ch01-导论/d1-2回归分析(糖尿病人).ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 实例：线性回归\n",
    "- 预测出连续值\n",
    "- scikit-learn\n",
    "- [糖尿病人数据](http://scikit-learn.org/stable/datasets/index.html#datasets)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 利用 diabetes(糖尿病人)数据集来学习线性回归\n",
    "- [Linear Regression Example](http://scikit-learn.org/dev/auto_examples/linear_model/plot_ols.html)\n",
    "包括442个病人的生理数据及一年以后的病情发展情况\n",
    "10项特征值如下:  \n",
    "年龄  \n",
    "性别  \n",
    "体质指数  \n",
    "血压\n",
    "\n",
    "s1,s2,s3,s4,s4,s6 (六种血清的化验数据)  \n",
    " \n",
    "#### sklearn"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "import numpy as np\n",
    "from sklearn import datasets, linear_model\n",
    "from sklearn.metrics import mean_squared_error, r2_score"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['age', 'sex', 'bmi', 'bp', 's1', 's2', 's3', 's4', 's5', 's6']\n"
     ]
    }
   ],
   "source": [
    "# 加载糖尿病数据\n",
    "diabetes = datasets.load_diabetes()\n",
    "\n",
    "# 列举一下特征名称\n",
    "# ['age', 'sex', 'bmi', 'bp', 's1', 's2', 's3', 's4', 's5', 's6']\n",
    "# 年龄，性别，身体质量指数，血压，s1-s6是6中血清的化验数据\n",
    "print(diabetes.feature_names)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "diabetes.DESCR\n",
    "diabetes.feature_names\n",
    "print(diabetes.data,'\\n')\n",
    "print(diabetes.data.shape,'\\n')\n",
    "print(diabetes.data.dtype)\n",
    "print(diabetes.target.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "均方误差是:2548.07\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXoAAAD8CAYAAAB5Pm/hAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDIuMi4zLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvIxREBQAAHRBJREFUeJzt3X2QXHWd7/H3d0IEJqgQE2JIMt1xCaxR3AgjxF0u8rgKWxpRWNiaheBlHSzh1lpLWaBzfeCWUyC6urgu7A7CmpCGKGFVruKtBRYEtxR2Jk8mcCEhzEwy5IkHFZjcxGS+949