c31_PR curve.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Precision-Recall Curve"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "import matplotlib.pyplot as plt"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "from sklearn import datasets\n",
    "\n",
    "digits = datasets.load_digits()\n",
    "X = digits.data\n",
    "y = digits.target.copy()\n",
    "\n",
    "y[digits.target==9] = 1\n",
    "y[digits.target!=9] = 0"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.model_selection import train_test_split\n",
    "\n",
    "X_train,X_test, y_train,y_test = train_test_split(X,y, random_state=666)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.linear_model import LogisticRegression\n",
    "\n",
    "log_reg = LogisticRegression()\n",
    "log_reg.fit(X_train, y_train)\n",
    "decision_scores = log_reg.decision_function(X_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.metrics import precision_score,recall_score\n",
    "\n",
    "precisions = []\n",
    "recalls = []\n",
    "thresholds = np.arange(decision_scores.min(), decision_scores.max(), 0.1)\n",
    "for threshold in thresholds:\n",
    "    y_predict = np.array(decision_scores>=threshold, dtype='int')\n",
    "    precisions.append(precision_score(y_test, y_predict))\n",
    "    recalls.append(recall_score(y_test, y_predict))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x29598a5b8d0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(thresholds, precisions)\n",
    "plt.plot(thresholds, recalls)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Precision-Recall 曲线"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x29598a722b0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(precisions, recalls)\n",
    "plt.show()\n",
    "# 急剧下降的点, 便是精准率和召回率最佳点"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Scikit-learn中的Precision-Recall曲线"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.metrics import precision_recall_curve"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "# decision_scores = log_reg.decision_function(X_test)\n",
    "precisions, recalls, thresholds = precision_recall_curve(y_test, decision_scores)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(145,)"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# decision_score的步长是sklearn自动确定最佳的\n",
    "precisions.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(145,)"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "recalls.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(144,)"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# the last precision and recall values are 1 and 0 respectively\n",
    "# don't have a corresponding threshold\n",
    "thresholds.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x29598c56518>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(thresholds, precisions[:-1])\n",
    "plt.plot(thresholds, recalls[:-1])\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x29598c50160>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(precisions, recalls)\n",
    "plt.show()"
   ]
  },
  {
   "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": "3.6.4"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}