predict-cancer.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Binary Prediction Metrics\n",
    "\n",
    "As a dataset for introducing binary prediction metrics, we'll use the (real) [Wisconsin Breast Cancer dataset](https://archive.ics.uci.edu/ml/machine-learning-databases/breast-cancer-wisconsin/wdbc.names), that is included in the `sklearn` library. The purpose of this prediction is classifying tumors as **benign** or **malignant**.\n",
    "\n",
    "It uses **logistic regression** as a mechanism for performing the _classification task_. "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Set up"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Import packages\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "import seaborn as sns # for visualiation\n",
    "import statsmodels.formula.api as smf # linear modeling\n",
    "import statsmodels.api as sm\n",
    "import matplotlib.pyplot as plt # plotting\n",
    "from sklearn import metrics\n",
    "\n",
    "# Import the dataset from sklearn\n",
    "from sklearn.datasets import load_breast_cancer\n",
    "data = load_breast_cancer()\n",
    "\n",
    "# Convert to Pandas dataframe for some exploration\n",
    "df = pd.DataFrame(columns=data.feature_names, data=data.data)\n",
    "df.columns = [c.replace(' ', '_') for c in df.columns]\n",
    "df['outcome'] = data.target"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Flip benign and malignant codes (currently 0 for malignant, 1 for benign)\n",
    "df.outcome = np.where(df.outcome == 1, 0, 1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "smoothness_error          -0.067016\n",
       "mean_fractal_dimension    -0.012838\n",
       "texture_error             -0.008303\n",
       "symmetry_error            -0.006522\n",
       "fractal_dimension_error    0.077972\n",
       "concavity_error            0.253730\n",
       "compactness_error          0.292999\n",
       "worst_fractal_dimension    0.323872\n",
       "mean_symmetry              0.330499\n",
       "mean_smoothness            0.358560\n",
       "concave_points_error       0.408042\n",
       "mean_texture               0.415185\n",
       "worst_symmetry             0.416294\n",
       "worst_smoothness           0.421465\n",
       "worst_texture              0.456903\n",
       "area_error                 0.548236\n",
       "perimeter_error            0.556141\n",
       "radius_error               0.567134\n",
       "worst_compactness          0.590998\n",
       "mean_compactness           0.596534\n",
       "worst_concavity            0.659610\n",
       "mean_concavity             0.696360\n",
       "mean_area                  0.708984\n",
       "mean_radius                0.730029\n",
       "worst_area                 0.733825\n",
       "mean_perimeter             0.742636\n",
       "worst_radius               0.776454\n",
       "mean_concave_points        0.776614\n",
       "worst_perimeter            0.782914\n",
       "worst_concave_points       0.793566\n",
       "outcome                    1.000000\n",
       "Name: outcome, dtype: float64"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# What is the correlation between each column and the outcome of interest\n",
    "corrs = df.corr().outcome\n",
    "corrs.sort_values()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Data Exploration\n",
    "\n",
    "This section briefly explores the dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Distribution of outcomes\n",
    "plt.hist(df.outcome)\n",
    "plt.xticks((0, 1), (\"Benign\", \"Malignent\"))\n",
    "plt.xlabel(\"Cancer Status\")\n",
    "plt.ylabel(\"Count\")\n",
    "plt.title(\"Distribution of Cancer Status in the Dataset\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# What is the distribution of \"worst_concave_points\" (for benign, malignant)?\n",
    "malignant = df[df.outcome == 1]\n",
    "benign = df[df.outcome != 1]\n",
    "\n",
    "# Draw histograms of the distribution\n",
    "plt.hist(malignant[\"worst_concave_points\"], alpha=0.5, label='Malignant', color=\"red\")\n",
    "plt.hist(benign[\"worst_concave_points\"], alpha=0.5, label='Benign', color=\"blue\")\n",
    "plt.legend(loc='upper right')\n",
    "plt.xlabel(\"Worst Concave Points in Tumor\")\n",
    "plt.ylabel(\"Count\")\n",
    "plt.title(\"Distribution of Worst Concave Points (by cancer status)\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Using Logistic Regression for Prediction"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Use the `smf.glm` function to fit `outcome` to `worst concave point`\n",
    "logistic_model = smf.glm(formula = 'outcome ~ worst_concave_points', data=df, family=sm.families.Binomial()).fit()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Generate a set of predicted probabilities using your model\n",
