.\n",
      "7 iterations complete.\n",
      "8 iterations complete.\n",
      "9 iterations complete.\n",
      "10 iterations complete.\n",
      "11 iterations complete.\n",
      "12 iterations complete.\n",
      "13 iterations complete.\n",
      "14 iterations complete.\n",
      "15 iterations complete.\n",
      "16 iterations complete.\n",
      "17 iterations complete.\n",
      "18 iterations complete.\n",
      "19 iterations complete.\n",
      "20 iterations complete.\n",
      "21 iterations complete.\n",
      "22 iterations complete.\n",
      "23 iterations complete.\n",
      "24 iterations complete.\n",
      "25 iterations complete.\n",
      "26 iterations complete.\n",
      "27 iterations complete.\n",
      "28 iterations complete.\n",
      "29 iterations complete.\n",
      "30 iterations complete.\n",
      "31 iterations complete.\n",
      "32 iterations complete.\n",
      "33 iterations complete.\n",
      "34 iterations complete.\n",
      "35 iterations complete.\n",
      "36 iterations complete.\n",
      "37 iterations complete.\n",
      "38 iterations complete.\n",
      "39 iterations complete.\n",
      "40 iterations complete.\n",
      "41 iterations complete.\n",
      "42 iterations complete.\n",
      "43 iterations complete.\n",
      "44 iterations complete.\n",
      "45 iterations complete.\n",
      "46 iterations complete.\n",
      "47 iterations complete.\n",
      "48 iterations complete.\n",
      "49 iterations complete.\n",
      "50 iterations complete.\n",
      "51 iterations complete.\n",
      "52 iterations complete.\n",
      "53 iterations complete.\n",
      "54 iterations complete.\n",
      "55 iterations complete.\n",
      "56 iterations complete.\n",
      "57 iterations complete.\n",
      "58 iterations complete.\n",
      "59 iterations complete.\n",
      "60 iterations complete.\n",
      "61 iterations complete.\n",
      "62 iterations complete.\n",
      "63 iterations complete.\n",
      "64 iterations complete.\n",
      "65 iterations complete.\n",
      "66 iterations complete.\n",
      "67 iterations complete.\n",
      "68 iterations complete.\n",
      "69 iterations complete.\n",
      "70 iterations complete.\n",
      "71 iterations complete.\n",
      "72 iterations complete.\n",
      "73 iterations complete.\n",
      "74 iterations complete.\n",
      "75 iterations complete.\n",
      "76 iterations complete.\n",
      "77 iterations complete.\n",
      "78 iterations complete.\n",
      "79 iterations complete.\n",
      "80 iterations complete.\n",
      "81 iterations complete.\n",
      "82 iterations complete.\n",
      "83 iterations complete.\n",
      "84 iterations complete.\n",
      "85 iterations complete.\n",
      "86 iterations complete.\n",
      "87 iterations complete.\n",
      "88 iterations complete.\n",
      "89 iterations complete.\n",
      "90 iterations complete.\n",
      "91 iterations complete.\n",
      "92 iterations complete.\n",
      "93 iterations complete.\n",
      "94 iterations complete.\n",
      "95 iterations complete.\n",
      "96 iterations complete.\n",
      "97 iterations complete.\n",
      "98 iterations complete.\n",
      "99 iterations complete.\n",
      "100 iterations complete.\n"
     ]
    }
   ],
   "source": [
    "#iterating through the modeling and scoring process 100x to get a range of results to observe scoring distribution\n",
    "#instantiating lists of resulting scores\n",
    "DSM_log_reg_roc_auc = []\n",
    "DSM_xgb_roc_auc = []\n",
    "DSM_catboost_roc_auc = []\n",
    "\n",
    "DSM_log_reg_f_1 = []\n",
    "DSM_xgb_f_1 = []\n",
    "DSM_catboost_f_1 = []\n",
    "\n",
    "for i in range(100):\n",
    "\n",
    "    #pulling in 100,000 synthetic data points from DSM\n",
    "    sample = model.sample(100000)\n",
    "    churn_samples = sample[sample['Churn'] == 'Yes']\n",
    "\n",
    "    #pulling out the number of minority class samples needed\n",
    "    new_samples = churn_samples[:n_samples_needed]\n",
    "    original_samples = data_train[modeling_columns].copy()\n",
    "    #adding our synthetic data to our original training data\n",
    "    new_model_train_data = pd.concat([original_samples, new_samples], axis=0).reset_index(drop=True)\n",
    "\n",
    "    #encoding augmented data\n",
    "    X_train_aug_encoded = one_hot_encode(new_model_train_data[input_columns], column_to_one_hot_encoder)\n",
    "    #making SeniorCitizen an float since it got skipped over\n",
    "    X_train_aug_encoded['SeniorCitizen']=X_train_aug_encoded['SeniorCitizen'].astype(float)\n",
    "    y_aug_train_encoded = new_model_train_data[target].apply(lambda x: int(x == 'Yes'))\n",
    "\n",
    "    #running models \n",
    "    #DSM\n",
    "    #log reg\n",
    "        #model\n",
    "    DSM_log_reg = log_reg.fit(X_train_aug_encoded, y_aug_train_encoded)\n",
    "        #y proba predictions\n",
    "    DSM_log_reg_pred_proba = DSM_log_reg.predict_proba(X_test_oh_encoded)[:,1]\n",
    "        #y predictions \n",
    "    DSM_log_reg_pred = DSM_log_reg.predict(X_test_oh_encoded)\n",
    "\n",
    "    #xgb\n",
    "        #model\n",
    "    DSM_xgb = xgb_class.fit(X_train_aug_encoded, y_aug_train_encoded)\n",
    "        #y proba predictions\n",
    "    DSM_xgb_pred_proba = DSM_xgb.predict_proba(X_test_oh_encoded)[:,1]\n",
    "        #y predictions \n",
    "    DSM_xgb_pred = DSM_xgb.predict(X_test_oh_encoded)\n",
    "\n",
    "    #catb\n",
    "    X_train_aug = new_model_train_data[input_columns]\n",
    "        #model\n",
    "    DSM_catboost = catboost.fit(X_train_aug, y_aug_train_encoded, cat_features=[col for col in categorical_columns if col!=target])\n",
    "      #y proba predictions\n",
    "    DSM_catboost_pred_proba = unaug_catboost.predict_proba(X_test)[:,1]\n",
    "      #y predictions\n",
    "    DSM_catboost_pred = unaug_catboost.predict(X_test)\n",
    "\n",
    "    #score the models \n",
    "    DSM_log_reg_roc_auc.append(roc_auc_score(y_test_encoded, DSM_log_reg_pred_proba))\n",
    "    DSM_log_reg_f_1.append(roc_auc_score(y_test_encoded, DSM_log_reg_pred))\n",
    "\n",
    "    DSM_xgb_roc_auc.append(roc_auc_score(y_test_encoded, DSM_xgb_pred_proba))\n",
    "    DSM_xgb_f_1.append(roc_auc_score(y_test_encoded, DSM_xgb_pred))\n",
    "\n",
    "    DSM_catboost_roc_auc.append(roc_auc_score(y_test_encoded, DSM_catboost_pred_proba))\n",
    "    DSM_catboost_f_1.append(roc_auc_score(y_test_encoded, DSM_catboost_pred))\n",
    "    \n",
    "    print(f\"{i+1} iterations complete.\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "id": "a6f93229-7317-4690-a2fb-b8d5c530c2f5",
   "metadata": {},
   "outputs": [],
   "source": [
    "#saving all scores as a dataframe\n",
    "DSM_log_reg_roc_auc_df = pd.DataFrame(DSM_log_reg_roc_auc, columns=['log_reg_roc_auc'])\n",
    "DSM_xgb_roc_auc_df = pd.DataFrame(DSM_xgb_roc_auc, columns=['xgb_roc_auc'])\n",
    "DSM_catboost_roc_auc_df = pd.DataFrame(DSM_catboost_roc_auc, columns=['catboost_roc_auc'])\n",
    "\n",
    "DSM_log_reg_f_1_df = pd.DataFrame(DSM_log_reg_f_1, columns=['log_reg_f_1'])\n",
    "DSM_xgb_f_1_df = pd.DataFrame(DSM_xgb_f_1, columns=['xgb_f_1'])\n",
    "DSM_catboost_f_1_df = pd.DataFrame(DSM_catboost_f_1, columns=['catboost_f_1'])\n",
    "\n",
    "DSM_dfs = [DSM_log_reg_roc_auc_df, DSM_xgb_roc_auc_df, DSM_catboost_roc_auc_df, DSM_log_reg_f_1_df, DSM_xgb_f_1_df, DSM_catboost_f_1_df]\n",
    "DSM_scores = pd.concat(DSM_dfs, axis=1)\n",
    "\n",
    "#saving scores as a csv\n",
    "DSM_scores.to_csv('DSM_scores_accomp.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "id": "c9e02e52-a1e2-4519-a522-ab70ca13683f",
   "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>log_reg_roc_auc</th>\n",
       "      <th>xgb_roc_auc</th>\n",
       "      <th>catboost_roc_auc</th>\n",
       "      <th>log_reg_f_1</th>\n",
       "      <th>xgb_f_1</th>\n",
       "      <th>catboost_f_1</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.833973</td>\n",
       "      <td>0.819379</td>\n",
       "      <td>0.845538</td>\n",
       "      <td>0.755699</td>\n",
       "      <td>0.740298</td>\n",
       "      <td>0.759313</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.834613</td>\n",
       "      <td>0.826723</td>\n",
       "      <td>0.843180</td>\n",
       "      <td>0.756715</td>\n",
       "      <td>0.751204</td>\n",
       "      <td>0.763025</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.835787</td>\n",
       "      <td>0.821356</td>\n",
       "      <td>0.845381</td>\n",
       "      <td>0.760580</td>\n",
       "      <td>0.741467</td>\n",
       "      <td>0.751788</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.834200</td>\n",
       "      <td>0.823340</td>\n",
       "      <td>0.845268</td>\n",
       "      <td>0.753895</td>\n",
       "      <td>0.730891</td>\n",
       "      <td>0.761957</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0.834746</td>\n",
       "      <td>0.829999</td>\n",
       "      <td>0.845396</td>\n",
       "      <td>0.756894</td>\n",
       "      <td>0.747874</td>\n",
       "      <td>0.764804</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   log_reg_roc_auc  xgb_roc_auc  catboost_roc_auc  log_reg_f_1   xgb_f_1  \\\n",
       "0         0.833973     0.819379          0.845538     0.755699  0.740298   \n",
       "1         0.834613     0.826723          0.843180     0.756715  0.751204   \n",
       "2         0.835787     0.821356          0.845381     0.760580  0.741467   \n",
       "3         0.834200     0.823340          0.845268     0.753895  0.730891   \n",
       "4         0.834746     0.829999          0.845396     0.756894  0.747874   \n",
       "\n",
       "   catboost_f_1  \n",
       "0      0.759313  \n",
       "1      0.763025  \n",
       "2      0.751788  \n",
       "3      0.761957  \n",
       "4      0.764804  "
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#loading saved DSM scores\n",
    "DSM_scores = pd.read_csv('DSM_scores_accomp.csv', index_col=0)\n",
    "DSM_scores.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "id": "3b9f3f3b-586a-452e-8840-4ce3ff2ee6ba",
   "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>log_reg_roc_auc</th>\n",
       "      <th>xgb_roc_auc</th>\n",
       "      <th>catboost_roc_auc</th>\n",
       "      <th>log_reg_f_1</th>\n",
       "      <th>xgb_f_1</th>\n",
       "      <th>catboost_f_1</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>100.000000</td>\n",
       "      <td>100.000000</td>\n",
       "      <td>100.000000</td>\n",
       "      <td>100.000000</td>\n",
       "      <td>100.000000</td>\n",
       "      <td>100.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>0.834862</td>\n",
       "      <td>0.824836</td>\n",
       "      <td>0.845948</td>\n",
       "      <td>0.758285</td>\n",
       "      <td>0.741009</td>\n",
       "      <td>0.759538</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>0.000936</td>\n",
       "      <td>0.003504</td>\n",
       "      <td>0.001350</td>\n",
       "      <td>0.002726</td>\n",
       "      <td>0.005744</td>\n",
       "      <td>0.004545</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>0.832299</td>\n",
       "      <td>0.815871</td>\n",
       "      <td>0.842024</td>\n",
       "      <td>0.751225</td>\n",
       "      <td>0.725679</td>\n",
       "      <td>0.746551</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>0.834300</td>\n",
       "      <td>0.822435</td>\n",
       "      <td>0.845262</td>\n",
       "      <td>0.756360</td>\n",
       "      <td>0.737209</td>\n",
       "      <td>0.757044</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>0.834816</td>\n",
       "      <td>0.824619</td>\n",
       "      <td>0.845942</td>\n",
       "      <td>0.758051</td>\n",
       "      <td>0.740717</td>\n",
       "      <td>0.759618</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>0.835443</td>\n",
       "      <td>0.827521</td>\n",
       "      <td>0.846903</td>\n",
       "      <td>0.759901</td>\n",
       "      <td>0.744880</td>\n",
       "      <td>0.762999</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>0.837533</td>\n",
