12-EX-Classifier-Performance-Measurement-v1.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Creative Commons CC BY 4.0 Lynd Bacon & Associates, Ltd. Not warranted to be suitable for any particular purpose. (You're on your own!)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Binary Classifier Performance Measurement\n",
    "\n",
    "The classification of observations into just two categories is a special case of the broader array of applications that involve predicting three or more ordered or unordered target variable labels.  And, there are many different performance metrics used in several different disciplines."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Confusion Matrix\n",
    "\n",
    "Many frequently used binary classifier performance metrics are based on the [confusion matrix](https://en.wikipedia.org/wiki/Confusion_matrix), which is basically just a cross-tabulation of true vs. predicted class labels.\n",
    "\n",
    "Other common measures include the _Area Under the Curve_ (AUC), where the curve is a _Receiver Operating Characteristic_ (ROC) curve.\n",
    "\n",
    "Let's get the data and results from the binary logistic regression notebook, and apply some performance metrics to it.\n",
    "\n",
    "But first we need to get lots of stuff in modules:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "from IPython.core.interactiveshell import InteractiveShell\n",
    "InteractiveShell.ast_node_interactivity = \"all\"\n",
    "from IPython.display import display, Markdown\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "\n",
    "import matplotlib.pyplot as plt\n",
    "import scikitplot as skplt\n",
    "from sklearn import linear_model\n",
    "from sklearn.model_selection import cross_val_predict\n",
    "from sklearn import preprocessing\n",
    "from sklearn.metrics import accuracy_score, confusion_matrix, roc_auc_score\n",
    "from sklearn.metrics import roc_curve, precision_score, recall_score\n",
    "import pickle\n",
    "import os\n",
    "import shelve\n",
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#  Unpickling the Data\n",
    "\n",
    "This assumes that the pickle file is in the DATA subdir."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "dict_keys(['X_train', 'y_train', 'X_test', 'y_test', 'yTestPredLabels', 'yTestPredProbs', 'yTrainPredLabels', 'yTrainPredProbs'])"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "with open('DATA/ML/rBinLogData2.pkl','rb') as pickleInFile:\n",
    "    breastResDict=pickle.load(pickleInFile)\n",
    "breastResDict.keys()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Basic Confusion\n",
    "\n",
    "A _confusion matrix_ is basically just a two-way table tabulating comments of \"true\" classifier label values by \"predicted\" labels."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[260,   7],\n",
       "       [  9, 150]])"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "## Training Data Confusion Matrix\n",
    "\n",
    "confusion_matrix(breastResDict['y_train'], breastResDict['yTrainPredLabels'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[84,  6],\n",
       "       [ 3, 50]])"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Test Data Confusion Matrix\n",
    "\n",
    "confusion_matrix(breastResDict['y_test'], breastResDict['yTestPredLabels'])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Kinder Confusion"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#Let's import a script for a \"better looking\" confusion matrix from [Matt Hancock](https://notmatthancock.github.io/2015/10/28/confusion-matrix.html), It's in the file `show-confusion.py`.  \n",
    "\n",
    "Assuming that it's in the pwd:\n",
    " "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "from show_confusion import show_confusion_matrix"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "    C: ndarray, shape (2,2) as given by scikit-learn confusion_matrix function\n",
      "    class_labels: list of strings, default simply labels 0 and 1.\n",
      "\n",
      "    Draws confusion matrix with associated metrics.\n",
      "    \n"
     ]
    }
   ],
   "source": [
    "print(show_confusion_matrix.__doc__)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "collapsed": true
   },
   "outputs": [
    {
     "data": {
      "text/markdown": [
       "<strong>Training Data Confusion Matrix</strong><br/>"
      ],
      "text/plain": [