    "df['logistic_preds'] = logistic_model.predict()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x720 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Visualize the predicted probabilities across balances (bonus: add the observed values as well)\n",
    "sorted = df.sort_values('worst_concave_points')\n",
    "plt.figure(figsize=(20,10))\n",
    "plt.scatter(x=sorted.worst_concave_points, y=sorted.outcome, c=\"blue\", label=\"data\")\n",
    "plt.plot(sorted.worst_concave_points, sorted.logistic_preds, c=\"black\", label=\"Logistic predictions\")\n",
    "plt.xlabel(\"Worst Concave Point\")\n",
    "plt.ylabel(\"Malignancy\")\n",
    "plt.title(\"Worst Concave Point Versus\")\n",
    "plt.legend(loc='upper left')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Evaluation Metrics"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Sensitivity and Specificity\n",
    "The concepts of sensitivity and specificity express how well your model accurately predicts positive and negative classes. In order to compute these, it is easy to first build a **confusion matrix** to show the total number of cases that were accurately predicted (and those that were not).\n",
    "\n",
    "  <table>\n",
    "  <tr>\n",
    "    <td/>\n",
    "    <td/>\n",
    "    <td colspan=\"2\"><em>Predicted Class</em></td>\n",
    "  </tr>\n",
    "  <tr>\n",
    "    <td/>\n",
    "    <td/>\n",
    "    <td><strong>Negative (-)</strong></td>\n",
    "    <td><strong>Positive (+)</strong></td>\n",
    "    <td><strong>Total</strong></td>\n",
    "  </tr>\n",
    "  <tr>\n",
    "    <td rowspan=\"2\"><em>True Class</em></td>\n",
    "    <td><strong>Negative (-)</strong></td>\n",
    "    <td>True Negative (TN)</td>\n",
    "    <td>False Positives (FP)</td>\n",
    "    <td>Total non-cases (N)</td>\n",
    "  </tr>\n",
    "  <tr>\n",
    "    <td><strong>Positive (+)</strong></td>\n",
    "    <td>False Negatives (FN)</td>\n",
    "    <td>True Positives (TP)</td>\n",
    "    <td>Total cases (P)</td>\n",
    "  </tr>\n",
    "</table>\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "In the matrix above, the rows represent the **data**, while the columns represent the **predictions**. For example, the top left cell (_True Negatives_) represents the number of instances in which someone did not default, and the model predicted that they did not default. Here are the definitions of each cell:\n",
    "\n",
    "- **True Positives**: The instances in which your model **predicts** a case (and the case **is true** in the data)\n",
    "- **False Positives**: The instances in which your model **predicts** a case (and the case **is not** true in the data)\n",
    "- **False Negatives**: The instances in which your model **does not** predict a case (and the case **is true** in the data)\n",
    "- **True Negatives**: The instances in which your model **does not** predict a case (and the case **is not** true in the data)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Using those values, we can compute the **sensitivity** and **specificity** of our model:\n",
    "\n",
    "\\begin{equation*}\n",
    "Sensitivity = \\frac { True Positives }{ True Positives+False Negatives } \n",
    "\\end{equation*}\n",
    "\n",
    "\\begin{equation*}\n",
    "Specificity = \\frac { TrueNegatives }{ TrueNegatives+FalsePositives } \n",
    "\\end{equation*}"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "These two metrics lend important insights regarding the quality of our model:\n",
    "\n",
    "> _**Sensitivity**_, also referred to as the _true positive rate_, tells us, of all of the **cases in the data**, how many did we accurately predict? This indicates the model's **ability to detect cases**. In other words, how **sensitively** does the model pick up on cases?\n",
    "\n",
    "> _**Specificity**_, also referred to as the _true negative rate_, tells us, of all of the **non-cases in the data**, how many did we accurately predict? This indicates the model's ability to assign non-cases.\n",
    "\n",
    "These metrics are directly used to calculate **Type I and Type II error rate**, which are analagous to Type I and Type II errors in statistical tests (incorrectly reject a true null hypothesis, incorrectly accept a false null hypothesis). \n",
    "\n",
    "> **Type I Error** rate is the proportion of instances which are **incorrectly classified as positive cases** (relative to the total number of **negative cases**). It is calculated as $1-specificity$, or simply the false positives relative to the total non-cases in the data, $FP/N$.\n",
    "\n",
    "> **Type II Error** rate is the proportion of instances which are **incorrectly classified as negative cases** (relative to the total number of **positive cases**). It is calculated as $1-sensitivity$, or simply the false negatives relative to the total cases in the data, $FN/P$.\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "collapsed": true