       "      <td>0.833289</td>\n",
       "      <td>0.849591</td>\n",
       "      <td>0.764776</td>\n",
       "      <td>0.755576</td>\n",
       "      <td>0.768313</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       log_reg_roc_auc  xgb_roc_auc  catboost_roc_auc  log_reg_f_1  \\\n",
       "count       100.000000   100.000000        100.000000   100.000000   \n",
       "mean          0.834862     0.824836          0.845948     0.758285   \n",
       "std           0.000936     0.003504          0.001350     0.002726   \n",
       "min           0.832299     0.815871          0.842024     0.751225   \n",
       "25%           0.834300     0.822435          0.845262     0.756360   \n",
       "50%           0.834816     0.824619          0.845942     0.758051   \n",
       "75%           0.835443     0.827521          0.846903     0.759901   \n",
       "max           0.837533     0.833289          0.849591     0.764776   \n",
       "\n",
       "          xgb_f_1  catboost_f_1  \n",
       "count  100.000000    100.000000  \n",
       "mean     0.741009      0.759538  \n",
       "std      0.005744      0.004545  \n",
       "min      0.725679      0.746551  \n",
       "25%      0.737209      0.757044  \n",
       "50%      0.740717      0.759618  \n",
       "75%      0.744880      0.762999  \n",
       "max      0.755576      0.768313  "
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#results \n",
    "DSM_scores.describe()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "8391523e-e2d7-4b15-bf7a-c1da48e4e938",
   "metadata": {},
   "source": [
    "## Modeling with SMOTE and SMOTE-NC"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "4576a94b-3e11-421f-a09c-4f3412798eb5",
   "metadata": {},
   "source": [
    "### SMOTE"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "id": "a2ce75e3-295d-4b1c-bf7c-8d15d09aae3f",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1 iterations complete.\n",
      "2 iterations complete.\n",
      "3 iterations complete.\n",
      "4 iterations complete.\n",
      "5 iterations complete.\n",
      "6 iterations complete.\n",
      "7 iterations complete.\n",
      "8 iterations complete.\n",
      "9 iterations complete.\n",
      "10 iterations complete.\n",
      "11 iterations complete.\n",
      "12 iterations complete.\n",
      "13 iterations complete.\n",
      "14 iterations complete.\n",
      "15 iterations complete.\n",
      "16 iterations complete.\n",
      "17 iterations complete.\n",
      "18 iterations complete.\n",
      "19 iterations complete.\n",
      "20 iterations complete.\n",
      "21 iterations complete.\n",
      "22 iterations complete.\n",
      "23 iterations complete.\n",
      "24 iterations complete.\n",
      "25 iterations complete.\n",
      "26 iterations complete.\n",
      "27 iterations complete.\n",
      "28 iterations complete.\n",
      "29 iterations complete.\n",
      "30 iterations complete.\n",
      "31 iterations complete.\n",
      "32 iterations complete.\n",
      "33 iterations complete.\n",
      "34 iterations complete.\n",
      "35 iterations complete.\n",
      "36 iterations complete.\n",
      "37 iterations complete.\n",
      "38 iterations complete.\n",
      "39 iterations complete.\n",
      "40 iterations complete.\n",
      "41 iterations complete.\n",
      "42 iterations complete.\n",
      "43 iterations complete.\n",
      "44 iterations complete.\n",
      "45 iterations complete.\n",
      "46 iterations complete.\n",
      "47 iterations complete.\n",
      "48 iterations complete.\n",
      "49 iterations complete.\n",
      "50 iterations complete.\n",
      "51 iterations complete.\n",
      "52 iterations complete.\n",
      "53 iterations complete.\n",
      "54 iterations complete.\n",
      "55 iterations complete.\n",
      "56 iterations complete.\n",
      "57 iterations complete.\n",
      "58 iterations complete.\n",
      "59 iterations complete.\n",
      "60 iterations complete.\n",
      "61 iterations complete.\n",
      "62 iterations complete.\n",
      "63 iterations complete.\n",
      "64 iterations complete.\n",
      "65 iterations complete.\n",
      "66 iterations complete.\n",
      "67 iterations complete.\n",
      "68 iterations complete.\n",
      "69 iterations complete.\n",
      "70 iterations complete.\n",
      "71 iterations complete.\n",
      "72 iterations complete.\n",
      "73 iterations complete.\n",
      "74 iterations complete.\n",
      "75 iterations complete.\n",
      "76 iterations complete.\n",
      "77 iterations complete.\n",
      "78 iterations complete.\n",
      "79 iterations complete.\n",
      "80 iterations complete.\n",
      "81 iterations complete.\n",
      "82 iterations complete.\n",
      "83 iterations complete.\n",
      "84 iterations complete.\n",
      "85 iterations complete.\n",
      "86 iterations complete.\n",
      "87 iterations complete.\n",
      "88 iterations complete.\n",
      "89 iterations complete.\n",
      "90 iterations complete.\n",
      "91 iterations complete.\n",
      "92 iterations complete.\n",
      "93 iterations complete.\n",
      "94 iterations complete.\n",
      "95 iterations complete.\n",
      "96 iterations complete.\n",
      "97 iterations complete.\n",
      "98 iterations complete.\n",
      "99 iterations complete.\n",
      "100 iterations complete.\n"
     ]
    }
   ],
   "source": [
    "#iterating through modeling and scoring 100x to account for randomness in SMOTE and look at the distribution\n",
    "#instantiating lists of scores \n",
    "SMOTE_log_reg_roc_auc = []\n",
    "SMOTE_xgb_roc_auc = []\n",
    "SMOTE_catboost_roc_auc = []\n",
    "\n",
    "SMOTE_log_reg_f_1 = []\n",
    "SMOTE_xgb_f_1 = []\n",
    "SMOTE_catboost_f_1 = []\n",
    "\n",
    "#iterating\n",
    "for i in range(100):\n",
    "\n",
    "    #augmenting data with SMOTE\n",
    "    X_train_oh_encoded_smoted, y_train_encoded_smoted = SMOTE().fit_resample(X_train_oh_encoded, y_train_encoded)\n",
    "\n",
    "    #running models\n",
    "    #log reg\n",
    "        #model\n",
    "    smote_log_reg = log_reg.fit(X_train_oh_encoded_smoted, y_train_encoded_smoted)\n",
    "        #y proba predictions\n",
    "    smote_log_reg_pred_proba = smote_log_reg.predict_proba(X_test_oh_encoded)[:,1]\n",
    "        #y predictions \n",
    "    smote_log_reg_pred = smote_log_reg.predict(X_test_oh_encoded)\n",
    "\n",
    "    #xgb\n",
    "        #model\n",
    "    smote_xgb = xgb_class.fit(\n",
    "        np.array(X_train_oh_encoded_smoted), \n",
    "        np.array(y_train_encoded_smoted))\n",
    "        #y proba predictions\n",
    "    smote_xgb_pred_proba = smote_xgb.predict_proba(\n",
    "        np.array(X_test_oh_encoded))[:,1]\n",
    "        #y predictions \n",
    "    smote_xgb_pred = smote_xgb.predict(\n",
    "        np.array(X_test_oh_encoded))\n",
    "\n",
    "    #catb\n",
    "        #model\n",
    "    smote_catboost = catboost.fit(X_train_oh_encoded_smoted, y_train_encoded_smoted)\n",
    "        #y proba predictions\n",
    "    smote_catboost_pred_proba = smote_catboost.predict_proba(X_test_oh_encoded)[:,1]\n",
    "        #y predictions \n",
    "    smote_catboost_pred = smote_catboost.predict(X_test_oh_encoded)\n",
    "\n",
    "    #scoring models\n",
    "    SMOTE_log_reg_roc_auc.append(roc_auc_score(y_test_encoded, smote_log_reg_pred_proba))\n",
    "    SMOTE_log_reg_f_1.append(roc_auc_score(y_test_encoded, smote_log_reg_pred))\n",
    "\n",
    "    SMOTE_xgb_roc_auc.append(roc_auc_score(y_test_encoded, smote_xgb_pred_proba))\n",
    "    SMOTE_xgb_f_1.append(roc_auc_score(y_test_encoded, smote_xgb_pred))\n",
    "\n",
    "    SMOTE_catboost_roc_auc.append(roc_auc_score(y_test_encoded, smote_catboost_pred_proba))\n",
    "    SMOTE_catboost_f_1.append(roc_auc_score(y_test_encoded, smote_catboost_pred))\n",
    "\n",
    "    print(f\"{i+1} iterations complete.\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "id": "fab5c89f-413d-4c7c-93ab-ac7a66e7fcb9",
   "metadata": {},
   "outputs": [],
   "source": [
    "#saving all scores as a dataframe\n",
    "SMOTE_log_reg_roc_auc_df = pd.DataFrame(SMOTE_log_reg_roc_auc, columns=['log_reg_roc_auc'])\n",
    "SMOTE_xgb_roc_auc_df = pd.DataFrame(SMOTE_xgb_roc_auc, columns=['xgb_roc_auc'])\n",
    "SMOTE_catboost_roc_auc_df = pd.DataFrame(SMOTE_catboost_roc_auc, columns=['catboost_roc_auc'])\n",
    "\n",
    "SMOTE_log_reg_f_1_df = pd.DataFrame(SMOTE_log_reg_f_1, columns=['log_reg_f_1'])\n",
    "SMOTE_xgb_f_1_df = pd.DataFrame(SMOTE_xgb_f_1, columns=['xgb_f_1'])\n",
    "SMOTE_catboost_f_1_df = pd.DataFrame(SMOTE_catboost_f_1, columns=['catboost_f_1'])\n",
    "\n",
    "SMOTE_dfs = [SMOTE_log_reg_roc_auc_df, SMOTE_xgb_roc_auc_df, SMOTE_catboost_roc_auc_df, SMOTE_log_reg_f_1_df, SMOTE_xgb_f_1_df, SMOTE_catboost_f_1_df]\n",
    "SMOTE_scores = pd.concat(SMOTE_dfs, axis=1)\n",
    "\n",
    "#saving scores as a CSV\n",
    "SMOTE_scores.to_csv('SMOTE_scores_accomp.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "id": "b9ceb8b2-63f7-4c65-b42b-f63642ff9065",
   "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>log_reg_roc_auc</th>\n",
       "      <th>xgb_roc_auc</th>\n",
       "      <th>catboost_roc_auc</th>\n",
       "      <th>log_reg_f_1</th>\n",
       "      <th>xgb_f_1</th>\n",
       "      <th>catboost_f_1</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.834644</td>\n",
       "      <td>0.818818</td>\n",
       "      <td>0.832924</td>\n",
       "      <td>0.758192</td>\n",
       "      <td>0.705630</td>\n",
       "      <td>0.713943</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.832635</td>\n",
       "      <td>0.819938</td>\n",
       "      <td>0.831937</td>\n",
       "      <td>0.755981</td>\n",
       "      <td>0.696681</td>\n",
       "      <td>0.703088</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.835041</td>\n",
       "      <td>0.818962</td>\n",
       "      <td>0.834047</td>\n",
       "      <td>0.759362</td>\n",
       "      <td>0.701537</td>\n",
       "      <td>0.713943</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.836977</td>\n",
       "      <td>0.823315</td>\n",
       "      <td>0.832789</td>\n",
       "      <td>0.768692</td>\n",
       "      <td>0.704486</td>\n",
       "      <td>0.713969</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0.833158</td>\n",
       "      <td>0.820852</td>\n",
       "      <td>0.833642</td>\n",
       "      <td>0.759692</td>\n",
       "      <td>0.701460</td>\n",
       "      <td>0.710181</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   log_reg_roc_auc  xgb_roc_auc  catboost_roc_auc  log_reg_f_1   xgb_f_1  \\\n",
       "0         0.834644     0.818818          0.832924     0.758192  0.705630   \n",
       "1         0.832635     0.819938          0.831937     0.755981  0.696681   \n",
       "2         0.835041     0.818962          0.834047     0.759362  0.701537   \n",
       "3         0.836977     0.823315          0.832789     0.768692  0.704486   \n",
       "4         0.833158     0.820852          0.833642     0.759692  0.701460   \n",
       "\n",
       "   catboost_f_1  \n",
       "0      0.713943  \n",
       "1      0.703088  \n",
       "2      0.713943  \n",
       "3      0.713969  \n",
       "4      0.710181  "
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#loading saved SMOTE scores\n",
    "SMOTE_scores = pd.read_csv('SMOTE_scores_accomp.csv', index_col=0)\n",
    "SMOTE_scores.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "id": "791b56ff-b1e8-4c78-972a-ef48a3bd07bb",
   "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>log_reg_roc_auc</th>\n",
       "      <th>xgb_roc_auc</th>\n",
       "      <th>catboost_roc_auc</th>\n",
       "      <th>log_reg_f_1</th>\n",
       "      <th>xgb_f_1</th>\n",
       "      <th>catboost_f_1</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>100.000000</td>\n",
       "      <td>100.000000</td>\n",
       "      <td>100.000000</td>\n",
       "      <td>100.000000</td>\n",
       "      <td>100.000000</td>\n",
       "      <td>100.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>0.835332</td>\n",