       "<IPython.core.display.Markdown object>"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x576 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "Markdown('<strong>{}</strong><br/>'.format(\"Training Data Confusion Matrix\"))\n",
    "\n",
    "show_confusion_matrix(confusion_matrix(breastResDict['y_train'], \n",
    "                                       breastResDict['yTrainPredLabels']))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "collapsed": true
   },
   "outputs": [
    {
     "data": {
      "text/markdown": [
       "<strong>Test Data Confusion Matrix</strong><br/>"
      ],
      "text/plain": [
       "<IPython.core.display.Markdown object>"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x576 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "Markdown('<strong>{}</strong><br/>'.format(\"Test Data Confusion Matrix\"))\n",
    "\n",
    "show_confusion_matrix(confusion_matrix(breastResDict['y_test'], \n",
    "                                       breastResDict['yTestPredLabels']))\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# A Few Useful Binary Classifier Scores  \n",
    "\n",
    "Scikit-learn provide a number of [metrics for classifiers](https://scikit-learn.org/stable/modules/model_evaluation.html#classification-metrics). Among these are:\n",
    "\n",
    "* AUC - area under the ROC (\"receiver operating characteristics\") curve.\n",
    "* Precision (AKA positive predictive value).\n",
    "* Recall (AKA sensitivity).\n",
    "* F1 (harmonic mean of Precision and Recall).\n",
    "* Accuracy (proportion correct).\n",
    "* Hit Rate (AKA sensitivity).\n",
    "* False Alarm Rate (1-selectivity).\n",
    "\n",
    "This page on [wikipedia about Precision and Recall](https://en.wikipedia.org/wiki/Precision_and_recall) includes a table that's helpful in disambiguating the terminology above.\n",
    "\n",
    "The Hit Rate (HR) and the False Alarm Rate (FAR) are used to plot an ROC curve."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# ROC Curve and AUC\n",
    "\n",
    "Let's first get the data out of the dict breastResDict to make things more convenient."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "y_train=breastResDict['y_train']\n",
    "y_test=breastResDict['y_test']\n",
    "yTrainPredProbs=breastResDict['yTrainPredProbs']\n",
    "yTestPredProbs=breastResDict['yTestPredProbs']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 131,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(426,)"
      ]
     },
     "execution_count": 131,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "text/plain": [
       "(143,)"
      ]
     },
     "execution_count": 131,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Just checking\n",
    "\n",
    "yTrainPredProbs.shape\n",
    "yTestPredProbs.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## ROC Plot and AUC for the Training Data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Plotting using maplotlib, a \"reusable\" ROC plotting fcn\n",
    "\n",
    "def plotROC(fpr,tpr,label=None):\n",
    "    plt.subplots(figsize=(6,6))\n",
    "    plt.plot(fpr,tpr, linewidth=3,label=label)\n",
    "    plt.plot([0,1],[0,1], 'k--')\n",
    "    plt.axis([0,1,0,1])\n",
    "    plt.xlabel('False Pos Rate')\n",
    "    plt.ylabel('True Pos Rate')\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {
    "collapsed": true
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# For the Training data results\n",
    "\n",
    "fprTrain, tprTrain, _ = roc_curve(y_train, yTrainPredProbs)\n",
    "\n",
    "plotROC(fprTrain, tprTrain)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training AUC: 0.995\n"
     ]
    }
   ],
   "source": [
    "trainingAUC=roc_auc_score(y_train, yTrainPredProbs)\n",
    "print('Training AUC: {0:4.3f}'.format(trainingAUC))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {
    "collapsed": true
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# For the Test data results\n",
    "\n",
    "fprTest, tprTest, _ = roc_curve(y_test, yTestPredProbs)\n",
    "\n",
    "plotROC(fprTest, tprTest)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Test AUC: 0.987\n"
     ]
    }
   ],
   "source": [
    "testAUC=roc_auc_score(y_test, yTestPredProbs)\n",
    "print('Test AUC: {0:4.3f}'.format(testAUC))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Zounds!  What a Fit!\n",
    "\n",