zOtOZ9MPpnn485/Oq6pru07/u+c7J9Ce/+f1+5xxzd0REJL7aGl2AiIjUloJeRCTmFPQiIjGnoBcRiTkFvYhIzCnoRURiTkEvIhJzCnoRkZhT0IuIxNwRjS4AYMaMGZ5OpxtdhohISxkYGHjZ3WeWatcUQZ9Op+nv7290GSIiLcXMhqK009CNiEjMKehFRGJOQS8iEnMKehGRmFPQi4jEnIJeRKRCmUyGdDpNW1sb6XSaTCbT6JLyaorllSIirSaTydDd3c3o6CgAQ0NDdHd3A9DV1dXI0g6jHr2ISAV6enoOhnzW6OgoPT09DaqoMAW9iEgFhoeHy9reSAp6EZEKdHR0lLW9kRT0IiIV6O3tpb29/ZBt7e3t9Pb2NqiiwhT0IiIV6Orqoq+vj1QqhZmRSqXo6+truolYAHP34g3MjgKeAI4kWKWzyt2/YmbzgZXAO4AB4Ap332dmRwLLgdOAV4DL3H2w2Pfo7Ox0ndRMRKQ8Zjbg7p2l2kXp0e8FznX3PwEWAR8xs8XA14Fvu/uJwGvA1WH7q4HXwu3fDtuJiEiDlAx6D7wRPpwa3hw4F1gVbl8GfDy8vyR8TPj8eWZmVatYRETKEmmM3symmNlaYBfwMPAC8Ft33x822QbMCe/PAbYChM//jmB4R0REGiBS0Lv7AXdfBMwFTgf+eLLf2My6zazfzPp379492bcTEZECylp14+6/BR4DPggca2bZUyjMBUbC+yPAPIDw+bcTTMpOfK8+d+90986ZM0teCUtERCpUMujNbKaZHRvePxq4AHiWIPAvCZstBX4S3n8wfEz4/H94qaU9IiJSM1FOajYbWGZmUwj+Y/ihu//UzJ4BVprZ14A1wF1h+7uAe8xsM/AqcHkN6hYRkYhKBr27rwfen2f7FoLx+onb/x9waVWqExGRSdORsSIiMaegFxGJOQW9iEjMKehFRGJOQS8iEnMKehGRmFPQi4jEnIJeRCTmFPQiIjGnoBcRiTkFvYhIzCnoRURiTkEvIhJzCnoRkZhT0IuIxJyCXkQk5hT0IiIxp6AXEYk5Bb2ISMwp6EVEYk5BLyIScwp6EZGYU9CLiMScgl5EJOYU9CIiMaegFxGJOQW9iEjMKehFRGJOQS8iEnMlg97M5pnZY2b2jJltNLO/Dbd/1cxGzGxteLso5zVfMLPNZvacmX24lj+AiIgUd0SENvuB6919tZm9FRgws4fD577t7t/MbWxmC4HLgfcAJwCPmNlJ7n6gmoWLiEg0JXv07r7d3VeH918HngXmFHnJEmClu+919xeBzcDp1ShWRETKV9YYvZmlgfcDT4WbrjOz9WZ2t5kdF26bA2zNedk2iv/HICIiNRQ56M3sGOAB4HPu/nvgDuCPgEXAduDvy/nGZtZtZv1m1r979+5yXioiImWIFPRmNpUg5DPu/m8A7r7T3Q+4+xhwJ+PDMyPAvJyXzw23HcLd+9y90907Z86cOZmfQUREioiy6saAu4Bn3f1bOdtn5zS7GNgQ3n8QuNzMjjSz+cAC4OnqlSwiIuWIsurmz4ArgN+Y2dpw2xeBvzKzRYADg8A1AO6+0cx+CDxDsGLnWq24ERFpnJJB7+6/BCzPUw8VeU0v0DuJukREpEp0ZKyISMwp6EVEYk5BLyIScwp6EZEKPf443HYbbNzY6EqKU9CLSOJkMhnS6TRtbW2k02kymUzk1x44ANdfD2Zwzjnwuc/BmWfCm2/WsOBJirK8UkQkNjKZDN3d3YyOjgIwNDREd3c3AF1dXQVfNzQEZ58Ng4OHP7dvH7jXoNgqUY9eRBKlp6fnYMhnjY6O0tPTk7f9vfcGvfd0On/IA9x/PxxzTHXrrCYFvYgkyvDwcMnto6Nw6aVBwBfp5PPTnwY9+YsuKtymGSjoRSRROjo6Cm5fvRqmTIFp02DVqvyvP/VU2LEjCPi/+IsaFlpFCnoRSZTe3l7a29sP2TZ16pcYGhrktNNgbKzQ64LnBgZg1qw6FFpFmowVkUTJTrjecMPtjIz8AJjLH/5QuP1//Rd0dtantlpRj15EEuW734W//usuRkb+k+As6oe7+GJ4441geKbVQx7UoxeRBNi7F046CQrMwx60bBlceWV9aqonBb2IxNbTT8MZZ5Rut2ULzJ9f+3oaRUM3IgkymSNCW8l11wVLI4uF/Mknjx/oFOeQB/XoRRKj0iNCW8Wrr8I73lG6XV8ffPrTta+nmZg3wXG7nZ2d3t/f3+gyRGItnU4zNDR02PZUKsVgoUM+W8ADD8All5Rut3UrzM0/99qyzGzA3UtOF2voRiQhohwR2irGxuBDHwqGZ4qF/CWXBG3d4xfy5VDQiyREsSNCW8WmTUG4T5kCTzxRuN2//3sQ7vffH7RPOgW9SELkOyK0vb2d3t7mv7zzrbcGgX3SSYXbvOUt8PrrQcBfcEH9amsFCnqRhOjq6qKvr49UKoWZkUql6Ovra9qJ2D17YPr0IOBvuKFwuy9/OQj3vXub+wySjaTJWBFpKk8+CWedVbrdxo2wcGHt62lmmowVkZZy1VVB771YyH/gA7B/f9CDT3rIl0Pr6EWkYXbuhHe+s3S7FSuKnxdeilPQi0jd3XNPtHPK7NwJxx9f+3riTkM3IlIXBw4EF+0wKx7yn/pUMDTjrpCvFvXoRaSmNmyAU04p3e7JJ+HMM2tfTxIp6EWkJk4+GZ5/vnibGTOCUwcffXR9akoqDd2ISNXs3h0MzZgVD/lbbw2GZnbvVsjXQ8mgN7N5ZvaYmT1jZhvN7G/D7dPN7GEz2xR+PS7cbmb2HTPbbGbrzezUWv8QItJY3/hGEO6lxtQ3bQoC/vOfr09dEogydLMfuN7dV5vZW4EBM3sYuAp41N1vMbMbgRuBG4ALgQXh7QzgjvCriMSIO7RFHBM4cCB6W6m+krve3be7++rw/uvAs8AcYAmwLGy2DPh4eH8JsNwDvwaONbPZVa9cRBpi3bqg914quG+4YXz1jEK+scqajDWzNPB+4ClglrtvD5/aAcwK788Btua8bFu4bXvONsysG+iG1jp7nkhSXXoprFpVut3wMMybV/t6JLrIQW9mxwAPAJ9z999bzrk/3d3NrKyT5rh7H9AHwbluynmtiNTHnj0w4YSXeR1zTHDmSGlOkf6gMrOpBCGfcfd/CzfvzA7JhF93hdtHgNz/z+eG20SkRaxaFQzPlAr5++4LhmYU8s2tZI/egq77XcCz7v6tnKceBJYCt4Rff5Kz/TozW0kwCfu7nCEeEWliM2bAK6+Ubvfmm9F6+tIcovTo/wy4AjjXzNaGt4sIAv4CM9sEnB8+BngI2AJsBu4EPlv9skWkWrZtG1/7XizkP/ax8clVhXxrKdmjd/dfAoUuxnVenvYOXDvJukSkxr70Jfja10q3GxgIzlEjrUunQBBJkAMH4IiIn/qxMV1vNS60ulUkAX784yC0S4X8zTePD88o5ONDPXqRGIsa1jt2wKxZpdtJa1KPXiRmXnllfHK1mDlzxnvvCvl4U9DXUSaTIZ1O09bWRjqdJpPJNLokiZEbbwzCfcaM4u1uvz0I923b6lOXNJ6Gbuokk8nQ3d3N6OgoAENDQ3R3dwPQpYthyiREHZ7ZsweOOqq2tUhzUo++Tnp6eg6GfNbo6Cg9PT0Nqkha2Zo10YZn5s8fH55RyCeXgr5OhoeHy9oukk/2mqul1rX/6ldBuG/ZUp+6pLkp6Ouk0Bk6debO5Kh0jmbfvvHe+5o1xduOjQUBv3hxFQqW2FDQ10lvby/tE44bb29vp7e3t0EVST1l52iGhoZw94NzNMXCfvnyINyPPLL4e197rda+S3EWnLGgsTo7O72/v7/RZdRcJpOhp6eH4eFhOjo66O3t1URsQqTTaYaGhg7bnkqlGBwcPGSb1r5LVGY24O6dJdsp6EVqr62tjXyfNTNjbGyMl14K1rVH0QQfWWkSUYNeQzcidVBoLmbatBWYlQ75e+8dH55pFTpupHloHb1IHfT29h5yHAUEif3GG8Vft28fTJ1a29pqQceNNBf16EXqoKuri8985mcEAV+8W7548XjvvRVDHnTcSLNRj16kxsYnV88u2m79ejjllFpXUx86bqS5qEcvUgOvvx7tyFUY773HJeRBx400GwW9SBVdf30Q7m97W/F2f/d3rTe5Wg4dN9JcNHQjUgVR177v2gUzZ9a2lmaQnXDVcSPNQevoRSq0cSO8973R2jbBx0xiSOvoRWokO/ZeKuTvvjvewzPSOjR0IxJBORfV3r8fpkypbT0i5VCPXqSIb34z2kW1cy/Lp5CXZqMevUgeUSdX162D972vtrWITJaCXiS0YwfMnh2trcbdpZVo6EYS74MfDHrwpUJ+yRJNrkprUo9eEivq8Myrr8Jxx9W2FpFaUo8+pnSK2Px+9KPyT02gkJdWVzLozexuM9tlZhtytn3VzEbMbG14uyjnuS+Y2WYze87MPlyrwqWwSi5bF3fZcP/EJ4q3u+suDc9I/JQ8MtbMzgLeAJa7+3vDbV8F3nD3b05ouxC4DzgdOAF4BDjJ3Q8U+x46Mra6yrlsXZzt3QtHHRWt7diYrrcqradqR8a6+xPAqxG/7xJgpbvvdfcXgc0EoS91lPRTxF55ZRDaUUJeF9WWJJjMGP11ZrY+HNrJjmLOAbbmtNkWbpM6SuopYrPDM/fcU7zdwICGZyRZKg36O4A/AhYB24G/L/cNzKzbzPrNrH/37t0VliH5JOkUsc8+W/7k6qmn1r4ukWZSUdC7+053P+DuY8CdjA/PjADzcprODbfle48+d+90986ZSThvax11dXXR19dHKpXCzEilUvT19cXqFLHZcF+4sHi7j35UvXeRitbRm9lsd98ePrwYyK7IeRC418y+RTAZuwB4etJVStm6urpiFewQhHVbxK7JG2/AtGm1rUekVZQMejO7j+BilzPMbBvwFeBsM1tEcJXjQeAaAHffaGY/BJ4B9gPXllpxI1LKnXdCd3e0tuq5ixxOFx6RphV1Jczy5XDFFbWtRaQZRV1eqVMgSFP5/e/h7W+P1rYJ+igiLUGnQJCmcN55QQ8+SshrclWkPOrRS0NFHZ7ZtAlOPLG2tYjElXr0Unfr1pW/9l0hL1I5Bb3UTTbcFy0q3u6aazQ8I1JNGrqRmhobi34N1X37YOrU2tYjkkTq0UtN3HFH0HuPEvKpVBqzNhYs0HnzRWpBPXqpqqiTq7/4BWzdmj1v/ijAwfPmA7E7qlekkXTAlEzayy9D1NMV5f666bz5IpNTtfPRixRywQVBD75UyJ9/fv7J1aSfN1+kXjR0I2WLOjyza1fx/wQ6Ojry9ujjft58kXpTj14ieeKJ8te+l+rpJ+m8+SKNpKCXorLh/qEPFW/3T/9U/tr3JJw3X6QZaDJWDrN/f/T17AcORD9HvIhUlyZjpWw33RT03qOEfLb3rpAXaX6ajJXIk6tr1pQ+fYGINB8FfUIND0MqFa1tE4zuicgk6A/vhHn3u4MefKmQnzbtAZ1YTCQm1KNPiKjDM3AM8CajowaM1a4gEakb9ehj7PHHo699BwtvbwI6aEkkThT0MbRoURDu55xTvN3KlbBiRYb29mmHbNdBSyLxoqCPiX37xnvv69YVb5sde7/sMh20VIlMJkM6naatrY10WqdWluanA6Za3LJlcNVVpdstWADPP1/zcmIvkwlOrTw6OnpwW3t7u/5zlIaIesCUgr5FRZ1c3bEDZs2qbS1JolMrSzPRkbExNDJS/onFFPLVpVMrSytS0LeAT386CPe5c4u3u/deXVS71gqtRtIqJWlmWkffpMo5j8wf/gBH6F+yLnp7e/OO0WuVkjQz9eibzJNPBr33UiH/p3863ntv5ZBvtRUsWqUkrUiTsU3ihBNg+/bS7davh1NOqX099aAVLCKTU7XJWDO728x2mdmGnG3TzexhM9sUfj0u3G5m9h0z22xm683s1Mn9GPE2Ojo+uVoq5LO997iEPEBPT88hIQ8wOjpKT09PgyoSiacoQzffBz4yYduNwKPuvgB4NHwMcCGwILx1A3dUp8x4ue22INynTSve7itfiffkqlawiNRHydFdd3/CzNITNi8Bzg7vLwMeB24Ity/3YDzo12Z2rJnNdvcIgxLxF3Xt+2uvwbHH1raWZqCLg4vUR6WTsbNywnsHkF2tPQfYmtNuW7gtsbZvL3/texJCHnRxcJF6mfSqm7D3Xvbggpl1m1m/mfXv3r17smU0nZtvDsL9hBOKt3vooXgPzxSjFSwi9VFp0O80s9kA4ddd4fYRYF5Ou7nhtsO4e5+7d7p758yZMysso7m4w0UXBQH/xS8Wb3vgQND+wgvrU1sUjVjq2NXVxeDgIGNjYwwODirkRWqg0qB/EFga3l8K/CRn+5Xh6pvFwO+SMD4/ODi+9v3nPy/c7qabmvei2tmljkNDQ7g7Q0NDdHd3N/26dhEpLcryyvuAXwEnm9k2M7sauAW4wMw2AeeHjwEeArYAm4E7gc/WpOom8Y//GAT8/PnF2730UhDuX/5yferKFbWXrqWOIvGlA6bKtHcvnHgibNtWvN3nPw+33lqfmgop54CktrY28v0umBljY7qkoEgz0tkrq+ypp4Le+1FHFQ/5NWuC3nujQx7K66XrZF0i8aWgL+Haa4OAX7y4cJuFC4MrPLkHl/FrFuUckKSljiLxpaDP45VXxte+33574Xbf+14Q7hs3wtSp9asvqnJ66VrqKBJfCvoc998fhPuMGcXbbdsWBPzVV9enrkqV20vXUkeReEp80I+NwZlnBgH/l39ZuN1llwVt3WFOixzrq166iECCV93s2QMTOrt5PfIInHde7esRESlX1FU3LXzJism54orCzx19NOzaBcccU796RERqJbFDN+vXH74te+Tq6KhCXkTiI7FBf889sDQ8icMzzzTuyFURkVpLbNCfcQZ8//tBwL/73Y2uRkSkdhIb9CIiSaGgFxGJOQW9iEjMKehroBEX8BARKSSx6+hrZeKpgbMX8AB0RKqINIR69FWmC3iISLNR0FdZOacGFhGpBwV9lcXlAh5JnWdI6s8tMefuDb+ddtppHhcrVqzw9vZ2Bw7e2tvbfcWKFY0uLbI4/AyVSOrPLa0L6PcIGdvwkPeYBb17EBipVMrNzFOpVMsFRSqVOiTssrdUKlX0dUn9uUUaJWrQJ/Y0xVJYJRcKL+dC5M1KF0iXVqOLg0vFKplniMNqo7jMr4hMpKCXw1RyofA4rDbSBdIlrhT0cphKLkEYh96wLr0osRVlIL/Wt0ZNxrb65GEz0YoVkfoj4mRsYnv02cnDoaEh3P3gqQq0broy6g2LNK/ErrpJp9MMDQ0dtj2VSjE4OFjXWkREKqFVNyXEYfJQRCSKxAZ9HCYPRUSimFTQm9mgmf3GzNaaWX+4bbqZPWxmm8Kvx1Wn1OrSUjoRSYpq9OjPcfdFOeNENwKPuvsC4NHwcdPR5GFy6cRlkjhRluYUugGDwIwJ254DZof3ZwPPlXqfRiyv1NLKZNIyUIkT6nGuGzN7EXgt/MD8i7v3mdlv3f3Y8HkDXss+nvDabqAboKOj47R8K2BqJQ7nZZHKaLWVxEnUVTeTDfo57j5iZscDDwP/A3gwN9jN7DV3LzpOX+/llfqwJ5dOXCZxUpflle4+En7dBfwIOB3YaWazwyJmA7sm8z1qQUsrk0urrSSJKg56M5tmZm/N3gf+HNgAPAgsDZstBX4y2SKrTR/25NJqK0miyfToZwG/NLN1wNPAz9z9/wC3ABeY2Sbg/PBxU9GHPbm02kqSKLGnQMhkMvT09DA8PExHRwe9vb36sItIS6nLZGy16ApTIiLl07luREQEUNCLiMSegl5EJOYU9CIiMaegFxGJuaZYdWNmu4H6newmMAN4uc7fsxKqs7pUZ/W1Sq1xrDPl7jNLNWqKoG8EM+uPsiyp0VRndanO6muVWpNcp4ZuRERiTkEvIhJzSQ76vkYXEJHqrC7VWX2tUmti60zsGL2ISFIkuUcvIpIIsQt6M5tuZg+b2abwa96rW5nZ0rDNJjNbGm57q5mtzbm9bGb/ED53lZntznnubxpVZ7j9cTN7Lqee48PtR5rZD8xss5k9ZWbpRtVpZu1m9jMz+79mttHMbslpX5X9aWYfCffDZjM77EL0xfaHmX0h3P6cmX046nvWs04zu8DMBszsN+HXc3Nek/d3oEF1ps1sT04t/5zzmtPC+jeb2XfMzBpYZ9eEz/iYmS0Kn2vE/jzLzFab2X4zu2TCc4U+++XvzygXlm2lG3ArcGN4/0bg63naTAe2hF+PC+8fl6fdAHBWeP8q4LvNUifwONCZ5zWfBf45vH858ING1Qm0A+eEbd4CPAlcWK39CUwBXgDeFb7/OmBhlP0BLAzbHwnMD99nSpT3rHOd7wdOCO+/FxjJeU3e34EG1ZkGNhR436eBxYABP8/+DjSizgltTgFeaPD+TAPvA5YDl5T6TFW6P2PXoweWAMvC+8uAj+dp82HgYXd/1d1fI7je7UdyG5jZScDxBOHUtHWWeN9VwHmT7EFVXKe7j7r7YwDuvg9YDcydRC0TnQ5sdvct4fuvDOstVH/u/lgCrHT3ve7+IrA5fL8o71m3Ot19jbu/FG7fCBxtZkdOsp6q11noDS24nOjb3P3XHqTUcvL/DjWizr8KX1srJet090F3Xw9MvGBx3s9UpfszjkE/y923h/d3EFwJa6I5wNacx9vCbbmyvYDc2epPmtl6M1tlZvOaoM5/Df/E/FLOL/HB17j7fuB3wDsaXCdmdizwUeDRnM2T3Z9R/h0L7Y9Cr43ynvWsM9cngdXuvjdnW77fgUbVOd/M1pjZL8zsv+W031biPetdZ9ZlwH0TttV7f5b72or25xERv2lTMbNHgHfmeaon94G7u5lVuqzocuCKnMf/G7jP3fea2TUEvYVz876yPnV2ufuIBdftfSCsdXmZ71GPOjGzIwg+UN9x9y3h5rL3Z5KZ2XuArxNcmzmrar8DVbAd6HD3V8zsNODHYc1NyczOAEbdfUPO5mban1XVkkHv7ucXes7MdprZbHffHv6ZsytPsxHg7JzHcwnG57Lv8SfAEe4+kPM9X8lp/z2CseuG1enuI+HX183sXoI/E5eHr5kHbAsD9u1Abu11rTPUB2xy93/I+Z5l788C3zf3L4G54bZ8bSbuj2KvLfWe9awTM5