       "      <td>0.821945</td>\n",
       "      <td>0.832473</td>\n",
       "      <td>0.761509</td>\n",
       "      <td>0.706448</td>\n",
       "      <td>0.709815</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>0.000975</td>\n",
       "      <td>0.002728</td>\n",
       "      <td>0.001068</td>\n",
       "      <td>0.003639</td>\n",
       "      <td>0.005512</td>\n",
       "      <td>0.004676</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>0.832635</td>\n",
       "      <td>0.816550</td>\n",
       "      <td>0.829499</td>\n",
       "      <td>0.753260</td>\n",
       "      <td>0.691444</td>\n",
       "      <td>0.699402</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>0.834710</td>\n",
       "      <td>0.819875</td>\n",
       "      <td>0.831836</td>\n",
       "      <td>0.759260</td>\n",
       "      <td>0.702649</td>\n",
       "      <td>0.706463</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>0.835322</td>\n",
       "      <td>0.821736</td>\n",
       "      <td>0.832540</td>\n",
       "      <td>0.761509</td>\n",
       "      <td>0.706583</td>\n",
       "      <td>0.709456</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>0.836035</td>\n",
       "      <td>0.823700</td>\n",
       "      <td>0.833103</td>\n",
       "      <td>0.764516</td>\n",
       "      <td>0.710072</td>\n",
       "      <td>0.713549</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>0.837265</td>\n",
       "      <td>0.829109</td>\n",
       "      <td>0.835059</td>\n",
       "      <td>0.768692</td>\n",
       "      <td>0.725841</td>\n",
       "      <td>0.719815</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       log_reg_roc_auc  xgb_roc_auc  catboost_roc_auc  log_reg_f_1  \\\n",
       "count       100.000000   100.000000        100.000000   100.000000   \n",
       "mean          0.835332     0.821945          0.832473     0.761509   \n",
       "std           0.000975     0.002728          0.001068     0.003639   \n",
       "min           0.832635     0.816550          0.829499     0.753260   \n",
       "25%           0.834710     0.819875          0.831836     0.759260   \n",
       "50%           0.835322     0.821736          0.832540     0.761509   \n",
       "75%           0.836035     0.823700          0.833103     0.764516   \n",
       "max           0.837265     0.829109          0.835059     0.768692   \n",
       "\n",
       "          xgb_f_1  catboost_f_1  \n",
       "count  100.000000    100.000000  \n",
       "mean     0.706448      0.709815  \n",
       "std      0.005512      0.004676  \n",
       "min      0.691444      0.699402  \n",
       "25%      0.702649      0.706463  \n",
       "50%      0.706583      0.709456  \n",
       "75%      0.710072      0.713549  \n",
       "max      0.725841      0.719815  "
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#results\n",
    "SMOTE_scores.describe()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "7c9fbebb-e06e-4c0c-b139-01d563e83390",
   "metadata": {},
   "source": [
    "### SMOTE-NC"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "id": "b3e71c13-aeb8-4a99-a316-448c2e5c06a3",
   "metadata": {},
   "outputs": [],
   "source": [
    "#subsetting categorical columns\n",
    "categorical_columns = list(column_to_one_hot_encoder.keys())\n",
    "categorical_features = []\n",
    "for col in X_train.columns:\n",
    "    if col in categorical_columns:\n",
    "        categorical_features.append(True)\n",
    "    else:\n",
    "        categorical_features.append(False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "id": "fb370b22-54dc-4a89-aaee-86b04f23c7ba",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1 iterations complete.\n",
      "2 iterations complete.\n",
      "3 iterations complete.\n",
      "4 iterations complete.\n",
      "5 iterations complete.\n",
      "6 iterations complete.\n",
      "7 iterations complete.\n",
      "8 iterations complete.\n",
      "9 iterations complete.\n",
      "10 iterations complete.\n",
      "11 iterations complete.\n",
      "12 iterations complete.\n",
      "13 iterations complete.\n",
      "14 iterations complete.\n",
      "15 iterations complete.\n",
      "16 iterations complete.\n",
      "17 iterations complete.\n",
      "18 iterations complete.\n",
      "19 iterations complete.\n",
      "20 iterations complete.\n",
      "21 iterations complete.\n",
      "22 iterations complete.\n",
      "23 iterations complete.\n",
      "24 iterations complete.\n",
      "25 iterations complete.\n",
      "26 iterations complete.\n",
      "27 iterations complete.\n",
      "28 iterations complete.\n",
      "29 iterations complete.\n",
      "30 iterations complete.\n",
      "31 iterations complete.\n",
      "32 iterations complete.\n",
      "33 iterations complete.\n",
      "34 iterations complete.\n",
      "35 iterations complete.\n",
      "36 iterations complete.\n",
      "37 iterations complete.\n",
      "38 iterations complete.\n",
      "39 iterations complete.\n",
      "40 iterations complete.\n",
      "41 iterations complete.\n",
      "42 iterations complete.\n",
      "43 iterations complete.\n",
      "44 iterations complete.\n",
      "45 iterations complete.\n",
      "46 iterations complete.\n",
      "47 iterations complete.\n",
      "48 iterations complete.\n",
      "49 iterations complete.\n",
      "50 iterations complete.\n",
      "51 iterations complete.\n",
      "52 iterations complete.\n",
      "53 iterations complete.\n",
      "54 iterations complete.\n",
      "55 iterations complete.\n",
      "56 iterations complete.\n",
      "57 iterations complete.\n",
      "58 iterations complete.\n",
      "59 iterations complete.\n",
      "60 iterations complete.\n",
      "61 iterations complete.\n",
      "62 iterations complete.\n",
      "63 iterations complete.\n",
      "64 iterations complete.\n",
      "65 iterations complete.\n",
      "66 iterations complete.\n",
      "67 iterations complete.\n",
      "68 iterations complete.\n",
      "69 iterations complete.\n",
      "70 iterations complete.\n",
      "71 iterations complete.\n",
      "72 iterations complete.\n",
      "73 iterations complete.\n",
      "74 iterations complete.\n",
      "75 iterations complete.\n",
      "76 iterations complete.\n",
      "77 iterations complete.\n",
      "78 iterations complete.\n",
      "79 iterations complete.\n",
      "80 iterations complete.\n",
      "81 iterations complete.\n",
      "82 iterations complete.\n",
      "83 iterations complete.\n",
      "84 iterations complete.\n",
      "85 iterations complete.\n",
      "86 iterations complete.\n",
      "87 iterations complete.\n",
      "88 iterations complete.\n",
      "89 iterations complete.\n",
      "90 iterations complete.\n",
      "91 iterations complete.\n",
      "92 iterations complete.\n",
      "93 iterations complete.\n",
      "94 iterations complete.\n",
      "95 iterations complete.\n",
      "96 iterations complete.\n",
      "97 iterations complete.\n",
      "98 iterations complete.\n",
      "99 iterations complete.\n",
      "100 iterations complete.\n"
     ]
    }
   ],
   "source": [
    "#iterating through SMOTE-NC resampling 100x to account for randomness and evaluate on score distributions\n",
    "#instantiating lists of resulting scores\n",
    "SMOTE_NC_log_reg_roc_auc = []\n",
    "SMOTE_NC_xgb_roc_auc = []\n",
    "SMOTE_NC_catboost_roc_auc = []\n",
    "\n",
    "SMOTE_NC_log_reg_f_1 = []\n",
    "SMOTE_NC_xgb_f_1 = []\n",
    "SMOTE_NC_catboost_f_1 = []\n",
    "\n",
    "for i in range(100):\n",
    "    \n",
    "    #augmenting data with SMOTE-NC\n",
    "    X_train_smotenced, y_train_smotenced = SMOTENC(\n",
    "        categorical_features=np.array(categorical_features)\n",
    "    ).fit_resample(X_train, y_train)\n",
    "\n",
    "    #one hot encoding smotenced data\n",
    "    X_train_oh_encoded_smotenced = one_hot_encode(X_train_smotenced, column_to_one_hot_encoder)\n",
    "    y_train_encoded_smotenced = y_train_smotenced.apply(lambda x : int(x == 'Yes'))\n",
    "\n",
    "    #Modeling smote-nced data\n",
    "    #log reg\n",
    "        #model\n",
    "    smotenc_log_reg = log_reg.fit(X_train_oh_encoded_smotenced, y_train_encoded_smotenced)\n",
    "        #y proba predictions\n",
    "    smotenc_log_reg_pred_proba = smotenc_log_reg.predict_proba(X_test_oh_encoded)[:,1]\n",
    "        #y predictions\n",
    "    smotenc_log_reg_pred = smotenc_log_reg.predict(X_test_oh_encoded)\n",
    "\n",
    "    #xgb\n",
    "        #model\n",
    "    smotenc_xgb = xgb_class.fit(\n",
    "        np.array(X_train_oh_encoded_smotenced), \n",
    "        np.array(y_train_encoded_smotenced))\n",
    "        #y proba predictions\n",
    "    smotenc_xgb_pred_proba = smotenc_xgb.predict_proba(\n",
    "        np.array(X_test_oh_encoded))[:,1]\n",
    "        #y predictions \n",
    "    smotenc_xgb_pred = smotenc_xgb.predict(\n",
    "        np.array(X_test_oh_encoded))\n",
    "    #catb\n",
    "        #model\n",
    "    smotenc_catboost = catboost.fit(X_train_smotenced, y_train_encoded_smotenced, cat_features=[col for col in categorical_columns if col!=target])\n",
    "        #y proba predictions\n",
    "    smotenc_catboost_pred_proba = smotenc_catboost.predict_proba(X_test)[:,1]\n",
    "        #y predictions \n",
    "    smotenc_catboost_pred = smotenc_catboost.predict(X_test)\n",
    "\n",
    "    #scoring models\n",
    "    SMOTE_NC_log_reg_roc_auc.append(roc_auc_score(y_test_encoded, smotenc_log_reg_pred_proba))\n",
    "    SMOTE_NC_log_reg_f_1.append(roc_auc_score(y_test_encoded, smotenc_log_reg_pred))\n",
    "\n",
    "    SMOTE_NC_xgb_roc_auc.append(roc_auc_score(y_test_encoded, smotenc_xgb_pred_proba))\n",
    "    SMOTE_NC_xgb_f_1.append(roc_auc_score(y_test_encoded, smotenc_xgb_pred))\n",
    "\n",
    "    SMOTE_NC_catboost_roc_auc.append(roc_auc_score(y_test_encoded, smotenc_catboost_pred_proba))\n",
    "    SMOTE_NC_catboost_f_1.append(roc_auc_score(y_test_encoded, smotenc_catboost_pred))\n",
    "\n",
    "    print(f\"{i+1} iterations complete.\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "id": "bcef1f9e-22a6-483a-901d-44ec540641a3",
   "metadata": {},
   "outputs": [],
   "source": [
    "#saving all scores as a dataframe\n",
    "SMOTE_NC_log_reg_roc_auc_df = pd.DataFrame(SMOTE_NC_log_reg_roc_auc, columns=['log_reg_roc_auc'])\n",
    "SMOTE_NC_xgb_roc_auc_df = pd.DataFrame(SMOTE_NC_xgb_roc_auc, columns=['xgb_roc_auc'])\n",
    "SMOTE_NC_catboost_roc_auc_df = pd.DataFrame(SMOTE_NC_catboost_roc_auc, columns=['catboost_roc_auc'])\n",
    "\n",
    "SMOTE_NC_log_reg_f_1_df = pd.DataFrame(SMOTE_NC_log_reg_f_1, columns=['log_reg_f_1'])\n",
    "SMOTE_NC_xgb_f_1_df = pd.DataFrame(SMOTE_NC_xgb_f_1, columns=['xgb_f_1'])\n",
    "SMOTE_NC_catboost_f_1_df = pd.DataFrame(SMOTE_NC_catboost_f_1, columns=['catboost_f_1'])\n",
    "\n",
    "SMOTE_NC_dfs = [SMOTE_NC_log_reg_roc_auc_df, SMOTE_NC_xgb_roc_auc_df, SMOTE_NC_catboost_roc_auc_df, SMOTE_NC_log_reg_f_1_df, SMOTE_NC_xgb_f_1_df, SMOTE_NC_catboost_f_1_df]\n",
    "SMOTE_NC_scores = pd.concat(SMOTE_NC_dfs, axis=1)\n",
    "\n",
    "#saving scores as a CSV\n",
    "SMOTE_NC_scores.to_csv('SMOTE_NC_scores_accomp.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "id": "5cf0f52e-39db-4d6a-a93b-e16f2767e789",
   "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>log_reg_roc_auc</th>\n",
       "      <th>xgb_roc_auc</th>\n",
       "      <th>catboost_roc_auc</th>\n",
       "      <th>log_reg_f_1</th>\n",
       "      <th>xgb_f_1</th>\n",
       "      <th>catboost_f_1</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.827874</td>\n",
       "      <td>0.813448</td>\n",
       "      <td>0.826228</td>\n",
       "      <td>0.757230</td>\n",
       "      <td>0.705855</td>\n",
       "      <td>0.718717</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.826619</td>\n",
       "      <td>0.815603</td>\n",
       "      <td>0.827801</td>\n",
       "      <td>0.756086</td>\n",
       "      <td>0.723422</td>\n",
       "      <td>0.728124</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.826953</td>\n",