    "This model works too well.  Suspiciously so, in fact.  To get a sense of what typical ROC's look like, do the following UDU."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# UDU: Train a Binary Logistic Classifier Using Features of Your Choice\n",
    "\n",
    "Pick out just a couple of features, and train a binary logistic classifier using them. Use a hold out validation sample as your test data.  Then, plot ROCs for results based on your training data and your test data.\n",
    "\n",
    "The scikit-learn method [model_selection.cross_val_precict()](https://scikit-learn.org/stable/modules/generated/sklearn.model_selection.cross_val_predict.html) can be used to \"quickly\" get predictions from a cross-validated model. Give it a try, here."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Alternative Ways to Do an ROC Plot  \n",
    "\n",
    "There's more than one way to do it, even if it isn't in Perl.  The [scikit-plot](https://scikit-plot.readthedocs.io/en/stable/) package has some clever ways of producing plot to assess classifiers and other kinds of models.  Here's a set-up for using it with the logistic results, above.\n",
    "\n",
    "`scikit-plot` likes a 2D numpy array for the predicted label probabilities:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [],
   "source": [
    "yTestProbs=np.c_[1-yTestPredProbs,yTestPredProbs]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Then we can do:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {
    "collapsed": true
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 504x504 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "skplt.metrics.plot_roc(y_test, yTestProbs, title=\"Test Data ROC\",\n",
    "                      figsize=(7,7),\n",
    "                      plot_micro=False,\n",
    "                      classes_to_plot=1)\n",
    "plt.show();"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Other ways to produce ROC plots include using the Python [ggplot](http://ggplot.yhathq.com/) package, and the [seaborn](https://seaborn.pydata.org/) package. "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Precision and Recall\n",
    "\n",
    "There is an opinion that Precision and Recall are more useful than the ROC curve and AUC for evaluating classifiers where the distribution of the target variable's values is \"unbalanced,\" or skewed. \n",
    "\n",
    "Some definitions:\n",
    "\n",
    "_Precision_ = proportion of cases classified as positive that are actually positive  \n",
    "_Recall_ = proportion of positive cases that are classified as positive  \n",
    "\n",
    "Let:\n",
    "\n",
    "TP=true positive, FP=false positive, TN=true negative, and FN=false negative.\n",
    "\n",
    "Then:\n",
    "\n",
    "_Precision_  =  TP/(TP+FP)  \n",
    "_Recall_     =  TP/(TP+FN)\n",
    "\n",
    "\n",
    "Here's a [PLOS One Paper](https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4349800/) about Precision-Recall plots vs ROC plots when target categories are unbalanced.\n",
    "\n",
    "Precision and recall for the present data are as follows. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Precision: 0.955 Training Recall: 0.943\n"
     ]
    }
   ],
   "source": [
    "print('Training Precision: {:5.3f} Training Recall: {:5.3f}'.format(\n",
    "        precision_score(y_train,yTrainPredProbs>0.5),\n",
    "        recall_score(y_train,yTrainPredProbs>0.5)))"
   ]
  }
 ],
 "metadata": {
  "hide_input": false,
  "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": "3.6.8"
  },
  "latex_envs": {
   "LaTeX_envs_menu_present": true,
   "autoclose": false,
   "autocomplete": true,
   "bibliofile": "biblio.bib",
   "cite_by": "apalike",
   "current_citInitial": 1,
   "eqLabelWithNumbers": true,
   "eqNumInitial": 1,
   "hotkeys": {
    "equation": "Ctrl-E",
    "itemize": "Ctrl-I"
   },
   "labels_anchors": false,
   "latex_user_defs": false,
   "report_style_numbering": false,
   "user_envs_cfg": false
  },
  "toc": {
   "base_numbering": 1,
   "nav_menu": {},
   "number_sections": false,
   "sideBar": false,
   "skip_h1_title": false,
   "title_cell": "Table of Contents",
   "title_sidebar": "Contents",
   "toc_cell": false,
   "toc_position": {
    "height": "calc(100% - 180px)",
    "left": "10px",
    "top": "150px",
    "width": "330.2px"
   },
   "toc_section_display": false,
   "toc_window_display": false
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}