sL/Ai40t1fyL6gyO9A3esM//LdG9YzYGYvACeF7XOH6xq+P0OXM6E336D9Wey1Z0947eNUuj+rNfHQLDfgGxw6eXhrnjbTgRcJJjmOC+9Pz3n+FuCmCa+ZnXP/YuDXjaqT4D/oGWGbqQRjkJ8JH1/LoZNQP2zk/gS+RtA7aqv2/gz3wxaCydTsZNd7JrTJuz+A93DoZOwWgsmzku9Z5zqPDdt/Is975v0daFCdM4Ep4f13EYRP9ndg4uThRY2qM3zcFtb3rkbvz5y23+fwydhCn6my92fFP0Sz3gjG4R4FNgGP5OycTuB7Oe3+O8EE3GbgUxPeYwvwxxO23UwwGbYOeGzi8/WsE5hGsCJofVjTbTkfsqOA+8P2T+f+MjegzrmAA88Ca8Pb31RzfwIXAc8TrG7oCbf9L+BjpfYHwdDUC8Bz5KxcyPeeVfi9rKhO4H8Cb+bsv7UEiwQK/g40qM5PhnWsJZh0/2jOe3YCG8L3/C7hgZqNqDN87mwmdCwauD8/QDDO/ibBXxwbi32mKt2fOjJWRCTm4rjqRkREcijoRURiTkEvIhJzCnoRkZhT0IuIxJyCXkQk5hT0IiIxp6AXEYm5/w+JhdvNe0dPuAAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# 选择bmi作为预测的指标（特征）\n",
    "diabetes_X = diabetes.data[:, np.newaxis, 2]\n",
    "\n",
    "# 除了最后20个都是训练集\n",
    "diabetes_X_train = diabetes_X[:-20]\n",
    "# 最后20个是测试集\n",
    "diabetes_X_test = diabetes_X[-20:]\n",
    "\n",
    "# 训练集的标注结果\n",
    "diabetes_y_train = diabetes.target[:-20]\n",
    "# 测试集的标注结果\n",
    "diabetes_y_test = diabetes.target[-20:]\n",
    "\n",
    "\n",
    "# 创建线性回归分类器\n",
    "regr = linear_model.LinearRegression()\n",
    "# 使用训练数据训练\n",
    "regr.fit(diabetes_X_train, diabetes_y_train)\n",
    "# 使用测试数据，让模型生成对应的结果\n",
    "diabetes_y_pred = regr.predict(diabetes_X_test)\n",
    "\n",
    "# 比较测试集的真实结果与模型的预测结果，使用均方误差作为评估标准\n",
    "print(\"均方误差是:%.2f\"% mean_squared_error(diabetes_y_test, diabetes_y_pred))\n",
    "\n",
    "# 画点（真实数据）\n",
    "plt.scatter(diabetes_X_test, diabetes_y_test,  color='black')\n",
    "# 画线（模型的直线）\n",
    "plt.plot(diabetes_X_test, diabetes_y_pred, color='blue', linewidth=3)\n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "![](images/d1-1.png)"
   ]
  }
 ],
 "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": "3.6.4rc1"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}