       "      <td>0.811313</td>\n",
       "      <td>0.828496</td>\n",
       "      <td>0.755018</td>\n",
       "      <td>0.708829</td>\n",
       "      <td>0.721743</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.826386</td>\n",
       "      <td>0.819962</td>\n",
       "      <td>0.826794</td>\n",
       "      <td>0.750695</td>\n",
       "      <td>0.716710</td>\n",
       "      <td>0.723547</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0.826593</td>\n",
       "      <td>0.813284</td>\n",
       "      <td>0.826050</td>\n",
       "      <td>0.750086</td>\n",
       "      <td>0.719303</td>\n",
       "      <td>0.723701</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   log_reg_roc_auc  xgb_roc_auc  catboost_roc_auc  log_reg_f_1   xgb_f_1  \\\n",
       "0         0.827874     0.813448          0.826228     0.757230  0.705855   \n",
       "1         0.826619     0.815603          0.827801     0.756086  0.723422   \n",
       "2         0.826953     0.811313          0.828496     0.755018  0.708829   \n",
       "3         0.826386     0.819962          0.826794     0.750695  0.716710   \n",
       "4         0.826593     0.813284          0.826050     0.750086  0.719303   \n",
       "\n",
       "   catboost_f_1  \n",
       "0      0.718717  \n",
       "1      0.728124  \n",
       "2      0.721743  \n",
       "3      0.723547  \n",
       "4      0.723701  "
      ]
     },
     "execution_count": 32,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#loading SMOTE scores dataset\n",
    "SMOTE_NC_scores = pd.read_csv('SMOTE_NC_scores_accomp.csv', index_col=0)\n",
    "SMOTE_NC_scores.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "id": "7095bb9c-f05a-4e9f-8094-98817d820879",
   "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>log_reg_roc_auc</th>\n",
       "      <th>xgb_roc_auc</th>\n",
       "      <th>catboost_roc_auc</th>\n",
       "      <th>log_reg_f_1</th>\n",
       "      <th>xgb_f_1</th>\n",
       "      <th>catboost_f_1</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>100.000000</td>\n",
       "      <td>100.000000</td>\n",
       "      <td>100.000000</td>\n",
       "      <td>100.000000</td>\n",
       "      <td>100.000000</td>\n",
       "      <td>100.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>0.827298</td>\n",
       "      <td>0.815250</td>\n",
       "      <td>0.827261</td>\n",
       "      <td>0.752273</td>\n",
       "      <td>0.714635</td>\n",
       "      <td>0.724852</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>0.001142</td>\n",
       "      <td>0.002832</td>\n",
       "      <td>0.001238</td>\n",
       "      <td>0.002406</td>\n",
       "      <td>0.006124</td>\n",
       "      <td>0.004284</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>0.824557</td>\n",
       "      <td>0.807556</td>\n",
       "      <td>0.823600</td>\n",
       "      <td>0.747009</td>\n",
       "      <td>0.700643</td>\n",
       "      <td>0.714752</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>0.826612</td>\n",
       "      <td>0.813597</td>\n",
       "      <td>0.826441</td>\n",
       "      <td>0.750651</td>\n",
       "      <td>0.709999</td>\n",
       "      <td>0.721736</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>0.827311</td>\n",
       "      <td>0.815433</td>\n",
       "      <td>0.827362</td>\n",
       "      <td>0.752310</td>\n",
       "      <td>0.714587</td>\n",
       "      <td>0.724704</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>0.827921</td>\n",
       "      <td>0.817247</td>\n",
       "      <td>0.827931</td>\n",
       "      <td>0.754108</td>\n",
       "      <td>0.718922</td>\n",
       "      <td>0.727368</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>0.829905</td>\n",
       "      <td>0.820515</td>\n",
       "      <td>0.829977</td>\n",
       "      <td>0.758195</td>\n",
       "      <td>0.731659</td>\n",
       "      <td>0.735598</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       log_reg_roc_auc  xgb_roc_auc  catboost_roc_auc  log_reg_f_1  \\\n",
       "count       100.000000   100.000000        100.000000   100.000000   \n",
       "mean          0.827298     0.815250          0.827261     0.752273   \n",
       "std           0.001142     0.002832          0.001238     0.002406   \n",
       "min           0.824557     0.807556          0.823600     0.747009   \n",
       "25%           0.826612     0.813597          0.826441     0.750651   \n",
       "50%           0.827311     0.815433          0.827362     0.752310   \n",
       "75%           0.827921     0.817247          0.827931     0.754108   \n",
       "max           0.829905     0.820515          0.829977     0.758195   \n",
       "\n",
       "          xgb_f_1  catboost_f_1  \n",
       "count  100.000000    100.000000  \n",
       "mean     0.714635      0.724852  \n",
       "std      0.006124      0.004284  \n",
       "min      0.700643      0.714752  \n",
       "25%      0.709999      0.721736  \n",
       "50%      0.714587      0.724704  \n",
       "75%      0.718922      0.727368  \n",
       "max      0.731659      0.735598  "
      ]
     },
     "execution_count": 33,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#results\n",
    "SMOTE_NC_scores.describe()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "746675df-ceea-46bc-a04c-5a90a6b8624f",
   "metadata": {},
   "source": [
    "## Comparing scores "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "632b275c-5b35-431b-b609-ef524929763c",
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "DSM_scores = pd.read_csv('djinn_scores_accomp.csv')\n",
    "SMOTE_scores = pd.read_csv('SMOTE_scores_accomp.csv')\n",
    "SMOTE_NC_scores = pd.read_csv('SMOTE_NC_scores_accomp.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "id": "fcc7b5df-3cbe-4da2-a9cf-35f34241ab50",
   "metadata": {
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1000x700 with 3 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "#distributions of ROC AUC scores for each model by augmentation method\n",
    "roc_auc_fig, (ax1, ax2, ax3) = plt.subplots(1,3, figsize=(10,7), sharey=True)\n",
    "roc_auc_fig.suptitle('Distribution of ROC AUC scores')\n",
    "\n",
    "log_reg_box = [DSM_scores['log_reg_roc_auc'], SMOTE_scores['log_reg_roc_auc'], SMOTE_NC_scores['log_reg_roc_auc']]\n",
    "xgb_box = [DSM_scores['xgb_roc_auc'], SMOTE_scores['xgb_roc_auc'], SMOTE_NC_scores['xgb_roc_auc']]\n",
    "catboost_box= [DSM_scores['catboost_roc_auc'], SMOTE_scores['catboost_roc_auc'], SMOTE_NC_scores['catboost_roc_auc']]\n",
    "\n",
    "bplot1 = ax1.boxplot(log_reg_box, patch_artist=True, notch=True, widths=1, medianprops={'color':'black'})\n",
    "#changing color\n",
    "for patch in bplot1['boxes']:\n",
    "    patch.set_facecolor('mediumturquoise')\n",
    "#other graph settings\n",
    "ax1.set_xticks([1,2,3], ['DSM', 'SMOTE', 'SMOTE_NC'])\n",
    "ax1.set_title('Logistic Regression ROC AUC Scores')\n",
    "ax1.grid(axis='y')\n",
    "ax1.axhline(roc_auc_score(y_test_encoded, unaug_log_reg_pred_proba), color='mediumpurple', label='Unaugmented score')\n",
    "\n",
    "bplot2 = ax2.boxplot(xgb_box, patch_artist=True, notch=True, widths=1, medianprops={'color':'black'})\n",
    "#changing color\n",
    "for patch in bplot2['boxes']:\n",
    "    patch.set_facecolor('mediumturquoise')\n",
    "#other graph settings\n",
    "ax2.set_xticks([1,2,3], ['DSM', 'SMOTE', 'SMOTE_NC'])\n",
    "ax2.set_title('XGBoost ROC AUC Scores')\n",
    "ax2.grid(axis='y')\n",
    "ax2.axhline(roc_auc_score(y_test_encoded, unaug_xgb_pred_proba), color='mediumpurple')\n",
    "\n",
    "bplot3 = ax3.boxplot(catboost_box, patch_artist=True, notch=True, widths=1, medianprops={'color':'black'})\n",
    "#changing color\n",
    "for patch in bplot3['boxes']:\n",
    "    patch.set_facecolor('mediumturquoise')\n",
    "#other graph settings\n",
    "ax3.set_xticks([1,2,3], ['DSM', 'SMOTE', 'SMOTE_NC'])\n",
    "ax3.set_title('Catboost ROC AUC Scores')\n",
    "ax3.grid(axis='y')\n",
    "ax3.axhline(roc_auc_score(y_test_encoded, unaug_catboost_pred_proba), color='mediumpurple')\n",
    "\n",
    "roc_auc_fig.legend()\n",
    "\n",
    "plt.savefig('roc_auc_box_accomp.jpeg')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "id": "ccfed73b-252a-46e2-a61c-6f51288b3cbe",
   "metadata": {
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAA6AAAAKTCAYAAAAHXGUGAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjcuMSwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/bCgiHAAAACXBIWXMAAA9hAAAPYQGoP6dpAACLW0lEQVR4nOzdeZyNdf/H8ffMmX1lzM6YGcLYl1ESopIRipBUblFJIkl3i7qzlNIq3W2qu/BLi0ibLUVERIYSmbFkkFmYYWbMYtbr94fm5JjBDOM6Z3g9H4/zmDnXdZ3v+VxnznWdeZ/re30vJ8MwDAEAAAAAcIE527sAAAAAAMClgQAKAAAAADAFARQAAAAAYAoCKAAAAADAFC72LgAAAAC42JWUlKioqMjeZQAXnKurqywWy2nnE0ABAACAC8QwDKWmpiozM9PepQCmqVWrlkJDQ+Xk5FRuHgEUAAAAuEDKwmdwcLC8vLwq/IccuFgYhqG8vDwdOnRIkhQWFlZuGQIoAAAAcAGUlJRYw2edOnXsXQ5gCk9PT0nSoUOHFBwcXK47LoMQAQAAABdA2TmfXl5edq4EMFfZe76i854JoAAAAMAFRLdbXGrO9J4ngAIAAAAATEEABQAAAAAHERUVpRkzZti7jAuGAAoAAADARrdu3TRu3Lhy02fPnq1atWqZXo+ju9hDY3UigAIAAAAAyikpKVFpaWm1tkkABQAAAHBOhg0bpn79+unll19WWFiY6tSpo9GjR9uMfvrhhx+qffv28vX1VWhoqG6//XbrdSKlio+qfvnll+UGspk6daqCg4Pl6+ure+65R48//rjatGlTrpbnnntOISEhqlWrlp5++mkVFxfrkUceUUBAgOrVq6dZs2bZtHvgwAENGjRItWrVUkBAgPr27aukpKRKr2O3bt20b98+PfTQQ3JycrKpe+3aterSpYs8PT0VERGhsWPHKjc31zr/0KFDuvHGG+Xp6ano6Gh99NFHZ33NV61apSuuuELe3t6qVauWOnXqpH379lnnf/PNN7r88svl4eGhwMBA3XzzzdZ5R48e1dChQ1W7dm15eXnphhtu0K5du8r9Lb7++ms1a9ZM7u7u2r9/vwoKCvTvf/9bdevWlbe3tzp06KBVq1adtdaKEEABAAAAkxiGoaKCUrvcDMO4IOv0ww8/aM+ePfrhhx80Z84czZ49W7Nnz7bOLyoq0jPPPKPffvtNX375pZKSkjRs2LAqPcdHH32kZ599Vi+88ILi4+NVv359vf322+WWW7lypZKTk/Xjjz9q+vTpmjRpkvr06aPatWtrw4YNuu+++zRy5Ej99ddf1tri4uLk6+urNWvW6KeffpKPj4969uypwsLCSq3jwoULVa9ePT399NNKSUlRSkqKJGnPnj3q2bOnBgwYoK1bt2revHlau3atxowZY2132LBhOnDggH744QctWLBAb731lk04P1VxcbH69eunrl27auvWrVq/fr3uvfdea+hdvHixbr75ZvXq1UtbtmzRihUrdMUVV9g836ZNm/T1119r/fr1MgxDvXr1svnCIC8vTy+88IL+97//afv27QoODtaYMWO0fv16ffrpp9q6datuueUW9ezZ0ya8VpaTcaHeiQBwiZg8ebKmTJlywT7YT9atWzdJsn7ruGrVKl1zzTWaP3++Bg4ceMGff9iwYVq1apXNN8OOJicnR+PGjdOiRYuUlpamBx98kPNyANjF8ePHtXfvXkVHR8vDw0OSVFRQqnce/9Mu9Yx8voFc3St3/Klbt25q06ZNuf3n7NmzNW7cOGVmZkr653Nhz549slgskqRBgwbJ2dlZn376aYVtb9q0SZdffrmOHTsmHx+fcm1KJ46A3nzzzdbP1iuvvFLt27fXG2+8YV2mc+fOysnJ0a+//mpTy59//iln5xPrGRMTo+DgYP3444+STnQp9ff31//+9z8NHjxYc+fO1dSpU7Vjxw5riCssLFStWrX05ZdfqkePHpVax6ioKI0bN87mvNl77rlHFotF77zzjnXa2rVr1bVrV+Xm5mr//v1q0qSJNm7cqMsvv1ySlJCQoKZNm+rVV1+t8BzcI0eOqE6dOlq1apW6du1abv5VV12lBg0aaO7cueXm7dq1S40bN9ZPP/2kq666SpKUkZGhiIgIzZkzR7fccotmz56t4cOH69dff1Xr1q0lSfv371eDBg20f/9+hYeHW9vr3r27rrjiCj333HPlnqui934ZjoACwElmz55t7T7j5OQkDw8PhYeHKy4uTv/973917Nixanme5ORkTZ482fqh6UgcubbKeO655zR79myNGjVKH374of71r3+ddtmoqCibv/fJt+PHj0s6EWgnTZqknj17KiAgQE5OTjbf7APApa558+bWYCZJYWFhNkfx4uPjdeONN6p+/fry9fW1Bqf9+/dX+jkSExNtjuRJKne/rJay8ClJISEhatmypfW+xWJRnTp1rPX99ttv2r17t3x9feXj4yMfHx8FBATo+PHj2rNnT6XXsSK//fabZs+ebW3Xx8dHcXFxKi0t1d69e7Vjxw65uLgoNjbW+piYmJgzDvIUEBCgYcOGKS4uTjfeeKNee+016xFXSfr111913XXXVfjYsufr0KGDdVqdOnXUpEkT7dixwzrNzc1NrVq1st7//fffVVJSosaNG9usy+rVq21eo8pyqfIjAOAS8PTTTys6OlpFRUVKTU3VqlWrNG7cOE2fPl1ff/21zY75P//5jx5//PEqtZ+cnKwpU6YoKirK5vyVs1m+fHmVnudcnKm29957r9oHI6huK1eu1JVXXqlJkyZVavk2bdro4YcfLjfdzc1NkpSenq6nn35a9evXV+vWrc/5nBcAkCQXNyeNfL6B3Z67svz8/JSVlVVuemZmpvz9/W2mubq62tx3cnKyflbk5uYqLi5OcXFx+uijjxQUFKT9+/crLi7O2sXV2dm5XC+ik7uEVkVFtZypvpycHMXGxlZ47mVQUFCl1vF0cnJyNHLkSI0dO7bcvPr162vnzp1nXpnTmDVrlsaOHatly5Zp3rx5+s9//qPvvvtOV155pTw9Pc+pzZN5enranMeak5Mji8Wi+Ph4mxAuST4+PlVunwAKABW44YYb1L59e+v9CRMmaOXKlerTp49uuukm7dixw7qTd3FxkYvLhd2d5uXlycvLyxqK7OXUD2BHdOjQITVr1qzSy9etW1dDhgw57fywsDClpKQoNDTU2m2spsnNzZW3t7e9ywCgvwORe+WDoL00adKkwi89N2/erMaNG1e6nYSEBGVkZOj5559XRESEpBNdcE8WFBSkY8eO2eyrTu2F06RJE/3yyy8aOnSoddovv/xS6TpOp127dpo3b56Cg4Pl5+d3zu24ubmppKSkXNt//PGHLrvssgofExMTo+LiYsXHx1s/WxITE226Ip9O27Zt1bZtW02YMEEdO3bUxx9/rCuvvFKtWrXSihUrNHz48HKPadq0qYqLi7VhwwabLriJiYln/Nxs27atSkpKdOjQIXXp0uWstZ0NXXABoJKuvfZaPfXUU9q3b5/NuRWTJ08uN1Lfd999p86dO6tWrVry8fFRkyZN9MQTT0g6cd5m2QfN8OHDrV0+y7p1duvWTS1atFB8fLyuvvpqeXl5WR/brVs363mgJyspKdETTzyh0NBQeXt766abbtKBAwdslomKiqpw0IeT2zxbbcOGDVNUVJTN43Nzc/Xwww8rIiJC7u7uatKkiV5++eVy32Y7OTlpzJgx+vLLL9WiRQu5u7urefPmWrZsWcUv+CkOHTqku+++WyEhIfLw8FDr1q01Z84c6/xVq1bJyclJe/fu1eLFi621n+/5qu7u7goNDT3nx6empmr48OGqV6+e3N3dFRYWVm6ERUlaunSpunbtKl9fX/n5+enyyy/Xxx9/bLPM/PnzFRsbK09PTwUGBmrIkCE6ePCgzTLDhg2Tj4+P9uzZo169esnX11d33HGHJKm0tFQzZsxQ8+bN5eHhoZCQEI0cOVJHjx61aWPTpk2Ki4tTYGCgdWTGu+6665xfAwA1z6hRo7Rz506NHTtWW7duVWJioqZPn65PPvmkwl4jp1O/fn25ubnp9ddf159//qmvv/5azzzzjM0yHTp0sH7W7dmzRx9//HG5Ux0eeOABvf/++5ozZ4527dqlqVOnauvWreU+f6vqjjvuUGBgoPr27as1a9Zo7969WrVqlcaOHWsdqKgyoqKi9OOPP+rgwYNKT0+XJD322GNat26dxowZo19//VW7du3SV199ZR2EqEmTJurZs6dGjhypDRs2KD4+Xvfcc88Zj2Lu3btXEyZM0Pr167Vv3z4tX75cu3btUtOmTSVJkyZN0ieffKJJkyZpx44d+v333/XCCy9Ikho1aqS+fftqxIgRWrt2rX777TcNGTJEdevWVd++fU/7nI0bN9Ydd9yhoUOHauHChdq7d682btyoadOmafHixZV+jcoQQAGgCsrOJzxTV9jt27erT58+Kigo0NNPP61XXnlFN910k3766SdJJ76BfPrppyVJ9957rz788EN9+OGHuvrqq61tZGRk6IYbbrAOAHHNNdecsa5nn31Wixcv1mOPPaaxY8fqu+++U/fu3ZWfn1+l9atMbSczDEM33XSTXn31VfXs2VPTp09XkyZN9Mgjj2j8+PHlll+7dq3uv/9+DR48WC+++KKOHz+uAQMGKCMj44x15efnq1u3bvrwww91xx136KWXXpK/v7+GDRum1157zVr7hx9+qMDAQLVp08Za+8ldqCpSVFSk9PR0m1teXl5lXq5KGTBggL744gsNHz5cb731lsaOHatjx47ZnPs0e/Zs9e7dW0eOHNGECRP0/PPPq02bNjbhfPbs2Ro0aJAsFoumTZumESNGaOHChercuXO5b8uLi4sVFxen4OBgvfzyyxowYIAkaeTIkXrkkUfUqVMnvfbaaxo+fLg++ugjxcXFWbu7HTp0SD169FBSUpIef/xxvf7667rjjjv0888/V9trAsDxNWjQQD/++KMSEhLUvXt3dejQQZ999pnmz5+vnj17VrqdoKAgzZ49W/Pnz1ezZs30/PPP6+WXX7ZZJiAgQHPnztWSJUvUsmVLffLJJ5o8ebLNMnfccYcmTJigf//732rXrp327t2rYcOGlRvgpqq8vLz0448/qn79+urfv7+aNm2qu+++W8ePH6/SEdGnn35aSUlJatiwofVzp1WrVlq9erV27typLl26qG3btpo4caLNQD6zZs1SeHi4unbtqv79++vee+9VcHDwGetNSEjQgAED1LhxY917770aPXq0Ro4cKenEl8rz58/X119/rTZt2ujaa6/Vxo0bbZ4vNjZWffr0UceOHWUYhpYsWXLWHk6zZs3S0KFD9fDDD6tJkybq16+ffvnlF9WvX7/Sr5GVAQCwmjVrliHJ+OWXX067jL+/v9G2bVvr/UmTJhkn705fffVVQ5Jx+PDh07bxyy+/GJKMWbNmlZvXtWtXQ5Ixc+bMCud17drVev+HH34wJBl169Y1srOzrdM/++wzQ5Lx2muvWadFRkYad95551nbPFNtd955pxEZGWm9/+WXXxqSjKlTp9osN3DgQMPJycnYvXu3dZokw83NzWbab7/9ZkgyXn/99XLPdbIZM2YYkoy5c+dapxUWFhodO3Y0fHx8bNY9MjLS6N279xnbO3lZSeVukyZNqnD5M702FTl69KghyXjppZdOu0xmZqbh6+trdOjQwcjPz7eZV1paahjGiXUNDg42WrRoYbPMokWLDEnGxIkTrdPuvPNOQ5Lx+OOP27S1Zs0aQ5Lx0Ucf2UxftmyZzfQvvvjirNsAgMrJz883/vjjj3LbNqpH9+7djSFDhti7DFTgTO99joACQBX5+PiccTTcstHrvvrqq3MesMfd3b3C8zdOZ+jQofL19bXeHzhwoMLCwrRkyZJzev7KWrJkiSwWS7kBFh5++GEZhqGlS5faTO/evbsaNmxovd+qVSv5+fnpzz/PfEmCJUuWKDQ0VLfddpt1mqurq8aOHaucnBytXr36nNehQ4cO+u6772xuJ59jdD48PT3l5uamVatWlevmWua7777TsWPH9Pjjj5f7Jr+sa9mmTZt06NAh3X///TbL9O7dWzExMRV2gRo1apTN/fnz58vf31/XX3+9zdHe2NhY+fj46IcffpD0z/t30aJF5zwICABUt7y8PE2fPl3bt29XQkKCJk2apO+//1533nmnvUtDFRFAAaCKcnJybMLeqW699VZ16tRJ99xzj0JCQjR48GB99tlnVQqjdevWrdKAQ40aNbK57+TkpMsuu+yCX69z3759Cg8PL/d6lJ2Lsm/fPpvpFXXVqV279mnD2cnP06hRI5uh9c/0PFURGBio7t2729waNKieESrd3d31wgsvaOnSpQoJCdHVV1+tF198UampqdZlyoawb9GixWnbKVu/Jk2alJsXExNTbv1dXFxUr149m2m7du1SVlaWgoODFRQUZHPLycmxXk6ga9euGjBggKZMmWI9L2rWrFkqKCg4txcBAKqBk5OTlixZoquvvlqxsbH65ptv9Pnnn6t79+72Lg1VxCi4AFAFf/31l7Kysk47op104qjXjz/+qB9++EGLFy+2DpN+7bXXavny5eWGMD9dG9XtdAM1lJSUVKqm6nC65zFOGbDoYjJu3DjdeOON+vLLL/Xtt9/qqaee0rRp07Ry5Uq1bdv2gjynu7t7ubBeWlqq4ODgCi81IP1zuQEnJyctWLBAP//8s7755ht9++23uuuuu/TKK6/o559/Pqch9wHgfHl6eur777+3dxmoBhwBBYAq+PDDDyVJcXFxZ1zO2dlZ1113naZPn64//vhDzz77rFauXGnt5ni+o/adateuXTb3DcPQ7t27bUasrV27doVDu5969KwqtUVGRio5Oblcl+SEhATr/OoQGRmpXbt2lTuKXN3Pc6E0bNhQDz/8sJYvX65t27apsLBQr7zyinWeJG3btu20jy9bv8TExHLzEhMTK7X+DRs2VEZGhjp16lTuiG/37t3VunVrm+WvvPJKPfvss9q0aZM++ugjbd++XZ9++mml1xkAgIoQQAGgklauXKlnnnlG0dHR1staVOTIkSPlprVp00aSrN0Yy65zVplrfVXG//3f/9mEwAULFiglJUU33HCDdVrDhg31888/Wy/8LZ04z+/Uy7VUpbZevXqppKREb7zxhs30V199VU5OTjbPfz569eql1NRUzZs3zzqtuLhYr7/+unx8fNS1a9dqeZ7qlpeXp+PHj9tMa9iwoXx9fa3vhR49esjX11fTpk0rt2zZkeH27dsrODhYM2fOtOkKu3TpUu3YsUO9e/c+ay2DBg1SSUlJucsfSCdey7K/99GjR8sdkT71/QsAwLmiCy4AVGDp0qVKSEhQcXGx0tLStHLlSn333XeKjIzU119/fcZh359++mn9+OOP6t27tyIjI3Xo0CG99dZbqlevnjp37izpRAipVauWZs6cKV9fX3l7e6tDhw6Kjo4+p3oDAgLUuXNnDR8+XGlpaZoxY4Yuu+wyjRgxwrrMPffcowULFqhnz54aNGiQ9uzZo7lz59oMClTV2m688UZdc801evLJJ5WUlKTWrVtr+fLl+uqrrzRu3LhybZ+re++9V++8846GDRum+Ph4RUVFacGCBfrpp580Y8aMM56TWx3eeOMNZWZmKjk5WZL0zTffWK8P98ADD8jf37/Cx+3cuVPXXXedBg0apGbNmsnFxUVffPGF0tLSNHjwYEmSn5+fXn31Vd1zzz26/PLLdfvtt6t27dr67bfflJeXpzlz5sjV1VUvvPCChg8frq5du+q2225TWlqaXnvtNUVFRemhhx466zp07dpVI0eO1LRp0/Trr7+qR48ecnV11a5duzR//ny99tprGjhwoObMmaO33npLN998sxo2bKhjx47pvffek5+fn3r16lVNryhwaTnXAemAmupM73kCKABUYOLEiZIkNzc3BQQEqGXLlpoxY4aGDx9+1rBz0003KSkpSR988IHS09MVGBiorl27asqUKdag4urqqjlz5mjChAm67777VFxcrFmzZp1zAH3iiSe0detWTZs2TceOHdN1112nt956S15eXtZl4uLi9Morr2j69OkaN26c2rdvr0WLFpW7oHhVanN2dtbXX3+tiRMnat68eZo1a5aioqL00ksvVelC5Wfj6empVatW6fHHH9ecOXOUnZ2tJk2aaNasWRo2bFi1Pc/pvPzyyzZdlRcuXKiFCxdKkoYMGXLaABoREaHbbrtNK1as0IcffigXFxfFxMTos88+s16bU5LuvvtuBQcH6/nnn9czzzwjV1dXxcTE2ATLYcOGycvLS88//7wee+wxeXt76+abb9YLL7xgHbn2bGbOnKnY2Fi98847euKJJ+Ti4qKoqCgNGTJEnTp1knQiqG7cuFGffvqp0tLS5O/vryuuuEIfffTROb8/gUuVm5ubnJ2dlZycrKCgILm5uVX7KRiAIzEMQ4WFhTp8+LCcnZ0rHFDRybiYR34AAAAA7KiwsFApKSnKy8uzdymAaby8vBQWFkYABQAAAMxmGIaKi4tVUlJi71KAC85iscjFxeW0R/sJoAAAAAAAUzAKLgAAAADAFARQAAAAAIApCKAAAAAAAFMQQAEAAAAApiCAAgAAAABMQQAFAAAAAJiCAAoAAAAAMAUBFAAAAABgCgIoAAAAAMAUBFAAAAAAgCkIoAAAAAAAUxBAAQAAAACmIIACAAAAAExBAAUAAAAAmIIACgAAAAAwBQEUAAAAAGAKAigAAAAAwBQEUAAAAACAKQigAAAAAABTEEABAAAAAKYggAIAAAAATEEABQAAAACYggAKAAAAADAFARQAAAAAYAoCKAAAAADAFARQAAAAAIApCKAAAAAAAFMQQAEAAAAApiCAAgAAAABMQQAFAAAAAJiCAAoAAAAAMAUBtBK6deumbt26VVt7UVFRGjZsWLW1B8nJyUmTJ0+2dxkAAFyyZs+eLScnJ23atMnepQBwYDUqgNakHdu6des0efJkZWZmXtDniYqKkpOTk/Xm7e2tK664Qv/3f/93QZ8XJ0yePNnm9T/5NnPmTOty8+bN05AhQ9SoUSM5OTlV+QuNw4cP68EHH1RMTIw8PT0VHBysK664Qo899phycnKqea3gKIYMGSIPDw/t3Lmz3Lznn39eTk5OWrRokc30goICvf766+rcubNq164tNzc3hYeH66abbtInn3yikpIS67JJSUnl3rd+fn5q06aN3njjDZtl7eWtt97S7NmzK7386bbH0NBQ6zIpKSl6/PHHdc0118jX11dOTk5atWpVler65ptv1LVrVwUHB8vLy0sNGjTQoEGDtGzZsiq1A5hhz549GjlypBo0aCAPDw/5+fmpU6dOeu2115Sfn1/l9qq6XToK9ieAY3CxdwE1wfLly6v8mHXr1mnKlCkaNmyYatWqZTMvMTFRzs7Vl/3btGmjhx9+WNKJHeH//vc/3XnnnSooKNCIESOq7XkcWX5+vlxc7Pd2fvvtt+Xj42MzrUOHDjbz4+PjdfnllysjI6NKbR85ckTt27dXdna27rrrLsXExCgjI0Nbt27V22+/rVGjRpV7blwcpk+friVLlui+++7TypUrrdP37t2rp59+WgMGDFCfPn2s0w8fPqwbbrhB8fHxiouL03/+8x8FBAQoNTVV33//vW6//Xbt3r1bTz31lM3z3HbbberVq5ckKSsrS0uWLNEDDzygffv26aWXXjJnZU/jrbfeUmBgYJV6jVx//fUaOnSozTRPT0/r74mJiXrhhRfUqFEjtWzZUuvXr69STS+//LIeeeQRde3aVRMmTJCXl5d2796t77//Xp9++ql69uxZpfaAC2nx4sW65ZZb5O7urqFDh6pFixYqLCzU2rVr9cgjj2j79u169913q9TmuWyXjoD9CeAgjBpk1qxZhiTjl19+sXcpZ/XSSy8Zkoy9e/de0OeJjIw0evfubTPt0KFDho+Pj9G0adML+twVycnJMf057WnSpEmGJOPw4cNnXG7//v1GSUmJYRiG0bx5c6Nr166Vfo4XX3zRkGT89NNP5eZlZWUZ+fn5Var5fFxqf19H8O677xqSjNmzZ1un9ezZ0/Dz8zP++usvm2Xj4uIMZ2dn4/PPP6+wrV9++cWYO3eu9f7evXsNScZLL71ks1xpaalx+eWXG+Hh4dW4JuemqtuLJGP06NFnXCY7O9vIyMgwDMMw5s+fb0gyfvjhh0q1X1RUZPj5+RnXX399hfPT0tIqXev5KikpMXX7R83z559/Gj4+PkZMTIyRnJxcbv6uXbuMGTNmVLnd022Xjv5/GvuT02N/AjPVqC64lbVlyxbdcMMN8vPzk4+Pj6677jr9/PPP5ZbbunWrunbtKk9PT9WrV09Tp07VrFmz5OTkpKSkJOtyFZ0D+vrrr6t58+by8vJS7dq11b59e3388ceSTnTLfOSRRyRJ0dHR1i4bZW1WdA5oZmamHnroIUVFRcnd3V316tXT0KFDlZ6eXuX1DwoKUkxMjPbs2WMzvbS0VDNmzFDz5s3l4eGhkJAQjRw5UkePHi233OTJkxUeHi4vLy9dc801+uOPP8rVXdYlevXq1br//vsVHBysevXqWecvXbpUXbp0kbe3t3x9fdW7d29t377d5rlSU1M1fPhw1atXT+7u7goLC1Pfvn1tXv9NmzYpLi5OgYGB8vT0VHR0tO666y6bdio6B7Qy74Oydfjpp580fvx4BQUFydvbWzfffLMOHz5c2Zf8rCIiIs75qPeePXtksVh05ZVXlpvn5+cnDw8Pm2kbNmxQr169VLt2bXl7e6tVq1Z67bXXbJZZuXKl9W9Tq1Yt9e3bVzt27LBZpqx78R9//KHbb79dtWvXVufOna3z586dq9jYWHl6eiogIECDBw/WgQMHbNrYtWuXBgwYoNDQUHl4eKhevXoaPHiwsrKyzum1uBTdc8896tSpk/79738rIyNDn376qZYtW6apU6eqbt261uXWr1+vb7/9Vvfee6/69+9fYVvt27fXHXfccdbndHJyUkhISIW9Ct566y01b95c7u7uCg8P1+jRoys81WD+/PnW90dgYKCGDBmigwcP2ixztu0/KipK27dv1+rVq6370eo4H9/X11cBAQHn9Nj09HRlZ2erU6dOFc4PDg62uX/8+HFNnjxZjRs3loeHh8LCwtS/f3+b/XNubq4efvhhRUREyN3dXU2aNNHLL78swzBs2nJyctKYMWP00UcfWf8GZV30Dh48qLvuukshISFyd3dX8+bN9cEHH5Sr70yfXbj4vPjii8rJydH777+vsLCwcvMvu+wyPfjgg9b7s2bN0rXXXqvg4GC5u7urWbNmevvtt20eU5ntMi8vTyNHjlSdOnXk5+enoUOHlvtfQ2J/wv4El6qLrgvu9u3b1aVLF/n5+enRRx+Vq6ur3nnnHXXr1k2rV6+2dos8ePCgrrnmGjk5OWnChAny9vbW//73P7m7u5/1Od577z2NHTtWAwcO1IMPPqjjx49r69at2rBhg26//Xb1799fO3fu1CeffKJXX31VgYGBkk4Ew4rk5OSoS5cu2rFjh+666y61a9dO6enp+vrrr/XXX39ZH19ZxcXF+uuvv1S7dm2b6SNHjtTs2bM1fPhwjR07Vnv37tUbb7yhLVu26KeffpKrq6skacKECXrxxRd14403Ki4uTr/99pvi4uJ0/PjxCp/v/vvvV1BQkCZOnKjc3FxJ0ocffqg777xTcXFxeuGFF5SXl6e3335bnTt31pYtWxQVFSVJGjBggLZv364HHnhAUVFROnTokL777jvt37/fer9Hjx4KCgrS448/rlq1aikpKUkLFy4842tQ2fdBmQceeEC1a9fWpEmTlJSUpBkzZmjMmDGaN29epV7zI0eO2Ny3WCzlXv9zFRkZqZKSEutreibfffed+vTpo7CwMD344IMKDQ3Vjh07tGjRIus/Gd9//71uuOEGNWjQQJMnT1Z+fr5ef/11derUSZs3b7b+bcrccsstatSokZ577jnrB9izzz6rp556SoMGDdI999yjw4cP6/XXX9fVV1+tLVu2qFatWiosLFRcXJwKCgr0wAMPKDQ0VAcPHtSiRYuUmZkpf3//anl9LnZOTk5655131LZtW40aNUpr1qxR+/btNXr0aJvlvvnmG0knzhutqry8POuXXdnZ2Vq6dKmWLVumCRMm2Cw3efJkTZkyRd27d9eoUaOUmJiot99+W7/88ovNPqRsP3P55Zdr2rRpSktL02uvvaaffvrJ+v6Qzr79z5gxQw888IB8fHz05JNPSpJCQkLOuj7Hjx8v9+Wdr69vpfbvZxMcHCxPT0998803euCBB874j2dJSYn69OmjFStWaPDgwXrwwQd17Ngxfffdd9q2bZsaNmwowzB000036YcfftDdd9+tNm3a6Ntvv9UjjzyigwcP6tVXX7Vpc+XKlfrss880ZswYBQYGKioqSmlpabryyiut/1AGBQVp6dKluvvuu5Wdna1x48ZJOvtnFy4+33zzjRo0aKCrrrqqUsu//fbbat68uW666Sa5uLjom2++0f3336/S0lLrPqcy2+WYMWNUq1YtTZ482bqf2Ldvn1atWiUnJydJ7E8k9ie4hNn1+GsVVaZrR79+/Qw3Nzdjz5491mnJycmGr6+vcfXVV1unPfDAA4aTk5OxZcsW67SMjAwjICCgXNfZrl272nTZ6Nu3r9G8efMz1nqmLriRkZHGnXfeab0/ceJEQ5KxcOHCcsuWlpae8XkiIyONHj16GIcPHzYOHz5s/P7778a//vWvct1G1qxZY0gyPvroI5vHL1u2zGZ6amqq4eLiYvTr189mucmTJxuSbOou+3t07tzZKC4utk4/duyYUatWLWPEiBE2baSmphr+/v7W6UePHq2w+9/Jvvjii0p155FkTJo0yXq/su+DsnXo3r27zWv90EMPGRaLxcjMzDzj85Z1wT31FhkZedrHVLULUGpqqhEUFGRIMmJiYoz77rvP+Pjjj8vVVlxcbERHRxuRkZHG0aNHbeadvG5t2rQxgoODrV2GDMMwfvvtN8PZ2dkYOnRouXW77bbbbNpKSkoyLBaL8eyzz9pM//333w0XFxfr9C1bthiSjPnz51d6XXF6EyZMMCQZFovFiI+PLzf/5ptvNiSVe1/k5+db9w+HDx+2eW+UdcGt6DZq1Cib982hQ4cMNzc3o0ePHtbu5IZhGG+88YYhyfjggw8MwzCMwsJCIzg42GjRooVNd65FixYZkoyJEycahlG57d8wzq3LXEW3WbNmVbh8VbvMGcY/+2xvb2/jhhtuMJ599tkK/yYffPCBIcmYPn16uXllr+2XX35pSDKmTp1qM3/gwIGGk5OTsXv3bpt1c3Z2NrZv326z7N13322EhYUZ6enpNtMHDx5s+Pv7G3l5eYZhVO6zCxePrKwsQ5LRt2/fSj+m7L1ysri4OKNBgwY2087WBTc2NtYoLCy0Ti87leSrr74yDIP9ycnYn+BSdFF1wS0pKdHy5cvVr18/NWjQwDo9LCxMt99+u9auXavs7GxJ0rJly9SxY0e1adPGulxAQECluqfVqlVLf/31l3755Zdqqfvzzz9X69atdfPNN5ebV/ZN4ZksX75cQUFBCgoKUsuWLfXhhx9q+PDhNoOHzJ8/X/7+/rr++uuVnp5uvcXGxsrHx0c//PCDJGnFihUqLi7W/fffb/McDzzwwGmff8SIEbJYLNb73333nTIzM3XbbbfZPJfFYlGHDh2sz+Xp6Sk3NzetWrWqwq45kqzfbC5atEhFRUVnfS2kqr0Pytx77702r3WXLl1UUlKiffv2Veo5P//8c3333XfW20cffVSpx1VGSEiIfvvtN9133306evSoZs6cqdtvv13BwcF65plnrEclt2zZor1792rcuHHlBr4qW7eUlBT9+uuvGjZsmM03ra1atdL111+vJUuWlHv+++67z+b+woULVVpaqkGDBtn8fUNDQ9WoUSPr37fsCOe3336rvLy8ans9LlVlPSHCw8PVokWLcvPL3tOnDkg1c+ZM6/4hKCjIpht1mXvvvdf63v388881evRovfPOOxo/frx1me+//16FhYUaN26cTXfyESNGyM/PT4sXL5Z0osv8oUOHdP/999t0D+/du7diYmKsy1Vm+z9Xffv2tdkev/vuO8XFxVVb+1OmTNHHH3+stm3b6ttvv9WTTz6p2NhYtWvXzqYr++eff67AwMAK959l2+SSJUtksVg0duxYm/kPP/ywDMPQ0qVLbaZ37dpVzZo1s943DEOff/65brzxRhmGYbNNxsXFKSsrS5s3b5ZU/Z9dcGxl+wRfX99KP+bkwXWysrKUnp6url276s8//6zSqRP33nuv9QimJI0aNUouLi7Wzxj2J/9gf4JL0UUVQA8fPqy8vDw1adKk3LymTZuqtLTUeo7avn37dNlll5VbrqJpp3rsscfk4+OjK664Qo0aNdLo0aP1008/nXPde/bsqfAfysrq0KGDvvvuOy1btkwvv/yyatWqpaNHj8rNzc26zK5du5SVlaXg4GCbf0aDgoKUk5OjQ4cOSZI1cJ36OgQEBJy2S2l0dLTN/V27dkmSrr322nLPtXz5cutzubu764UXXtDSpUsVEhKiq6++Wi+++KJSU1OtbXXt2lUDBgzQlClTFBgYqL59+2rWrFkqKCg47etRlfdBmfr169vcL1vXyn6QXX311erevbv1drrzOc5VWFiY3n77baWkpCgxMVH//e9/rd2e33//fUmyngNypvdS2d/3dK9Nenq6tRt1mYr+voZhqFGjRuX+vjt27LD+faOjozV+/Hj973//U2BgoOLi4vTmm29y/uc5OHDggCZNmqQWLVrowIEDevHFF8stU/ZP5qmX5RkwYID1n6ZWrVpV2H6jRo2s793+/fvrjTfe0P33368ZM2bo999/l3T6946bm5saNGhgnX+m91hMTIx1fmW2/3NVr149m+2xe/fuFZ7/dj5uu+02rVmzRkePHtXy5ct1++23a8uWLbrxxhutpyvs2bNHTZo0OeMI3fv27VN4eHi5kNC0aVPr/JOduj0ePnxYmZmZevfdd8ttj8OHD5ck6zZZ3Z9dcGx+fn6SpGPHjlX6MT/99JO6d+9uHR8gKChITzzxhCRVad/dqFEjm/s+Pj4KCwuzno/J/sQW+xNcai66c0DN0LRpUyUmJmrRokVatmyZPv/8c7311luaOHGipkyZYno9gYGB6t69uyQpLi5OMTEx6tOnj1577TXrEYzS0lIFBwef9sjc6c5PrYyTvzEtey7pxHmgJ18rq8zJO89x48bpxhtv1Jdffqlvv/1WTz31lKZNm6aVK1eqbdu2cnJy0oIFC/Tzzz/rm2++0bfffqu77rpLr7zyin7++edqu/zIyUdwT2acctK+vTk5Oalx48Zq3LixevfurUaNGumjjz7SPffcc8Ges6K/r5OTk5YuXVrh63by3+SVV17RsGHD9NVXX2n58uUaO3aspk2bpp9//tlmwCqc2ZgxYySdGNhr/PjxevbZZ3X77bfbHOGPiYmRJG3bts3mC5CIiAhFRERIOvHFSmUHNrvuuuv0xhtv6Mcff1TLli2ra1VsnG37rwn8/Px0/fXX6/rrr5erq6vmzJmjDRs2qGvXrhfk+U63vx0yZMhpzxEv++LB0T67cGH5+fkpPDxc27Ztq9Tye/bs0XXXXaeYmBhNnz5dERERcnNz05IlS/Tqq69a32uOiv1J1bE/gb1cVEdAg4KC5OXlpcTExHLzEhIS5OzsbP1HLDIyUrt37y63XEXTKuLt7a1bb71Vs2bN0v79+9W7d289++yz1m+qKtN1tkzDhg0r/QFRGb1791bXrl313HPPWY9mNWzYUBkZGerUqVO5b/K6d++u1q1bSzrxukjlX4eMjIxKHw1s2LChpBMn11f0XKeOOtewYUM9/PDDWr58ubZt26bCwkK