   },
   "source": [
    "# Metric Calculation (15 minutes)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "In this section, you'll evaluate the fit of your model by calculating the following:\n",
    "\n",
    "- First, using a threshold of .5, use your model to predict a binary outcome (each case as 0 or 1)\n",
    "- What is the **accuracy** of your model (how often does the prediction match the data)?\n",
    "- What are the sensitivity, specificity, Type I error rate, and Type II error rates?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "# First, using a threshold of .5, use your model to predict a binary outcome (each case as 0 or 1)\n",
    "df['binary_preds'] = np.where(df.logistic_preds > .5, 1, 0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.9086115992970123"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# What is the accuracy of your model (how often does the prediction match the data)?\n",
    "(df.binary_preds == df.outcome).sum()/df.shape[0] # 91% accurate!"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Sensitivity: 0.8726415094339622 Specificity 0.9299719887955182\n"
     ]
    }
   ],
   "source": [
    "# What are the sensitivity, specificity, Type I error rate, and Type II error rates?\n",
    "cm = metrics.confusion_matrix(df.outcome, df.binary_preds)\n",
    "true_pos = cm[1,1]\n",
    "false_pos = cm[0,1]\n",
    "true_neg = cm[0,0]\n",
    "false_neg = cm[1,0]\n",
    "\n",
    "# Sensitivity, specificity, and error rates\n",
    "sensitivity = true_pos / (true_pos + false_neg)\n",
    "specificity = true_neg / (true_neg + false_pos)\n",
    "type_one_error = 1 - specificity\n",
    "type_two_error = 1 - sensitivity\n",
    "print(\"Sensitivity:\", sensitivity, \"Specificity\", specificity)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "What is your interpretation of these values?\n",
    "\n",
    "> The model is really good at predicting _who will not default_ (specificity). It has a Type I error rate of .4%. However, it is **really bad** at predicting _who will default_. The Type II error rate is 70% (meaning that it does not discover 70% of cases!). Note, the accuracy seems quite impressive -- it can accurately interpret **95%** of cases. However, recall that only **3%** of cases defaulted. This means that a naive model that predicts **no one defaults** would have an accuracy of **97%**. We can choose a specific threshold (default is .5) and assess the sensitivity and specificity of our model."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## ROC Curve (10 minutes)\n",
    "\n",
    "It is common to compare the _true positive rate_ (sensitivity) to the _false positive rate_ (1 - specificity) at each **threshold** for classification in an [ROC Curve](https://en.wikipedia.org/wiki/Receiver_operating_characteristic). This [interactive visualization](http://www.navan.name/roc/) may help you better understand the relationship between thresholds and the ROC. In this section, you'll do the following:\n",
    "\n",
    "- Generate data for the ROC curve using the `metrics.roc_curve` function\n",
    "- Draw your ROC curve\n",
    "- Calculate the area under your ROC curve using the `metrics.roc_auc_curve` function"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Generate data for the ROC curve using the `metrics.roc_curve` function\n",
    "fpr, tpr, thresholds  = metrics.roc_curve(df.outcome, df.logistic_preds)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0,0.5,'True Positive Rate')"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x720 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Draw your ROC curve\n",
    "plt.figure(figsize=(20,10))\n",
    "plt.title(\"ROC Curve for the model\")\n",
    "plt.plot(fpr, tpr)\n",
    "plt.plot(fpr, fpr, 'b--')\n",
    "plt.axis([-.02,1.02, -.02,1.02])\n",
    "plt.xlabel('False Positive Rate')\n",
    "plt.ylabel('True Positive Rate')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.9013067491147403"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Calculate the area under your ROC curve using the metrics.roc_auc_curve function\n",
    "metrics.roc_auc_score(df.outcome, df.binary_preds)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Write data\n",
    "df.to_csv(\"public/data/cancer_data.csv\", index=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "roc_data = pd.DataFrame({\"fpr\":fpr, \"tpr\":tpr, \"thresholds\":thresholds})"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "roc_data.to_csv(\"public/data/roc_data.csv\", index=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "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": "3.6.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}