{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "import sklearn.metrics"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Read in labels and performance data:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "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>Pneumonia</th>\n",
       "      <th>Atelectasis</th>\n",
       "      <th>Effusion</th>\n",
       "      <th>Pneumothorax</th>\n",
       "      <th>Infiltration</th>\n",
       "      <th>Cardiomegaly</th>\n",
       "      <th>Mass</th>\n",
       "      <th>Nodule</th>\n",
       "      <th>algorithm_output</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   Pneumonia  Atelectasis  Effusion  Pneumothorax  Infiltration  Cardiomegaly  \\\n",
       "0          1            1         1             0             0             0   \n",
       "1          1            1         0             0             0             1   \n",
       "2          1            0         1             0             0             0   \n",
       "3          1            1         1             0             0             1   \n",
       "4          0            1         0             0             0             0   \n",
       "\n",
       "   Mass  Nodule  algorithm_output  \n",
       "0     0       0                 1  \n",
       "1     0       0                 1  \n",
       "2     0       0                 1  \n",
       "3     0       0                 1  \n",
       "4     0       0                 0  "
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data = pd.read_csv('labels_and_performance.csv')\n",
    "data.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "First, look at the overall performance of the algorithm for the detection of pneumonia:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "tn, fp, fn, tp = sklearn.metrics.confusion_matrix(data.Pneumonia.values,\n",
    "                                                  data.algorithm_output.values,labels=[1,0]).ravel()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.8235294117647058"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sens = tp/(tp+fn)\n",
    "sens"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.8166666666666667"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "spec = tn/(tn+fp)\n",
    "spec"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Now, look at the algorithm's performance in the presence of the other diseases: "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Atelectasis\n",
      "Sensitivity: 0.8333333333333334\n",
      "Specificity: 0.782608695652174\n",
      "\n",
      "Effusion\n",
      "Sensitivity: 0.8571428571428571\n",
      "Specificity: 0.6521739130434783\n",
      "\n",
      "Pneumothorax\n",
      "Sensitivity: 0.6666666666666666\n",
      "Specificity: 0.8571428571428571\n",
      "\n",
      "Infiltration\n",
      "Sensitivity: 0.0\n",
      "Specificity: 0.3888888888888889\n",
      "\n",
      "Cardiomegaly\n",
      "Sensitivity: 1.0\n",
      "Specificity: 0.8888888888888888\n",
      "\n",
      "Mass\n",
      "Sensitivity: 0.8666666666666667\n",
      "Specificity: 0.9285714285714286\n",
      "\n",
      "Nodule\n",
      "Sensitivity: 0.5384615384615384\n",
      "Specificity: 1.0\n",
      "\n"
     ]
    }
   ],
   "source": [
    "for i in ['Atelectasis','Effusion','Pneumothorax','Infiltration','Cardiomegaly','Mass','Nodule']:\n",
    "\n",
    "    tn, fp, fn, tp = sklearn.metrics.confusion_matrix(data[data[i]==1].Pneumonia.values,\n",
    "                                                  data[data[i]==1].algorithm_output.values,labels=[1,0]).ravel()\n",
    "    sens = tp/(tp+fn)\n",
    "    spec = tn/(tn+fp)\n",
    "\n",
    "    print(i)\n",
    "    print('Sensitivity: '+ str(sens))\n",
    "    print('Specificity: ' +str(spec))\n",
    "    print()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Statement on algorithmic limitations:\n",
    "\n",
    "The results above indicate that the presence of infiltrations in a chest x-ray is a limitation of this algorithm, and that the algorithm performs very poorly on the accurate detection of pneumonia in the presence of infiltration. The presence of nodules and pneumothorax have a slight impact on the algorithm's sensitivity and may reduce the ability to detect pneumonia, while the presence of effusion has a slight impact on specificity and may increase the number of false positive pneumonia classifications."
   ]
  }
 ],
 "metadata": {
  "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": "3.7.6"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}