98sorNstceeWVevbZZ7Vp0yZ99NFH2r59uz799NMKn78q74OarEGDBqpdu7ZSUlIk/fO6n+m9VPb3Pd1rExgYKG9v7zM+b9lAB9HR0RX+fU8drbdly5b6z3/+ox9//FFr1qzRwYMHNXPmzCqt66Xsiy++0Ndff61nnnlG9erV04wZM+Tm5lZuEKKy64FWV/fv4uJiSf8cUT3de6ewsFB79+61zj/TeywxMdE6v8zZtv+q7EvtrX379pJks00mJiae8fSByMhIJScnlztKlZCQYJ1/JkFBQfL19VVJSUmF22P37t1tRtI822cXLi59+vTRnj17KnVtym+++UYFBQX6+uuvNXLkSPXq1Uvdu3cvF1Kks2+XZT2hyuTk5CglJcU6yB37k7Njf4KL2UUVQC0Wi3r06KGvvvrK5jIeaWlp+vjjj9W5c2drl5S4uDitX79ev/76q3W5I0eOVOqft4yMDJv7bm5uatasmQzDsO4Yyv6Jr2g48VMNGDBAv/32m7744oty8871CNxjjz2mjIwMvffee5KkQYMGqaSkRM8880y5ZYuLi611XnfddXJxcSk37Pobb7xR6eeOi4uTn5+fnnvuuQp3lGWXN8nLyyu3k2rYsKF8fX2tXWyPHj1a7jUoO2/3dN1wq/I+qAk2bNhQrlusJG3cuFEZGRnWrknt2rVTdHS0ZsyYUe59V/YahoWFqU2bNpozZ47NMtu2bdPy5cvVq1evs9bTv39/WSwWTZkypdzfxjAM6/aRnZ1tDTFlWrZsKWdn5zN2ocY/jh07prFjx6pt27bW837Cw8P1zDPPaNmyZZo/f7512U6dOun666/Xu+++q6+++qrC9qqyPykbVbfsy6nu3bvLzc1N//3vf23aef/995WVlaXevXtLOvFPU3BwsGbOnGnzd166dKl27NhhXa4y2790Yl9amf2oWfLy8k77z3zZ+VVl2+SAAQOUnp5e4f6z7DXs1auXSkpKyi3z6quvysnJSTfccMMZ67FYLBowYIA+//zzCr98OvlyUpX57MLF5dFHH5W3t7fuuecepaWllZu/Z88e62W6ynq0nLx9Z2VladasWeUed7bt8t1337V5T7399tsqLi62vp/Zn5zA/gSXqhrZBfeDDz6wXqvoZA8++KCmTp2q7777Tp07d9b9998vFxcXvfPOOyooKLA5b+rRRx/V3Llzdf311+uBBx6wXoalfv36OnLkyBm/JevRo4dCQ0PVqVMnhYSEaMeOHXrjjTfUu3dva7/72NhYSdKTTz6pwYMHy9XVVTfeeGOFR5ceeeQRLViwQLfccovuuusuxcbG6siRI/r66681c+ZM6z+AVXHDDTeoRYsWmj59ukaPHq2uXbtq5MiRmjZtmn799Vf16NFDrq6u2rVrl+bPn6/XXntNAwcOVEhIiB588EG98soruummm9SzZ0/99ttvWrp0qQIDAyv17aGfn5/efvtt/etf/1K7du00ePBgBQUFaf/+/Vq8eLE6deqkN954Qzt37tR1112nQYMGqVmzZnJxcdEXX3yhtLQ0DR48WJI0Z84cvfXWW7r55pvVsGFDHTt2TO+99578/PzOGJYq+z4wy48//qgff/xR0okdeG5urqZOnSrpxPmjV1999Wkf++GHH+qjjz7SzTffrNjYWLm5uWnHjh364IMP5OHhYT0/x9nZWW+//bZuvPFGtWnTRsOHD1dYWJgSEhK0fft2ffvtt5Kkl156STfccIM6duyou+++23oZFn9//3LXUq1Iw4YNNXXqVE2YMEFJSUnq16+ffH19tXfvXn3xxRe699579e9//1srV67UmDFjdMstt6hx48YqLi7Whx9+aP2Aw9n95z//UXJyshYuXGjT3Xn06NGaM2eOxo0bp549e1r3O3PnzlXPnj3Vr18/3XDDDerevbtq166t1NRUff/99/rxxx8r/Adk8+bNmjt3rqQToXfFihX6/PPPddVVV6lHjx6STnwzPmHCBE2ZMkU9e/bUTTfdpMTERL311lu6/PLLrZd/cXV11QsvvKDhw4era9euuu2226yXTYiKitJDDz0kSZXa/qUT+9K3335bU6dO1WWXXabg4GBde+215/3alm1/Zdcm/vDDD7V27Vrr6346eXl5uuqqq3TllVeqZ8+eioiIUGZmpr788kutWbNG/fr1s3b3Gzp0qP7v//5P48eP18aNG9WlSxfl5ubq+++/1/3336++ffvqxhtv1DXXXKMnn3xSSUlJat26tZYvX66vvvpK48aNs/ZsOJPnn39eP/zwgzp06KARI0aoWbNmOnLkiDZv3qzvv//eepmoynx24eLSsGFDffzxx7r11lvVtGlTDR06VC1atFBhYaHWrVun+fPnW6/v3aNHD7m5uenGG2/UyJEjlZOTo/fee0/BwcHWo3BlzrZdFhYWWrfvsv1E586dddNNN0lif1KG/QkuWeYNuHv+yob3Pt3twIEDhmEYxubNm424uDjDx8fH8PLyMq655hpj3bp15drbsmWL0aVLF8Pd3d2oV6+eMW3aNOO///2vIclITU21LnfqZVjeeecd4+qrrzbq1KljuLu7Gw0bNjQeeeQRIysry6b9Z555xqhbt67h7Oxsc0mWUy/DYhgnLgEzZswYo27duoabm5tRr14948477yw3DPapIiMjjd69e1c4b/bs2eWGC3/33XeN2NhYw9PT0/D19TVatmxpPProo0ZycrJ1meLiYuOpp54yQkNDDU9PT+Paa681duzYYdSpU8e47777yv09TneJlB9++MGIi4sz/P39DQ8PD6Nhw4bGsGHDjE2bNhmGYRjp6enG6NGjjZiYGMPb29vw9/c3OnToYHz22WfWNjZv3mzcdtttRv369Q13d3cjODjY6NOnj7WNMjrlMixljz3b++B06/DDDz9Uaij1skuVHD58uFLLVXQ7te5Tbd261XjkkUeMdu3aGQEBAYaLi4sRFhZm3HLLLcbmzZvLLb927Vrj+uuvN3x9fQ1vb2+jVatWxuuvv26zzPfff2906tTJ8PT0NPz8/Iwbb7zR+OOPP6q0bp9//rnRuXNnw9vb2/D29jZiYmKM0aNHG4mJiYZhGMaff/5p3HXXXUbDhg0NDw8PIyAgwLjmmmuM77///ozrixM2bdpkWCwWY8yYMRXO37hxo+Hs7GyMHTvWZnp+fr4xY8YMo2PHjoafn5/h4uJihIaGGn369DE++ugjm0smVXQZFhcXF6NBgwbGI488Yhw7dqzc877xxhtGTEyM4erqaoSEhBijRo0qd9kfwzCMefPmGW3btjXc3d2NgIAA44477jD++usv6/zKbP+GceIyRL179zZ8fX0NSWe9hIJOuQTVmZY73e1MioqKjPfee8/o16+fERkZabi7uxteXl5G27ZtjZdeeskoKCiwWT4vL8948sknjejoaMPV1dUIDQ01Bg4caHOJqGPHjhkPPfSQER4ebri6uhqNGjUyXnrppXKX4TrTuqWlpRmjR482IiIirM9z3XXXGe+++651mcp+duHis3PnTmPEiBFGVFSU4ebmZvj6+hqdOnUyXn/9deP48ePW5b7++mujVatWhoeHhxEVFWW88MIL1st/nHxZudNtl2WfqatXrzbuvfdeo3bt2oaPj49xxx132Fz6qwz7E/YnuDQ5GYaDjbJiZ+PGjdM777yjnJyc0w5McynKzMxU7dq1NXXqVOsFnAEAAACgKi6qc0CrKj8/3+Z+RkaGPvzwQ3Xu3PmSDp+nvi6SNGPGDEkqN4AQAAAAAFRWjTwHtLp07NhR3bp1U9OmTZWWlqb3339f2dnZeuqpp+xdml3NmzdPs2fPVq9eveTj46O1a9fqk08+UY8ePar9+pYAAAAALh2XdADt1auXFixYoHfffVdOTk5q166d3n///TMOCHMpaNWqlVxcXPTiiy8qOzvbOjBR2Un2AAAAAHAuOAcUAAAAAGCKS/ocUAAAAACAeS6KLrilpaVKTk6Wr69vpa5TCeDCMQxDx44dU3h4uJyda853XOxHAMfBfgTA+aip+5BLxUURQJOTkxUREWHvMgCc5MCBA6pXr569y6g09iOA42E/AuB81LR9yKXiogigvr6+kk68yfz8/OxcDXBpy87OVkREhHW7rCnYjwCOg/0IgPNRU/chl4qLIoCWdXPx8/Njhw84iJrW/Yz9COB42I8AOB81bR9yqaBTNAAAAADAFARQAAAAAIApCKAAAAAAAFMQQAEAAAAApiCAAgAAAABMQQAFAAAAAJiCAAoAAAAAMAUBFAAAAABgCgIoAAAAAMAUBFAAAAAAgCkIoAAAAAAAUxBAAQAAAACmIIACAAAAAExBAAUAAAAAmIIACgAAAAAwBQEUAAAAAGAKAigAAAAAwBQEUAAAAACAKQigAAAAAABTEEABAAAAAKYggAIAAAAATOFi7wJQs+Xl5SkhIaFa2srPz1dSUpKioqLk6elZLW3GxMTIy8urWtoCAAAAcH4IoDgvCQkJio2NtXcZpxUfH6927drZuwwAAKqEL3gBXKwIoDgvMTExio+Pr5a2duzYoSFDhmju3Llq2rRptbQZExNTLe0AAGAmvuAFcLEigOK8eHl5VfsHUNOmTflQAwBc0viCF8DFigAKAADgYPiCF8DFilFwAQAAAACmIIACAAAAAExBAAUAAAAAmIJzQE9xKQx7bhiG3njjDSUnJ1dLTdUlNTVVkvT6668rNDTUztXYiomJ0Z133mnvMgAAAIAajQB6ikth2PN169Zp7Nix8g4LlcXVtZoqO38lhYWSpPnLlsri5mbnav5RnH9ceYcP65prrlH9+vXtXQ4AAABQYxFAT3EpDHteUlIiSWo38015R0Wed3vVpST/uHKSkuQTFSWLp4e9y7HK2LBRv9x7v/V1AwAAAHBuCKCnYNhz+7F4esi/KdcVAwDUbC+99JI8PBzni9SyU27ef/99LV261M7V/KN+/fr617/+Ze8yAJiMAHoJK87Pl2EYcnJysncpDssoLVVxXp69ywAA1CCvvPWmnJwdZ5zH0uJiSdIHH38kZxfH+NevtLhYxzOz1LJlS7Vp08be5QAwkWPshWAX6wcPkcXNVR4BdeQWUFsuAbXlGhAg94Dacvt7mltAgNzrBMgtIEButWvL2bXmv2VKCwtVcOSICo8cVeGRIyrIOKLCsvsZGSo6elTFR46q4MgRHT9y1PrBDQBAZXT65gu5+PjYuwwrRzzFJTtxp9YNul1FRUX2LgWAyWp+mkCVtWvXTqtWrVJaWpoOHTpkvaWlpSklLU1p23for0OHlHvsWLnHetTyl3tAgFxq15ZrnQC5+PrKL6aJ/Fu2sMOanFnGho3KTdqn4uxsFR/NVFHGERUcOaKCCtbLt1YtBQUFKSokRGFhdRXcuq2Cg4MVEhKi4OBgBQcHKzw83A5rAQDA+eEUFwCOhAB6CfLx8VHXrl0lSUVFReVCaNnv+/bt07Zt25SYmGgdgOd4ZpaOZ2ZJ2nuiMWdnqbTUTmtSCU5OkmFUOMvNzU1NmzZV8+bNFRERYRM2y34PDAyUi4N0VwIAAABqOv6zvgRt2bJFg2+/Talph5R99Gi5+W7e3vIIrCPX2rXlEhKs8GYxJ7rgBtSWe51/uua61QmQs4urcvfvt8NaVI53RMQ/XW7/7mpbcOSoCo9kqPDIUaUcOaID235X0Y8/6viRIyo65XxPJycn1QoIUEhIiFZ89x1HQQEAZ3XwmyXyql9P7oGBcg8MlFvtWg51TqjZjNJSFWVmqSA9XccPp6vg8GEdS9xp77IA2MlFFUBXr14tb29ve5dhtXPniZ3rxo0blZmZad9iTrJ161btTEhU1J3/Uv0G0SfO8wyoLbc6deRWu5YsVRy5z/G79XjLLaC2dFnDsy5ZnJf/z/mgRzJUcOSosrb+roQvvlJBQYEJtQIAarodz79oc9/ZxUWegYFyD6wjl6BAazB1Dw468TMoUB5BgXKrXVtOFoudqq46o7RUhUePquDQYRWkp6vg8N8BMz1dhYfTVZSersL0dOWlZ6j0lHM9awUEqFnz5vJxoHNlAZjjogqgN910k71LqNCoUaPsXUKFIvr3c6jrgDoCFy9PuXjVlVe9utZpXnXD9dcXX9mxKgBATXL48GHl5eUpJSVFycnJSklJsf6enJKiv3buVurqNco4fFjGSaeJODk7yzOwjtwDAxUS10O1O1xux7Wo2KEVK3V4xQ8qPpaj/CNHZJxyjew6QUEKDQtTvfBw1b38CoWFhSksLEzh4eHW30NDQ+Xu7m6nNQBgbxdVAG3zygvyi2li7zKsSo4XKO+vv+RVr54sHo6zo01Z8q12vfm2fhpwq5ycHecSLIZhSIYkJznUpWGM0hOXquHDEgBQGU5OTpW+GaeOU2AYkmHoeEaG1g8eYp8VqASP2rXKjbFQlfUGcOm6qALowUVLlPnb73Lx8ZGLj49cfX2sv7v4+sj1758u3j6mXU7EtxLdPqtLaVGRio/lqCgnR8U5OSo+duLnqfdz9504Z/M/Tzyh4OBg0+o7mwMHDuiFF17QY489poiICHuXY6Nx48ac/wkAqJTAwECb+04Wi7xO7oJ7WQP5d7xCwUH/dMF1DwqSW+1a1ut0luQfV1ivnvYov1J8oqLk7OaqwqNH/z6v80TX24JDh5WVnq7D6enauOFnFaRnKD89vdwlzWrXqaPwsDDN/+wzNW3a1E5rAcAeLqoAGno0SzkHDupoVpaOZWer8Azn7Ll6esrt74Bq8fGRxcdbFh8fOXucGKq8Vts25hVeSRk/b1DOnj9Vcvy4SnJyVJqTq5KcHBUdO6bCYzkqPsP6unt4yNfPT37+/orw91er7t117733qm7duqd9jNk2b96sF154QYMGDVK7du3sXc4FkZeXp4SEhGppKz8/X0lJSYqKipKnp2e1tBkTEyMvL69qaQsALlVNH39EXhH1rOHSrXbtKg9CVFMunVJ2PqvOkCFPnCuaeSKg/h1SsxN3avsn85STk2NesQAcwkUVQPvccIMKCgqUlZWlrKwspR06pP0HDij5r7/KXei4KD9fRfn50qHD1mkuHh6yeHgo+etFZpdeaW7+/io9fvyMYdO6rJub6kZEqH5EhIKDguTv729z48R/8yUkJCg2NtbeZZxWfHz8RRv+AcAsdW/sLRc+Y62cnJ3lXidA7nUCpCaNJUnZiTu1/5N5dq4MgD1cVAH01Zkz5V7L/8SRzL+PaLq2aaW6na+ydsM90TXX17Zbro+PXHy85ezqqpL848pJSrL3qpyWT1SULJ4eKi0sVHFO7onutceOnehq+3cX27Jb0bEcFebkKCEnR9sP7FfpjhyV5OSqIPOojmdmqWfPnvL397f3Kl1SYmJiFB8fXy1t7dixQ0OGDNHcuXOrrftSTIzjf9sOAI5u9zvvydnNzd5lWJUWF6soM1Outf7p4mtvhRlH7F0CADtxjL1QNYl54jH5/v3NWmUZpaUqys5WUXa2dZqLA3dBPH7oULlpZee5eoRWro2sbdu19YmnqrkyVIaXl1e1H2Fs2rQpRy0BwIEUrVojZwe67mdhQYHSUlIUEhYmNwcZUM9JUqerr1bjxlX7vw1AzXdRBdBfH3nc3iXUGB5eXvLz87N3GQAAXHT++P13h/qM3bx5s2JjY7Vk0SK+sARgdxdVAF2yZIm8vb3tXYbVzp07NWLECL333nsO9w1feHi4QkMrecgUQJUw2BQAAEDFLqoA2qlTJ4f6xrFskJ927drxjSNwCWGwKQAAgIpdVAEUABwBg00BAABUjAAKnEVRUZGeeuoppaWl2bsUGxkZGZKkyZMnq06dOnau5h+FhYX2LsHuGGwKAACgYgRQ4Cx+/PFHvfDCC6rdrKmcHWT0QEkqzsuTJP24M9GhRm4uzMy0dwkAAABwUARQnJfqHGxlx44dNj+rQ3UOttJ86hR5N4iWk5NTtbR3vsquWVt2bVhHYJSWKmXZcm2d8B97lwIAAAAHRADFebkQg60MGTKk2tqqzsFW1vYfJCeLRW4+PnL7+9qrFm8vOfv4yOLjIxcfb7n4+MjVx0cuvj5y8T5x/+TfXf/+3cliOe96LJ4e8m9afefylRYVqzg3R8U5OSo+lqOinByV5Oaq6Njf03JzVfz372XzSnNyVZJTNi1XBTk5kmFUW00AAAC4uBBAcV6qc7CVC3W5ifPVtm1bLVq0SFlZWUpJSdGBAwe0f/9+HThwQAcOHKjyuaGRgwcpvN9N511Xddvz1js69OOaSi/v7OyssLAw1a9fXxFt2ykiIkL169eXj4+P7r777gtYKQAAAGoqAijOS3UPttKpU6dqa6u6/PHHHxryr3/pWHa2SkpKTrucq7eX3Hx8rUdCnb29Zfn7dxefE0c+i3NzlfR/c7Xv089MXIPKi/rX7bJ4e5844pmTq+KcHJXk5Ko098SRzqK/j3QW5eertLRUBw8e1MGDB7V+/XprGxYXdisAAACo2Dn9p/jmm2/qpZdeUmpqqlq3bq3XX39dV1xxRYXLduvWTatXry43vVevXlq8eLEkadiwYZozZ47N/Li4OC1btuxcyjsvl9I5jaicgoICZR49qkaj75NnvXr/dLO1CZdelepWW5J/XGG9eppQ9bmp7Pmk/3TXzf07rP7z+5H4zfpr4ZcXvlgAAADUOFUOoPPmzdP48eM1c+ZMdejQQTNmzFBcXJwSExMVHBxcbvmFCxfaXJYhIyNDrVu31i233GKzXM+ePTVr1izrfXc7jTZ6KZ3TiKo5uuU3ZSck2rsMK6OkRMW5eZUOv2Y5fuiQvUsAAACAg6pyAJ0+fbpGjBih4cOHS5JmzpypxYsX64MPPtDjjz9ebvmAgACb+59++qm8vLzKBVB3d3eFhoZWqoaCggIVFBRY72dnZ0s6cb3GoqKiKq3PqRo2bKgNGzacVxtl8vPztW/fPkVGRlbbOY0NGzY873VE1bRv315D77xTaamp9i7FRnZettb/skkdO3aUn1/1vL+qQ3FwqFZou73LOKsLuR+pTsXFxdafjlQXcCHVlPc6+xHAMfE+d2xVCqCFhYWKj4/XhAkTrNOcnZ3VvXt3m3PAzuT999/X4MGD5e3tbTN91apVCg4OVu3atXXttddq6tSpqlOnToVtTJs2TVOmTCk3ffny5Q7XPdXf31+ZmZnKrKZrI6akpFRLO6ia/jffbO8SytmzZ4/Wr1+vgQMHqmHDhvYuxyovL08rVqywdxlnVVP2I3v27JEkrV27lu0fl4y8v69z7OjYjwCOqabsQy5VToZR+WsmJCcnq27dulq3bp06duxonf7oo49q9erVZz1yuHHjRnXo0EEbNmywOWe07KhodHS09uzZoyeeeEI+Pj5av369LBV0LazoG8eIiAilp6fLz8+vsqsD1Ghbtmyxbk9t27a1dzlW2dnZCgwMVFZWlkNvjzVlP+Kof2fgQmI/Ur3Yj+BSU1P2IZcqU4erfP/999WyZctyAxYNHjzY+nvLli3VqlUrNWzYUKtWrdJ1111Xrh13d/cKzxF1dXWVq6tr9RcOOCCXv0ebdXFxcaj3vSPVciY1ZT/iqH9n4EKqKe/1C7kfqa5BEUtKSvTVV19Jkr766is5OTlV+OV+VTEoIhxZTdmHXKqqFEADAwNlsVjKXfcwLS3trOdv5ubm6tNPP9XTTz991udp0KCBAgMDtXv37goDKAAAwMXsQgyK+Nxzz+m5556rlrYYFBHAuapSAHVzc1NsbKxWrFihfv36SZJKS0u1YsUKjRkz5oyPnT9/vgoKCio1Iuxff/2ljIwMhYWFVaU8AACAi0JMTIzi4+PP+fErV67Uo48+qi5duuj222+Xi4uLiouL9fHHH2vNmjV68cUXde21155XfQBwLqrcBXf8+PG688471b59e11xxRWaMWOGcnNzraPiDh06VHXr1tW0adNsHvf++++rX79+5QYWysnJ0ZQpUzRgwACFhoZqz549evTRR3XZZZcpLi7uPFYNcDxcZxYAUBleXl7nfISxpKREAwYMUJ8+ffTll1/K2dnZOm/EiBHq16+f3nrrLT300EPV0h0XAKqiygH01ltv1eHDhzVx4kSlpqaqTZs2WrZsmUJCQiRJ+/fvt9nRSVJiYqLWrl2r5cuXl2vPYrFo69atmjNnjjIzMxUeHq4ePXromWeesdu1QIELhevMAgAutDVr1igpKUmffPJJuf/JnJ2dNWHCBF111VVas2aNunXrZp8iAVyyzmkQojFjxpy2y+2qVavKTWvSpIlON9iup6envv3223MpA6hxzrdL1cny8/OVlJSkqKioarvOLF2qAKDmK7vUSosWLSqcXzadS7IAsAdTR8EFLnXn06WqIp06daq2tgAAF4eyMTS2bdumK6+8stz8bdu22SwHAGYigALASQ4fPqzjx4/buwyrI0eOWH8eOnTIztXYCgoKkpOTk73LAHCKLl26KCoqSs8991y5c0BLS0s1bdo0RUdHq0uXLnasEsCligAKACe57LLL7F1Cha6//np7l1DOlClTNHHiRHuXAeAUFotFr7zyigYOHKh+/fppwoQJatGihbZt26Zp06Zp0aJFWrBgAQMQAbALAigAnKT1C8/K4uFh7zKs8g78pYSXX1XMvx+SV0Q9e5djtXvme/rtt9/sXQaA0+jfv78WLFighx9+WFdddZV1enR0tBYsWKD+/fvbsToAlzICKACcJKhzJ7n4+Ni7DKuS/OOq3T5WPlFRsng6TjA+MP9ze5cA4Cz69++vvn37as2aNUpJSVFYWJi6dOnCkU8AdkUABQAHZvH0kH9TRicGcG4sFguXWgHgUJzPvggAAAAAAOePAAoAJzndNYvxD8MwZBSX2LsMAABQA9EFFwBOsuq6nvKpV1du4WHyDA+XZ92/b3//7urnd9FfesQwDBVlZSk/OUX5Bw8q/2Cy9ffC5BTlHkxWcUGBxCAmAACgigigAHCSZyZNUmpqqv7cu1d7tu/QvsVLlZeTY53v5uMj77rhcgsPk8ffodTrpIDq4u1tx+orrzgnR/nJKcqrIGDmHUxWYW6udVkvHx9FRkWpVXS0GrRrr6ioKEVHR6tFixZ2XAMAAFATEUAB4CSjR4+Wn5+f9b5hGMrIyFBSUpL27t1r/fnnn39qz6bN2jP/cxUVFlqX96jlL8+wMF328ENy8XGsMFp87JgSp72ogowMFWRlW6e7urkpMjpaTRs0UPR13RUdHa3o6Ghr0AwICLjoj/oCAABzEEAB4AwKCwt19OjRcreMo0eVefSoTfh0dnGRm5+fDItFv9wz0o5Vn5lXSLCcc/NUWlwsSSoqLFTm0aPKOHJEtWvV0tGjR1WrVi3V+vt3X19fubm52blqAABwMSCAAsBJnn32WaWkpGj3n39q7969SktJsQ5M5OTsLO/QEHmEn+iCWyu2jcLC/zlH1CM4SE4Wi0ryjysnKcm+K3IGPlFRcnJ1UcHh9BPneP7dFfdgcor+3JmoglU/KCc1TSpbbycnhYaHq0F0tBo2aGA9Mtq2bVu1bt3azmsDAABqEgIoAJzk9Q/el1e9enILD5dX755qftJARB4hoXJ2Pftus6Zcu9MzLFSeYaEVzistKtLx1DTlJycr72Cy8pOT9efBZCX89quOL1uq3EOH1b9/f33++ecmVw0AAGoyAigAnKTTV5/LxcfH3mXYnbOrq7wi6skrop7qVDA/fvRY02sCAAA1H9cBBQAAAACYggAKAAAAADAFXXAB4CS/TXhKzi6Os2s0SkpUnJcnFy8vOVks9i7HKuuPBCm44vNHAQAATsdx/ssCAAfQ2s1dLq6u9i7DKisrS+t/2aSOV10lfy8Huq5ohw568MEH7V0FAACoYQigAHCSzz//XH5+fvYuw2rz5s2KjY3VG6+/rnbt2tm7HAAAgPNCAAUAAAAuEXl5eUpISKiWtvLz85WUlKSoqCh5enpWS5sxMTHy8vKqlrbgmAigAAAAwCUiISFBsbGx9i7jtOLj4+nxc5EjgAIAAACXiJiYGMXHx1dLWzt27NCQIUM0d+5cNW3atFrajImJqZZ24LgIoAAAAMAlwsvLq9qPMDZt2pSjlqg0rgMKAAAAADAFARQAAAAAYAoCKAAAAADAFARQAAAAAIApGIQIAIAahGv4AQBqMgIoAAA1CNfwAwDUZARQAKhm1XmEaseOHTY/qwNHqGo2ruEHAKjJCKAAUM0uxBGqIUOGVFtbHKGq2biGHwCgJiOAAkA1q84jVBfqHD0AAAB7IIACQDWr7iNUnTp1qra2AAAA7InLsAAAAAAATEEABQAAAACYgi64AOCgSkpKtGbNGqWkpCgsLExdunSRxWKxd1kOi9cLAADHRwAFAAe0cOFCPfzww0pKSrJOi4qK0iuvvKL+/fvbrzAHxesF4GJ29OhRXdejh/bv22fvUmwUFxVJkq697jq5uLrauZp/lJaW2rsEnAEBFAAczMKFCzVw4ED16dNHn3zyiVq0aKFt27bpueee08CBA7VgwQJC1Ul4vQBc7Pbs2aMtmzap/q23yD0o0N7lWB0/dFhZny2QX4/u8ggOsnc5Vrn7D+jo14vsXQZOw8kwDMPeRZyv7Oxs+fv7KysrS35+fvYuB7ik1dTt0VHqLikp0WWXXaaWLVvqyy+/lLPzP6fql5aWql+/ftq2bZt27dpF91Lxep2vzZs3KzY21uGuDeso22NV1dS64fg2bdqkyy+/XFfN+0h+MU3sXY5VSf5x5SQlyScqShZPD3uXY3X45w2KHzmabdFBMQgRADiQNWvWKCkpSU888YRNmJIkZ2dnTZgwQXv37tWaNWvsVKFj4fUCcCnJ3pGgnKQkFR3LkSMcQ7J4esi/aYxDhE/DMFSUfUw5e5OUs2u3vcvBGdAFFwAcSEpKiiSpRYsWFc4vm1623KWO1wvApWTb5Gesv7u4u8szMFBugXXkEhAgt8A6cg+sI7c6J3661wn8+36ALO7udqz6/JQcP66CjCMqzMhQQXqGCtLTVZCRocL0DBVkZKj473n5h9NVUlho73JRCQRQAHAgYWFhkqRt27bpyiuvLDd/27ZtNstd6ni9AFwKYmNjlZWVpdTU1ApvKampOrhrj9LW/KT0Q4fKDcLj7usrj8A6cq1TR5Za/vJv2UJ1Olxhp7U5vfSf1in7jwQVZ2aqOOOIjmdkqODYMZtlLBaLAoODFRIaqrphYQpr31ChoaE2N29vb8XGxtppLXA2BFAAcCBdunRRVFSUnnvuuQrPaZw2bZqio6PVpUsXO1bpOHi9AFwKnJyc5OfnJz8/PzVq1EhHjx61CaBpaWnW3w8cPKjt27frUGqq9fEFx46dCHJ7k+Tq46O071facW3OzNXHW0U5uRXOCw0PV4tmzVS3bl2FhISUC56hoaGqVauWjp0SWuFYCKAA4EAsFoteeeUVDRw4UP369dOECROso7pOmzZNixYt0oIFCxhQ52+8XgAuBYmJibpjyBAlp6bqcFqa9fInZVy9vOT59xFOlzoB8ujaRY0C68gtMFDuZV1yA+vILSBARnGJck66ZJWj8YmKkpPFWQVHjpzoZvt3V9uC9AwVZmTo9/R0bfntVxVmZOh4eoaK8vNtHu/q5qY6QY4zIi/KI4ACgIPp37+/FixYoIcfflhXXXWVdXp0dDSXFKkArxeAi92xY8cUv2mT6g24WY0aN/r7HM86cg8MlFudALl4eVW+MVdX+TeNuXDFVhPP0FB5hoaedbnivLwT4TQ9QwUZ6SpIz1Dm1t+lgwdNqBLnggAKAA6of//+6tu3r9asWaOUlBSFhYWpS5cuHMk7DV4vAJeC+oMGOtRlWByBi5eXXOp7ybt+hHWaV3SUUpYss2NVOBMCKAA4KIvFom7dutm7jBqD1wsAAMdHAAUAAABqgD3v/k+utWrZuwyr0uJiFR/LkYuvj5xdHCdWHD90yN4l4Awc550CAAAAoJymTZuqX//+Stq3T8rMtnc5Vnl5edq5Y4caN20qr6qch3qBlZSUKN3eReC0CKAAAACAA/P29tYXn39u7zLK2bx5s2JjY/XJ3Llq166dvcuxys7Olr+/v73LwGk4n30RAAAAAADOH0dAAQAAUKPl5eUpISGhWtrKz89XUlKSoqKi5Onped7txcTEOFT3VMDeCKAAALty5H8cJf55BGqChIQExcbG2ruMCsXHxztU91TA3gigAAC7cuR/HKXq+efxpZde0pRnnqmmiqpPSUmJJKmTg10z1TAMe5eAGiYmJkbx8fHV0taOHTs0ZMgQzZ07V02bNj3v9mJiYqqhKuDiQQAFANiVI//jKFXPP4/Lvv1WzuFhCu/Tqxoqqj7H0w5p39yPFdK/nzxCgu1djtWBTz6TcnLsXQZqEC8vr2o/yti0aVOOXAIXAAEUAGBXl8o/jt5RkYoeOsTeZdgoyT+u8D695BMVJYunh73LsTq8cpXykpPtXQYA4AIggAIAcImyeHrIvyndAwEA5uEyLAAAAAAAUxBAAQAAAACmoAsuAAAXmJOkQytXaf3Ng+QRGSGvyPryioyUd2R9eUfWl1udOnJycrJ3maYyDEMFhw4rd99+5e7bp7x9+5W3b7+O7z+g7H377F0eAOACIYACAHCBTZ06VZs2bdLOnTuVkJiohFVr9Me+udbLjbh5e8snKlLu9U+EU++/w6lX/fpy9fWxc/XnpzArS3l/h8zcv0Nmwb4Dytm/X0X5+ZIki8Wi+tHRat+kiZrcfLPq1aunhx9+2M6VAwAuBAIoAAAX2JVXXqkrr7zSZtrx48f1559/aufOndbbjsRE7Vz4lfYcOmRdziuwjjzr15dH/Qh5R9aXZ1iYPOvVlZOLY32Elx4v0NFff1XJ8ePKT07R8X37lbd/v44fzbQuExIerpgmjRXTrZsaN25svUVHR8vV1dW6XHZ2NgEUuEDy8vKUkJBQLW3t2LHD5md1iImJkZeXV7W1B8fjWJ9eAABcIjw8PNSsWTM1a9bMZnpRUZG2bt2qJUuWaNGiRdq4caPy0jOkzVvsVOm5c3Z2VufOndWnTx/16tVLzZo1k8VisXdZwCUtISFBsbGx1drmkCHVd4mp+Ph4h7uMFqoXARQAAJMZhqHk5GSbo5+Jfx8B3bd3r0qKiyVJLh4eqt2k8YmuufXryyuyvjxCQ+Tq7S05WJAziop0bPceGUVFyk9OsZ7PuX7jBq1du1aPP/64XN3cFN2woZo2aaImJx0Bbdy4sYKDgy+582AvdWlpacrMzLR3GeUkJSVZf3p7e9u3mJPUqlVLISEh591OTEyM4uPjq6EiKT8/X0lJSYqKipKnp2e1tBkTw6WhLnYEUAAALrBly5bpp59+0s6dO/VHYqL27N6t/NxcSZKTxSLfunXlERkhzw7t1WTQgBPnf0bWl0dwsJyca86A9bVatSw3zSgpUX5K6t/nf544D3Td/v1a+csvyklOtp4H6+Pnp0aNGqlpkyaqX7++2aXDZMnJyYpp1kzHsrLsXcppDRgwwN4l2PD199eO7dtVt27d82rHy8urWo8wdurUqdrawqWBAAoAwAX20ssva8369fJv3kwelzVQ/euuOREyo+rLq25dOZ90/uPFxslikVe9uvKqV1fq1NFmXklBgfL/OmgdCffwvv06kLBDRxcssFO1MMvhw4d1LCtLzZ96Qt4Nou1djo3cpH3aPmWqmk/6j7yjIu1djiQpd2+Stj/9rA4fPnzeARSwNwIoAAAmqNOlk9q8OM3eZTgUi7u7fBo2kE/DBjbTNw67R0e2/GqfomAqv5gm8m/R3N5l2PBv2lR+TWPkExUli6eHvcuRdGJbAS4WBFAAAADgbxZPD/k35TxE4EKpOSeWAAAAAABqNAIoAAAA7KKkoMDeJdQIvE64mNAFFwBQZUVFRdq5c6e9yyhnz5491p/uDnTOVG5uruTno9KiYnuXUiMYpaX2LgEm2XjXvXLz8ZFXeJjcQkLkHhYqz9BQeZz00z0oUM4uF++/rKXFxSo4nK7jKanKT021/ixISVVhWpryklNUmJNj7zKBanPxbs0AgAvmoYce0ptvvmnvMk5r0KBB9i7BlpOTZBhKXf69vSsBHEKTJk30008/af/+/dq/f7/27dunffv3a++ORP21/Htln3R9UCeLRd4hwXIPDZVbaIg8QkPlGRZq/ekZFioXHx/7rcxZFOfkKD8lVfkpqTqe+s/PwtQ0FaSmKjftkIySEuvyfrVqqV5EhFpFRiqyZSvVr19fkZGRql+/PtfIxEWBAAoAqLI9f/6p2u3bqfHYMfYuxUZu0j5tmzhFLZ6e5DCXT5CkHc8+rxj/Who9erS9S7GRlJSkSZMmacqUKYqKirJ3OVb5+fm677777F0GLiAPDw9dddVVuuqqqyqcn52drQMHDlgD6t69e/Xzzz9rw8pVOn78eLnlIwcNVHj/fhe46qrbN/v/lLxsebnpHh4euuKKK3TlkOvVoEED1a9fX/Xr11dERIT8/PzsUClgHgIoAOCcuPn7q3brVvYuw4Zf48bybdzIoS6fIEnutWsrOjJKQ4cOtXcpNjZv3qxJkyapT58+1Xph+vOVnZ1NAL3EZGdnW8Pmybe9fx8ZTT14UCUnHSV09/OTV1ioXP8+IuoVFan1g4fYcQ1OL/T67iotLFBRapryUlJVkJ2t48eP68cff9RPP/2k0Lp1FVm/vqL/Psp56o1AiosNARQAcNHg8glAzbB//37dP3q09u7bpwMHDujYqV1uQ0PkHhIit9BQefa4Tk1O7nYbGlKuy21J/nHVatvG3JWopFO/ECvOyVF+appNd9y9yalK3PGHClb9oNzUNJsuub61aikiIkLRkZF68403FBnpOL07gHNBAAUAnJO8vw4q9fuV8qwbLs+wULn6+8vJycneZTkEwzBUlJml/ORk5aekKD8tTYqMsndZgMM4evSoFi9apNAe1yvs2q5qEBZ2YuChsFC5BwbKyWKpUns16csnFx8f+V7mI9/LGlY43ygpUUF6+olwmpKq/JQUHdmRqD8WLdLUZ54hgKLGI4ACAKqsXt26Klq9Wr8+/Kh1mqu3l7zDw+UWGiqP8DB5hofJMzz8xD+V4eFyC6h90QRUwzBUmJGh/OSUv2/JJ/5ZTE5RUUqqcpOTVZSfb13ey8dHYWFhdqwYcEzRdw6Rf4vm9i7DoThZLPIICZFHSIjUprUkKWv7H0pd/p2dKwOqBwEUAFBl7733nt59912lp6dr3759SkpKsv5M2rdPe7bv0P4ly5R77Jj1MS4eHvIOD7MG1JNDaq0Wzat8xONCKy0qVta2bScCZsqJoFmQkqLC5FTlpqSo+KTr8vnWqqX69eurbXS0otrFKioqSpGRkdafAQEBF034BgDgfBBAAQDnxMnJSUFBQQoKClL79u3LzTcMQ+np6Vq3bp1Wr16t1atXa8uWLTL+3Ftu2cv/945cfLzNKLvSCg6na/MD48pNt1gsateunbp27aquXbuqY8eOqlOnjvkFAjWYm5ubJCn+vjGyONg1Pg3DkFFSIieLxWG+OCopPnEN4bLXDajJHGuLBwDUKIWFhTpw4IDNEdB9+/bpz717tXffvnIjV3oF1pFHWJjcwsLkGX6ia65RUqxf7hlpx7U4s6hhQyXD0PHkZBUmpyg/JVW//PKLfvnlF7388suyuLgovF49RUdFqkFUtM2Rz6ioKNWrV0+urq72Xg3AoTRt2lSffPKJ9u4t/4WUvR08eFBvvvmmRo8erbp169q7HKuoqCg1a9bM3mUA540ACgCosoceekiffPaZDqWkyDAM63Tv4GC5h4XKPTxMntdfpyZ/n//pGR4mz9DQCi+NcmL0yrZmll8lFV3SpTgv/8QIlsnJ1vNAdyYna/uWeOUvXqS8w+nWZZ2dnRUSHq67hw3TM888Y3b5gMMaPHiwvUuo0ObNm/Xmm2/qrrvucqjLEwEXCwIoAKDKEhITlevhoWYTn7Sex+kZGiLnc+geVpNGryzj4uUpnwbR8mkQXeH8koKCE6NX/n3+aNKsOdq1a5fJVQIA4HgIoACAc+IdHamI/v3sXYZDsri7yzsqUt5RJy6XcGj593auCAAAx+Bs7wIAAAAAAJcGAigAAAAAwBR0wQUAnJPCI0eVseEXe5dho6SgUMdTU+URGiqLu+NcrqAwK8veJQAA4BAIoACAKotp0kTLli7VL/eOsncpNUbMoFvtXQIAAHZHAAUAVNnLL7+sBx54wN5llLNt2zb17dtXX331lVq0aGHvcmxERUXZuwQAAOzunALom2++qZdeekmpqalq3bq1Xn/9dV1xxRUVLtutWzetXr263PRevXpp8eLFkiTDMDRp0iS99957yszMVKdOnfT222+rUaNG51IeAOACs1gsatCggb3LKCczM1OSVK9ePYesDwCAS12VA+i8efM0fvx4zZw5Ux06dNCMGTMUFxenxMREBQcHl1t+4cKFKiwstN7PyMhQ69atdcstt1invfjii/rvf/+rOXPmKDo6Wk899ZTi4uL0xx9/yMOj/EXLAQC4VOXl5SkhIaFa2tqxY4fNz+oQExMjLy+vamsPAHBxqXIAnT59ukaMGKHhw4dLkmbOnKnFixfrgw8+0OOPP15u+YCAAJv7n376qby8vKwB1DAMzZgxQ//5z3/Ut29fSdL//d//KSQkRF9++aUGDx5c5ZUCAOBilZCQoNjY2Gptc8iQIdXWVnx8vNq1a1dt7QEALi5VCqCFhYWKj4/XhAkTrNOcnZ3VvXt3rV+/vlJtvP/++xo8eLC8vb0lSXv37lVqaqq6d+9uXcbf318dOnTQ+vXrKwygBQUFKigosN7Pzs6WJBUVFamoqKgqqwSgmtWUbZD9yMWpuLjY+vNi/Ts2bNhQGzZsqJa28vPztW/fPkVGRsrT07Na2mzYsOF5v/Y15W/HfuTidCnsRy52/N0cW5UCaHp6ukpKShQSEmIzPSQkpFLdgTZu3Kht27bp/ffft05LTU21tnFqm2XzTjVt2jRNmTKl3PTly5fT7Qews7y8PHuXUCnsRy5Oe/bskSStXbtWKSkpdq6mZvD391dmZqb1/NnzVR2vO/sRVFVBQYH++uuvammrrJ1PP/1Ua9euPe/26tWrJ3d39/NuB5VXU/YhlyonwzCMyi6cnJysunXrat26derYsaN1+qOPPqrVq1ef9RvZkSNHav369dq6dat12rp169SpUyclJycrLCzMOn3QoEFycnLSvHnzyrVT0TeOERERSk9Pl5+fX2VXB8AFkJ2drcDAQGVlZTn09sh+5OK0ZcsWdejQQRs2bFDbtm3tXQ7OEfsRVFXZtu+I2B+Zr6bsQy5VVToCGhgYKIvForS0NJvpaWlpCg0NPeNjc3Nz9emnn+rpp5+2mV72uLS0NJsAmpaWpjZt2lTYlru7e4XfJLm6usrV1bUyqwLgAqkp2yD7kYuTi4uL9Sd/x5qrpvzt2I84jhYtWig+Pr5a2srPz1dSUpKioqKqpWt6TEwM7weT8Xo7tioFUDc3N8XGxmrFihXq16+fJKm0tFQrVqzQmDFjzvjY+fPnq6CgoNxAB9HR0QoNDdWKFSusgTM7O1sbNmzQqFFc4BwAAABn5uXlVa2DX3Xq1Kna2gJgq8qj4I4fP1533nmn2rdvryuuuEIzZsxQbm6udVTcoUOHqm7dupo2bZrN495//33169dPderUsZnu5OSkcePGaerUqWrUqJH1Mizh4eHWkAsAAAAAqPmqHEBvvfVWHT58WBMnTlRqaqratGmjZcuWWQcR2r9/v5ydnW0ek5iYqLVr12r58uUVtvnoo48qNzdX9957rzIzM9W5c2ctW7aMa4ACAAAAwEWkSoMQOars7Gz5+/tzojHgAGrq9lhT64atzZs3KzY2lmtR1nA1dXusqXUDFxu2RcfmfPZFAAAAAAA4fwRQAAAAAIApCKAAAAAAAFMQQAEAAAAApiCAAgAAAABMQQAFAFwUSkpKtGnTJknSpk2bVFJSYueKAADAqQigAIAab+HChbrssss0cuRISdLIkSN12WWXaeHChXauDAAAnIwACgCo0RYuXKiBAweqZcuWmj17tiRp9uzZatmypQYOHEgIBQDAgbjYuwAAwKUtLy9PCQkJ5/TYkpISPfDAA+rSpYsmTpyoxMRESZKLi4smTpyorKwsjR07VhEREbJYLOf0HDExMfLy8jqnxwIAAFsEUACAXSUkJCg2Nva82khOTtbll19uvT9kyBCb+VdcccU5tx0fH6927dqd8+MBAMA/CKAAALuKiYlRfHz8OT122bJlevLJJ7VmzRp5eXkpPz9fSUlJioqKkqenp3Jzc3X11Vfr2WefVc+ePc+5PgAAUD0IoAAAu/Ly8jrnI4zZ2dmSTnS5LWujU6dO1vnr16+XJF111VUcxQQAwAEwCBEAoMbq0qWLoqKi9Nxzz6m0tNRmXmlpqaZNm6bo6Gh16dLFThUCAICTEUABADWWxWLRK6+8okWLFqlfv35av369jh07pvXr16tfv35atGiRXn755XMegAgAAFQvuuACAGq0/v37a8GCBXr44Yd11VVXWadHR0drwYIF6t+/vx2rAwAAJyOAAgBqvP79+6tv375as2aNUlJSFBYWpi5dunDkEwAAB0MABQBcFCwWi7p162bvMgAAwBlwDigAAAAAwBQEUAAAAACAKQigAAAAAABTEEABAAAAAKYggAIAAAAATEEABQAAAACYggAKAAAAADAFARQAAAAAYAoCKAAAAADAFARQAAAAAIApCKAAAAAAAFMQQAEAAAAApiCAAgAAAABMQQAFAAAAAJiCAAoAAAAAMAUBFAAAAABgCgIoAAAAAMAUBFAAAAAAgCkIoAAAAAAAUxBAAQAAAACmIIACAAAAAExBAAUAAAAAmIIACgAAAAAwBQEUAAAAAGAKAigAAAAAwBQEUAAAAACAKQigAAAAAABTEEABAAAAAKYggAIAAAAATEEABQAAAACYggAKAAAAADAFARQAAAAAYAoCKAAAAADAFARQAAAAAIApCKAAAAAAAFMQQAEAAAAApiCAAgAAAABMQQAFAAAAAJiCAAoAAAAAMAUBFAAAAABgCgIoAAAAAMAUBFAAAAAAgCkIoAAAAAAAUxBAAQAAAACmIIACAAAAAExBAAUAAAAAmIIACgAAAAAwBQEUAAAAAGAKAigAAAAAwBQEUAAAAACAKQigAAAAAABTEEABAAAAAKYggAIAAAAATEEABQAAAACYggAKAAAAADAFARQAAAAAYAoCKAAAAADAFARQAAAAAIApCKAAAAAAAFMQQAEAAAAApiCAAgAAAABMQQAFAAAAAJiCAAoAAAAAMAUBFAAAAABgCgIoAAAAAMAUBFAAAAAAgCkIoAAAAAAAUxBAAQAAAACmIIACAAAAAExBAAUAAAAAmIIACgAAAAAwBQEUAAAAAGCKcwqgb775pqKiouTh4aEOHTpo48aNZ1w+MzNTo0ePVlhYmNzd3dW4cWMtWbLEOn/y5MlycnKyucXExJxLaQAAAAAAB+VS1QfMmzdP48eP18yZM9WhQwfNmDFDcXFxSkxMVHBwcLnlCwsLdf311ys4OFgLFixQ3bp1tW/fPtWqVctmuebNm+v777//pzCXKpcGAAAAAHBgVU5506dP14gRIzR8+HBJ0syZM7V48WJ98MEHevzxx8st/8EHH+jIkSNat26dXF1dJUlRUVHlC3FxUWhoaFXLAQAAAADUEFUKoIWFhYqPj9eECROs05ydndW9e3etX7++wsd8/fXX6tixo0aPHq2vvvpKQUFBuv322/XYY4/JYrFYl9u1a5fCw8Pl4eGhjh07atq0aapfv36FbRYUFKigoMB6Pzs7W5JUVFSkoqKiqqwSgGpWU7ZB9iOA46op2yD7EcAxsf05tioF0PT0dJWUlCgkJMRmekhIiBISEip8zJ9//qmVK1fqjjvu0JIlS7R7927df//9Kioq0qRJkyRJHTp00OzZs9WkSROlpKRoypQp6tKli7Zt2yZfX99ybU6bNk1TpkwpN3358uXy8vKqyioBqGZ5eXn2LqFS2I8Ajov9CIDzUVP2IZcqJ8MwjMounJycrLp162rdunXq2LGjdfqjjz6q1atXa8OGDeUe07hxYx0/flx79+61HvGcPn26XnrpJaWkpFT4PJmZmYqMjNT06dN19913l5tf0TeOERERSk9Pl5+fX2VXB8AFkJ2drcDAQGVlZTn09sh+BHBc7EcAnI+asg+5VFXpCGhgYKAsFovS0tJspqelpZ32/M2wsDC5urradLdt2rSpUlNTVVhYKDc3t3KPqVWrlho3bqzdu3dX2Ka7u7vc3d3LTXd1dbWeZwrAPmrKNsh+BHBcNWUbZD8COCa2P8dWpcuwuLm5KTY2VitWrLBOKy0t1YoVK2yOiJ6sU6dO2r17t0pLS63Tdu7cqbCwsArDpyTl5ORoz549CgsLq0p5AAAAAAAHVuXrgI4fP17vvfee5syZox07dmjUqFHKzc21joo7dOhQm0GKRo0apSNHjujBBx/Uzp07tXjxYj333HMaPXq0dZl///vfWr16tZKSkrRu3TrdfPPNslgsuu2226phFQEAAAAAjqDKl2G59dZbdfjwYU2cOFGpqalq06aNli1bZh2YaP/+/XJ2/ifXRkRE6Ntvv9VDDz2kVq1aqW7dunrwwQf12GOPWZf566+/dNtttykjI0NBQUHq3Lmzfv75ZwUFBVXDKgIAAAAAHEGVBiFyVNnZ2fL39+dEY8AB1NTtsabWDVyMaur2WFPrBi42bIuOrcpdcAEAAAAAOBcEUAAAAACAKQigAAAAAABTEEABAAAAAKYggAIAAAAATEEABQAAAACYggAKAAAAADAFARQAAAAAYAoCKAAAAADAFARQAAAAAIApCKAAAAAAAFMQQAEAAAAApiCAAgAAAABMQQAFAAAAAJiCAAoAAAAAMAUBFAAAAABgCgIoAAAAAMAUBFAAAAAAgCkIoAAAAAAAUxBAAQAAAACmIIACAAAAAExBAAUAAAAAmIIACgAAAAAwBQEUAAAAAGAKAigAAAAAwBQEUAAAAACAKQigAAAAAABTEEABAAAAAKYggAIAAAAATEEABQAAAACYggAKAAAAADAFARQAAAAAYAoCKAAAAADAFARQAAAAAIApCKAAAAAAAFMQQAEAAAAApiCAAgAAAABMQQAFAAAAAJiCAAoAAAAAMAUBFAAAAABgCgIoAAAAAMAUBFAAAAAAgCkIoAAAAAAAUxBAAQAAAACmIIACAAAAAExBAAUAAAAAmIIACgAAAAAwBQEUAAAAAGAKAigAAAAAwBQEUAAAAACAKQigAAAAAABTEEABAAAAAKYggAIAAAAATEEABQAAAACYggAKAAAAADAFARQAAAAAYAoCKAAAAADAFARQAAAAAIApCKAAAAAAAFMQQAEAAAAApiCAAgAAAABMQQAFAAAAAJiCAAoAAAAAMAUBFAAAAABgCgIoAAAAAMAUBFAAAAAAgCkIoAAAAAAAUxBAAQAAAACmIIACAAAAAExBAAUAAAAAmIIACgAAAAAwBQEUAAAAAGAKAigAAAAAwBQEUAAAAACAKQigAAAAAABTEEABAAAAAKYggAIAAAAATEEABQAAAACYggAKAAAAADAFARQAAAAAYAoCKAAAAADAFARQAAAAAIApCKAAAAAAAFMQQAEAAAAApiCAAgAAAABMQQAFAAAAAJiCAAoAAAAAMAUBFAAAAABgCgIoAAAAAMAUBFAAAAAAgCnOKYC++eabioqKkoeHhzp06KCNGzeecfnMzEyNHj1aYWFhcnd3V+PGjbVkyZLzahMAAAAAULNUOYDOmzdP48eP16RJk7R582a1bt1acXFxOnToUIXLFxYW6vrrr1dSUpIWLFigxMREvffee6pbt+45twkAAAAAqHmqHECnT5+uESNGaPjw4WrWrJlmzpwpLy8vffDBBxUu/8EHH+jIkSP68ssv1alTJ0VFRalr165q3br1ObcJAAAAAKh5XKqycGFhoeLj4zVhwgTrNGdnZ3Xv3l3r16+v8DFff/21OnbsqNGjR+urr75SUFCQbr/9dj322GOyWCzn1GZBQYEKCgqs97OzsyVJRUVFKioqqsoqAahmNWUbZD8COK6asg2yHwEcE9ufY6tSAE1PT1dJSYlCQkJspoeEhCghIaHCx/z5559auXKl7rjjDi1ZskS7d+/W/fffr6KiIk2aNOmc2pw2bZqmTJlSbvry5cvl5eVVlVUCUM3y8vLsXUKlsB8BHBf7EQDno6bsQy5VVQqg56K0tFTBwcF69913ZbFYFBsbq4MHD+qll17SpEmTzqnNCRMmaPz48db72dnZioiIUI8ePeTn51ddpQM4B2VHABwd+xHAcbEfAXA+aso+5FJVpQAaGBgoi8WitLQ0m+lpaWkKDQ2t8DFhYWFydXWVxWKxTmvatKlSU1NVWFh4Tm26u7vL3d293HRXV1e5urpWZZUAVLOasg2yHwEcV03ZBtmPAI6J7c+xVWkQIjc3N8XGxmrFihXWaaWlpVqxYoU6duxY4WM6deqk3bt3q7S01Dpt586dCgsLk5ub2zm1CQAAAACoeao8Cu748eP13nvvac6cOdqxY4dGjRql3NxcDR8+XJI0dOhQmwGFRo0apSNHjujBBx/Uzp07tXjxYj333HMaPXp0pdsEAAAAANR8VT4H9NZbb9Xhw4c1ceJEpaamqk2bNlq2bJl1EKH9+/fL2fmfXBsREaFvv/1WDz30kFq1aqW6devqwQcf1GOPPVbpNgEAAAAANZ+TYRiGvYs4X9nZ2fL391dWVhYn/QN2VlO3x5paN3AxqqnbY02tG7jYsC06tip3wQUAAAAA4FwQQAEAAAAApiCAAgAAAABMQQAFAAAAAJiCAAoAAAAAMAUBFAAAAABgCgIoAAAAAMAUBFAAAAAAgCkIoAAAAAAAUxBAAQAAAACmIIACAAAAAExBAAUAAAAAmIIACgAAAAAwBQEUAAAAAGAKAigAAAAAwBQEUAAAAACAKQigAAAAAABTEEABAAAAAKYggAIAAAAATEEABQAAAACYwsXeBVSnooJSFRWU2rsM4JLGNggAAIDTuagC6AeT9srT3dfeZQCXtPyCY/YuAQAAAA6KLrgAAAAAAFNcVEdA75oSLT8/P3uXAVzSsrOz9ejb9q4CAAAAjuiiCqCu7s5ydeegLmBPbIMAAAA4Hf5TBAAAAACYggAKAAAAADAFARQAAAAAYAoCKAAAAADAFARQAAAAAIApCKAAAAAAAFMQQAEAAAAApiCAAgAAAABMQQAFAAAAAJiCAAoAAAAAMAUBFAAAAABgCgIoAAAAAMAUBFAAAAAAgCkIoAAAAAAAUxBAAQAAAACmIIACAAAAAExBAAUAAAAAmIIACgAAAAAwBQEUAAAAAGAKAigAAAAAwBQEUAAAAACAKQigAAAAAABTuNi7AAAAgItJUUGpigpK7V0GcMli+3NsBFAAAIBq9MGkvfJ097V3GcAlK7/gmL1LwBnQBRcAAAAAYAqOgAIAAFSju6ZEy8/Pz95lAJes7OxsPfq2vavA6RBAAQAAqpGru7Nc3elkBtgL259j468DAAAAADAFARQAAAAAYAoCKAAAAADAFARQAAAAAIApCKAAAAAAAFMQQAEAAAAApiCAAgAAAABMQQAFAAAAAJiCAAoAAAAAMAUBFAAAAABgCgIoAAAAAMAUBFAAAAAAgCkIoAAAAAAAUxBAAQAAAACmIIACAAAAAExBAAUAAAAAmIIACgAAAAAwBQEUAAAAAGAKAigAAAAAwBQEUAAAAACAKQigAAAAAABTEEABAAAAAKYggAIAAAAATEEABQAAAACYggAKAAAAADAFARQAAAAAYAoCKAAAAADAFARQAAAAAIApCKAAAAAAAFMQQAEAAAAApiCAAgAAAABMQQAFAAAAAJiCAAoAAAAAMAUBFAAAAABgCgIoAAAAAMAUBFAAAAAAgCkIoAAAAAAAU5xTAH3zzTcVFRUlDw8PdejQQRs3bjztsrNnz5aTk5PNzcPDw2aZYcOGlVumZ8+e51IaAAAAAMBBuVT1AfPmzdP48eM1c+ZMdejQQTNmzFBcXJwSExMVHBxc4WP8/PyUmJhove/k5FRumZ49e2rWrFnW++7u7lUtDQAAAADgwKp8BHT69OkaMWKEhg8frmbNmmnmzJny8vLSBx98cNrHODk5KTQ01HoLCQkpt4y7u7vNMrVr165qaQAAAAAAB1alI6CFhYWKj4/XhAkTrNOcnZ3VvXt3rV+//rSPy8nJUWRkpEpLS9WuXTs999xzat68uc0yq1atUnBwsGrXrq1rr71WU6dOVZ06dSpsr6CgQAUFBdb72dnZkqSioiIVFRVVZZUAVLOasg2yHwEcV03ZBtmPAI6J7c+xVSmApqenq6SkpNwRzJCQECUkJFT4mCZNmuiDDz5Qq1atlJWVpZdffllXXXWVtm/frnr16kk60f22f//+io6O1p49e/TEE0/ohhtu0Pr162WxWMq1OW3aNE2ZMqXc9OXLl8vLy6sqqwSgmuXl5dm7hEphPwI4LvYjAM5HTdmHXKqcDMMwKrtwcnKy6tatq3Xr1qljx47W6Y8++qhWr16tDRs2nLWNoqIiNW3aVLfddpueeeaZCpf5888/1bBhQ33//fe67rrrys2v6BvHiIgIpaeny8/Pr7KrA+ACyM7OVmBgoLKyshx6e2Q/Ajgu9iMAzkdN2Ydcqqp0BDQwMFAWi0VpaWk209PS0hQaGlqpNlxdXdW2bVvt3r37tMs0aNBAgYGB2r17d4UB1N3dvcJBilxdXeXq6lqpOgBcGDVlG2Q/AjiumrINsh8BHBPbn2Or0iBEbm5uio2N1YoVK6zTSktLtWLFCpsjomdSUlKi33//XWFhYadd5q+//lJGRsYZlwEAAAAA1CxVHgV3/Pjxeu+99zRnzhzt2LFDo0aNUm5uroYPHy5JGjp0qM0gRU8//bSWL1+uP//8U5s3b9aQIUO0b98+3XPPPZJODFD0yCOP6Oeff1ZSUpJWrFihvn376rLLLlNcXFw1rSYAAAAAwN6qfB3QW2+9VYcPH9bEiROVmpqqNm3aaNmyZdaBifbv3y9n539y7dGjRzVixAilpqaqdu3aio2N1bp169SsWTNJksVi0datWzVnzhxlZmYqPDxcPXr00DPPPMO1QAEAAADgIlKlQYgcVXZ2tvz9/TnRGHAANXV7rKl1Axejmro91tS6gYsN26Jjq3IXXAAAAAAAzgUBFAAAAABgCgIoAAAAAMAUBFAAAAAAgCkIoAAAAAAAUxBAAQAAAACmIIACAAAAAExBAAUAAAAAmIIACgAAAAAwBQEUAAAAAGAKAigAAAAAwBQEUAAAAACAKQigAAAAAABTEEABAAAAAKYggAIAAAAATEEABQAAAACYggAKAAAAADAFARQAAAAAYAoCKAAAAADAFARQAAAAAIApCKAAAAAAAFMQQAEAAAAApiCAAgAAAABMQQAFAAAAAJiCAAoAAAAAMAUBFAAAAABgCgIoAAAAAMAUBFAAAAAAgCkIoAAAAAAAUxBAAQAAAACmIIACAAAAAExBAAUAAAAAmIIACgAAAAAwBQEUAAAAwP+3d/cxVZd9HMc/hw4ebkQUYaIogVMINKP6x8oalhmWD2whiYZpZ8P1jFt2r6GObtIZRtJqTVnFQ1uLZotGT/ZgUhmr1rJiDqdrspyCKWMDSnmQ6/7DgZwOKCfwx+H83q/t/OHv4eL6Xqfrs31/nEOAJWhAAQAAAACWoAEFAAAAAFiCBhQAAAAAYAkaUAAAAACAJWhAAQAAAACWoAEFAAAAAFiCBhQAAAAAYAkaUAAAAACAJZyjPQEA8CddHT3q6ugZ7WkAtsYeBIDARQMKAP2U5h/Xf1wTRnsagK2d62gb7SkAAK4SPoILAAAAALAEvwEFgH7c/5up8PDw0Z4GYGutra367+7RngUA4GqgAQWAfoJdQQp28eEQYDSxBwEgcJHwAAAAAABL0IACAAAAACxBAwoAAAAAsAQNKAAAAADAEjSgAAAAAABL0IACAAAAACxBAwoAAAAAsAQNKAAAAADAEjSgAAAAAABL0IACAAAAACxBAwoAAAAAsAQNKAAAAADAEjSgAAAAAABL0IACAAAAACxBAwoAAAAAsAQNKAAAAADAEjSgAAAAAABL0IACAAAAACxBAwoAAAAAsAQNKAAAAADAEs7RnsBIMMZIklpbW0d5JgB692HvvhwryBHAf5AjAIZjrGaIXQREA9rW1iZJio2NHeWZAOjV1tamiRMnjvY0howcAfwPOQJgOMZahtiFwwTAo4Genh6dOnVKEyZMkMPhGO3p9GltbVVsbKxOnDih8PDw0Z6O32O9fOOv62WMUVtbm2JiYhQUNHY+5U+OBAbWyzf+ul7kyMjy1/fZX7FeQ+evazVWM8QuAuI3oEFBQZoxY8ZoT2NQ4eHhfrUp/R3r5Rt/XK+x+LSRHAksrJdv/HG9yJGR54/vsz9jvYbOH9dqLGaIXfBIAAAAAABgCRpQAAAAAIAlaECvIpfLpfz8fLlcrtGeypjAevmG9bIH3mffsF6+Yb3sgffZN6zX0LFW+DcC4o8QAQAAAAD8H78BBQAAAABYggYUAAAAAGAJGlAAAAAAgCVoQAEAAAAAlqABBQAAAABYggZ0iNavXy+HwyGHw6Hg4GBFR0dr8eLFKi0tVU9PT991v/76q1asWKEpU6YoJCRE8fHxWrVqlf78809JUkNDgxwOh6655hqdPHnS42c0NjbK6XTK4XCooaHByvJ8dubMGT366KO69tpr5XK5NHXqVKWlpem7776TJMXHx8vhcKiystLr3rlz58rhcKi8vNzjeG1tre677z5FREQoJCRE8+bN065du3ThwgVJUnl5ed97MNiroaFBzz333IDnkpKSqPsyFi5cOODcX375ZcXHx3sc6+zs1M6dO5WSkqLQ0FBFRUVpwYIFKisrU1dX1xBX037IkUsCeS/ZuXZy5OoiQzwF8l6ya91kiD3QgPpgyZIlamxsVENDgz799FPdeeedys3N1bJly9Td3a0zZ85o0aJFmjx5sj777DPV19errKxMMTEx+uuvvzzGmj59ut566y2PYxUVFZo+fbqVJf1rGRkZOnTokCoqKnT06FFVV1dr4cKFam5u7rsmNjZWZWVlHvd9//33ampq0vjx4z2OV1VVKTU1VTNmzNCBAwd05MgR5ebmatu2bcrKypIxRqtWrVJjY2Pf69Zbb1VOTo7HsdjYWEkXA7b/8cbGRh08eJC6ryAkJERbtmy5bHB3dnYqLS1NL7zwgjZs2KDa2lr9+OOPevzxx/Xqq6/q8OHDQ/55dkSOXBToe8nOtZMjVxcZckmg7yW71k2G2IDBkKxbt86kp6d7Hd+/f7+RZF5//XVTVVVlnE6n6erqGnSc48ePG0lmy5YtJiEhweNcYmKi2bp1q5Fkjh8/PsIVjJyWlhYjydTU1Ax6TVxcnHn22WeNy+Uyf/zxR9/xnJwc8+STT5qJEyeasrIyY4wx7e3tJjIy0tx///1e41RXVxtJprKy0utcamqqyc3N9Tqen59vUlJSfK7rSgK97tTUVPPwww+byMhI89prr/UdLy4uNnFxcX3/LiwsNEFBQebnn3/2GqOzs9O0t7f/6zkEOnLkokDfS5cT6LWTI1cXGXJJoO+lwQR63WSIPfAb0GG66667lJKSovfff19Tp05Vd3e3qqqqZIy57H0rVqxQS0tL3xOhgwcPqqWlRcuXL7di2sMSFhamsLAwffDBB+ro6Bj0uujoaKWlpamiokKS9Pfff+vdd9+V2+32uO7zzz9Xc3OzNm3a5DXG8uXLlZiYqHfeeWdki/gX7FB3eHi4Nm/erIKCAq8n5b3efvtt3X333brpppu8zgUHB3s9WcWV2S1H7LCXBmOH2skR69ktQyR77KWB2KFuMiTw0YCOgKSkJDU0NOiWW25RXl6e1qxZo6ioKN1777168cUXdfr0aa97goODlZ2drdLSUklSaWmpsrOzFRwcbPX0feZ0OlVeXq6KigpNmjRJCxYsUF5enn777Teva91ut8rLy2WM0XvvvadZs2bpxhtv9Ljm6NGjkqTk5OQBf15SUlLfNUNVV1fXF9K9r0ceecSnMf7JLnU/9thjCgkJ0a5duwY8f+zYsRH7LhwusVOO2GUvDcQutZMj1rNThkj22Uv/ZJe6yZDARgM6AowxcjgckqTt27erqalJe/bs0dy5c7Vnzx4lJSWprq7O6z632629e/eqqalJe/fu9Xoq5c8yMjJ06tQpVVdXa8mSJaqpqdHNN9/s9aX2pUuXqr29Xd98841KS0svW+OVntT64rrrrtMvv/zi8SooKBj2uHao2+VyqaCgQEVFRTp79qzX+ZGcLy6xW47YYS8Nxg61kyPWs1uGSPbYSwOxQ91kSGCjAR0B9fX1mjlzZt+/IyMjlZmZqaKiItXX1ysmJkZFRUVe982bN09JSUlavXq1kpOTdf3111s57WELCQnR4sWLtXXrVtXW1mr9+vXKz8/3uMbpdGrt2rXKz8/XDz/8oAcffNBrnMTEREkX13Eg9fX1fdcM1bhx4zR79myP15QpU3waYzB2qDs7O1txcXHatm3bgPM+cuSIz2Pi8uyYI3bYS4OxQ+3kiLXsmCGSPfbSQOxQNxkSuGhAh+mrr75SXV2dMjIyBjw/btw4zZo1a9DPsLvdbtXU1IypJ46DmTNnzoB1ut1uff3110pPT1dERITX+XvuuUeTJ0/WSy+95HWuurpax44d0+rVq6/KnEdCINYdFBSkHTt2aPfu3V5/hn/NmjX68ssvdejQIa/7urq6Bv1vHYMjRy4KxL00VIFYOzliHTLkkkDcS0MRiHWTIYHLOdoTGEs6OjrU1NSkCxcu6PTp09q3b5927NihZcuW6aGHHtJHH32kyspKZWVlKTExUcYYffjhh/rkk0+8/hR2r5ycHGVmZmrSpEnWFjMMzc3NyszMlNvt1g033KAJEybop59+0s6dO5Wenu51fXJyss6ePavQ0NABxxs/frxKSkqUlZWlDRs26IknnlB4eLj279+vZ555RitXrtQDDzzg0xy7u7vV1NTkcczhcCg6OtqncfqzW91Lly7V/PnzVVJS4nH/xo0b9fHHH2vRokV6/vnndfvtt/etRWFhod58802v75jgEnLEfnupP7vVTo6MPDLkIrvtpV52q5sMCUw0oD7Yt2+fpk2bJqfTqYiICKWkpOiVV17RunXrFBQUpDlz5ig0NFRPP/20Tpw4IZfLpYSEBL3xxhtau3btgGM6nU5FRUVZXMnwhIWFaf78+SouLtbvv/+urq4uxcbGKicnR3l5eQPeExkZedkxV65cqQMHDmj79u264447dP78eSUkJGjz5s3auHFj3/dahurw4cOaNm2axzGXy6Xz58/7NE5/dqy7sLBQt912m9d4X3zxhYqLi1VSUqJNmzYpNDRUycnJeuqpp8bcx7esRo7Ycy/1smPt5MjIIkMusuNekuxZNxkSeByGb/ECAAAAACzAd0ABAAAAAJagAQVs4Ntvv/X6f3L1fwHAlZAjAIaDDEEvPoIL2MC5c+d08uTJQc/Pnj3bwtkAGIvIEQDDQYagFw0oAAAAAMASfAQXAAAAAGAJGlAAAAAAgCVoQAEAAAAAlqABBQAAAABYggYUAAAAAGAJGlAAAAAAgCVoQAEAAAAAlvg/WfUnTy5+//QAAAAASUVORK5CYII=",
      "text/plain": [
       "<Figure size 1000x700 with 3 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "#distributions of F1 scores for each model by augmentation method\n",
    "f_1_fig, (ax1, ax2, ax3) = plt.subplots(1,3, figsize=(10,7), sharey=True)\n",
    "f_1_fig.suptitle('Distribution of F1 scores')\n",
    "\n",
    "log_reg_box = [DSM_scores['log_reg_f_1'], SMOTE_scores['log_reg_f_1'], SMOTE_NC_scores['log_reg_f_1']]\n",
    "xgb_box = [DSM_scores['xgb_f_1'], SMOTE_scores['xgb_f_1'], SMOTE_NC_scores['xgb_f_1']]\n",
    "catboost_box= [DSM_scores['catboost_f_1'], SMOTE_scores['catboost_f_1'], SMOTE_NC_scores['catboost_f_1']]\n",
    "\n",
    "bplot1 = ax1.boxplot(log_reg_box, patch_artist=True, notch=True, widths=1, medianprops={'color':'black'})\n",
    "#set boxplot color\n",
    "for patch in bplot1['boxes']:\n",
    "    patch.set_facecolor('mediumturquoise')\n",
    "#other graph settings\n",
    "ax1.set_xticks([1,2,3], ['DSM', 'SMOTE', 'SMOTE_NC'])\n",
    "ax1.set_title('Logistic Regression F1 Scores')\n",
    "ax1.grid(axis='y')\n",
    "ax1.axhline(f1_score(y_test_encoded, unaug_log_reg_pred), color='mediumpurple', label='Unaugmented score')\n",
    "\n",
    "bplot2 = ax2.boxplot(xgb_box, patch_artist=True, notch=True, widths=1, medianprops={'color':'black'})\n",
    "#set boxplot color\n",
    "for patch in bplot2['boxes']:\n",
    "    patch.set_facecolor('mediumturquoise')\n",
    "#other graph settings\n",
    "ax2.set_xticks([1,2,3], ['DSM', 'SMOTE', 'SMOTE_NC'])\n",
    "ax2.set_title('XGBoost F1 Scores')\n",
    "ax2.grid(axis='y')\n",
    "ax2.axhline(f1_score(y_test_encoded, unaug_xgb_pred), color='mediumpurple')\n",
    "\n",
    "bplot3 = ax3.boxplot(catboost_box, patch_artist=True, notch=True, widths=1, medianprops={'color':'black'})\n",
    "#set boxplot color\n",
    "for patch in bplot3['boxes']:\n",
    "    patch.set_facecolor('mediumturquoise')\n",
    "#other graph settings\n",
    "ax3.set_xticks([1,2,3], ['DSM', 'SMOTE', 'SMOTE_NC'])\n",
    "ax3.set_title('Catboost F1 Scores')\n",
    "ax3.grid(axis='y')\n",
    "ax3.axhline(f1_score(y_test_encoded, unaug_catboost_pred), color='mediumpurple')\n",
    "\n",
    "f_1_fig.legend()\n",
    "\n",
    "plt.savefig('f_1_box_accomp.jpeg')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "a61f1260-d505-42e4-808f-dccd666d9fc2",
   "metadata": {},
   "source": [
    "### Looking at Catboost with DSM"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "id": "e37a329d-7b2a-493c-9374-f41c687a24b2",
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "#running a CatBoost model with DSM\n",
    "#pulling in 100,000 synthetic data points from DSM\n",
    "sample = model.sample(100000)\n",
    "churn_samples = sample[sample['Churn'] == 'Yes']\n",
    "\n",
    "    #pulling out the number of minority class samples needed\n",
    "new_samples = churn_samples[:n_samples_needed]\n",
    "original_samples = data_train[modeling_columns].copy()\n",
    "    #adding our synthetic data to our original training data\n",
    "new_model_train_data = pd.concat([original_samples, new_samples], axis=0).reset_index(drop=True)\n",
    "\n",
    "    #encoding augmented data\n",
    "y_aug_train_encoded = new_model_train_data[target].apply(lambda x: int(x == 'Yes'))\n",
    "X_train_aug = new_model_train_data[input_columns]\n",
    "\n",
    "    #model\n",
    "DSM_catboost = catboost.fit(X_train_aug, y_aug_train_encoded, cat_features=[col for col in categorical_columns if col!=target])\n",
    "      #y proba predictions\n",
    "DSM_catboost_pred_proba = unaug_catboost.predict_proba(X_test)[:,1]\n",
    "      #y predictions\n",
    "DSM_catboost_pred = unaug_catboost.predict(X_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "id": "6cf3c1f5-eff5-41db-badd-701a00910465",
   "metadata": {
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1300x400 with 4 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Building confusion matrices sub-plots\n",
    "\n",
    "catboost_fig, axes= plt.subplots(1,2, figsize=(13,4))\n",
    "axes = axes.flatten()\n",
    "catboost_fig.suptitle('Catboost Confusion Matrices - Unaugmented and DSM \\n')\n",
    "\n",
    "#unaugmented\n",
    "cf_matrix_plot(y_test_encoded, unaug_catboost_pred, axes[0])\n",
    "axes[0].set_title(f\"\\n Unaugmented - F1: {round(f1_score(y_test_encoded, unaug_catboost_pred) , 6)}\")\n",
    "axes[0].set_ylabel('True values')\n",
    "axes[0].set_xlabel('Predicted values')\n",
    "\n",
    "#DSM\n",
    "cf_matrix_plot(y_test_encoded, DSM_catboost_pred, axes[1])\n",
    "axes[1].set_title(f\"DSM - F1: {round(f1_score(y_test_encoded, DSM_catboost_pred) , 6)}\")\n",
    "axes[1].set_ylabel('True values')\n",
    "axes[1].set_xlabel('Predicted values')\n",
    "\n",
    "unaug_fig.tight_layout()\n",
    "plt.savefig('unaug_DSM_confusion_matrices_accomp.jpeg')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "f969d271-fbb6-4fd7-a47e-4a30ebcca701",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.10.10"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}