lectures/10_week.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {
    "slideshow": {
     "slide_type": "skip"
    }
   },
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "import pandas as pd\n",
    "import seaborn as sns\n",
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {
    "slideshow": {
     "slide_type": "skip"
    }
   },
   "outputs": [],
   "source": [
    "df = sns.load_dataset('iris')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "## COMP 3122 - Artificial Intelligence with Python\n",
    "__Week 10 lecture__\n",
    "\n",
    "### [github.com/kamrik/ML1](https://github.com/kamrik/ML1)\n",
    "\n",
    "### [slido.com/COMP3122](http://slido.com/COMP3122)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "## The plan for today\n",
    " - Model evaluation & train/test splitting \n",
    "   - Videos 5-7 from [Kevin Markham's playlist](https://www.dataschool.io/machine-learning-with-scikit-learn/)\n",
    " - Guest lecture (2nd hour)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "## How good is your model?"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "source": [
    "## Good for what?"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "## Visual estimates"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "metadata": {
    "slideshow": {
     "slide_type": "skip"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "LinearRegression(copy_X=True, fit_intercept=True, n_jobs=1, normalize=False)"
      ]
     },
     "execution_count": 56,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.linear_model import LinearRegression\n",
    "lm = LinearRegression()\n",
    "lm.fit(df[['petal_length']], df.petal_width)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "metadata": {
    "slideshow": {
     "slide_type": "-"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[<matplotlib.lines.Line2D at 0x164822310b8>]"
      ]
     },
     "execution_count": 63,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.scatter(df.petal_length, df.petal_width)\n",
    "x = np.arange(1,8)\n",
    "plt.plot(x, x*lm.coef_  + lm.intercept_ , 'r')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 71,
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[<matplotlib.lines.Line2D at 0x16482231668>]"
      ]
     },
     "execution_count": 71,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 444.75x360 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "sns.relplot('petal_length', 'petal_width', hue='species', data=df[df.species != 'setosa'])\n",
    "miss = df[(df.species=='virginica') & (df.petal_width<1.8)]\n",
    "plt.plot(miss.petal_length, miss.petal_width, 'xk', markersize=10)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "## But we want to quantify how good it is\n",
    " - What measures of model quality have we already talked about?"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "## We still need to know actual labels to calculated accuracy score or MSE\n",
    " - Where do we get data with labels for that?"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "## Option 1: Use the same data we used for training\n",
    " - Training accuracy"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "## Overfitting\n",
    " - Overfitting is like learning through memorization. Rather than understanding the concepts at play and making generalizations you simply memorize what you’ve seen.\n",
    " - We want the model (or student) to internalize the essence and to **generalize** from examples, not memorize them\n",
    " \n",
    " - Some good discussion [on Quora](https://www.quora.com/What-is-overfitting)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "## kNN with k=1 example"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "## Train / test split\n",
    "![split](https://raw.githubusercontent.com/justmarkham/scikit-learn-videos/master/images/05_train_test_split.png)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "## How to split?\n",
    " - Let's try to use 1/3 of the iris data for training"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 73,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(150, 5)"
      ]
     },
     "execution_count": 73,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 74,
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "outputs": [],
   "source": [
    "from sklearn.model_selection import train_test_split"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 81,
   "metadata": {},
   "outputs": [],
   "source": [
    "X = df[['sepal_length', 'sepal_width', 'petal_length', 'petal_width']]\n",
    "y = df.species"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 85,
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "outputs": [],
   "source": [
    "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.67)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 88,
   "metadata": {
    "slideshow": {
     "slide_type": "-"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(49,)"
      ]
     },
     "execution_count": 88,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_train.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "## Random state / seed"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 90,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([0.03805366, 0.75046049, 0.35330927])"
      ]
     },
     "execution_count": 90,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.random.rand(3)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "outputs": [],
   "source": [
    "rng1 = np.random.RandomState(seed=42)\n",
    "rng2 = np.random.RandomState(seed=42)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [],
   "source": [
    "a = rng1.rand(10)\n",
    "b = rng2.rand(10)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[0.37454012 0.95071431 0.73199394 0.59865848 0.15601864 0.15599452\n",
      " 0.05808361 0.86617615 0.60111501 0.70807258]\n",
      "[0.37454012 0.95071431 0.73199394 0.59865848 0.15601864 0.15599452\n",
      " 0.05808361 0.86617615 0.60111501 0.70807258]\n"
     ]
    }
   ],
   "source": [
    "print(a)\n",
    "print(b)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "![overfitting](https://raw.githubusercontent.com/justmarkham/scikit-learn-videos/master/images/05_overfitting.png)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "celltoolbar": "Slideshow",
  "kernelspec": {
   "display_name": "Python 3",
   "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": "3.6.6"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}