HW_波士頓房價_前處理實戰_Ans-checkpoint.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 基礎實戰範例-波士頓房價資料處理與視覺化\n",
    ">資料集：波士頓房產<br>\n",
    "特徵：房地產客觀數據，如年份、平面大小<br>\n",
    "預測目標：房地產價格<br>\n",
    "\n",
    "# 實戰一，載入所需要的模組\n",
    "* sklearn.dataset\n",
    "* pandas\n",
    "* numpy\n",
    "* matplotlib\n",
    "\n",
    "\n",
    "# 讀取boston資料集\n",
    "from sklearn.datasets import load_boston<br>\n",
    "boston = load_boston()\n",
    "\n",
    "### 基本資料集內容\n",
    "* 印出boston的key值與檔案位置\n",
    "* 查看前10筆資料\n",
    "* 查看資料型別\n",
    "* 印出標註的樣本類別資料\n",
    "* 取出幾種特徵值欄位assign給X，樣本類別assign給Y\n",
    "* 組成 pandas DataFrame (資料表的方式顯示)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 波士頓房價\n",
    "\n",
    "檔案中包含506條樣本資料，即有506條房價資料，而13表示有13個影響房價的特徵值\n",
    "特徵值解釋如下表\n",
    "![波士頓房價檔案部分中英文標題](boston_1.png)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "資料包內含資料的key值： dict_keys(['data', 'target', 'feature_names', 'DESCR', 'filename'])\n",
      "資料欄位名稱(資料特徵參數)： ['CRIM' 'ZN' 'INDUS' 'CHAS' 'NOX' 'RM' 'AGE' 'DIS' 'RAD' 'TAX' 'PTRATIO'\n",
      " 'B' 'LSTAT']\n",
      "檔案位置: C:\\Users\\User\\anaconda3\\lib\\site-packages\\sklearn\\datasets\\data\\boston_house_prices.csv\n",
      "第一筆資料: [[6.320e-03 1.800e+01 2.310e+00 0.000e+00 5.380e-01 6.575e+00 6.520e+01\n",
      "  4.090e+00 1.000e+00 2.960e+02 1.530e+01 3.969e+02 4.980e+00]]\n",
      "資料型別: <class 'numpy.ndarray'>\n",
      "資料size: 6578\n",
      "前10筆target_names(房價):  [24.  21.6 34.7 33.4 36.2 28.7 22.9 27.1 16.5 18.9]\n",
      "pandas DataFrame\n",
      "      CRIM    ZN  INDUS  CHAS    NOX     RM   AGE     DIS  RAD    TAX  \\\n",
      "0  0.00632  18.0   2.31   0.0  0.538  6.575  65.2  4.0900  1.0  296.0   \n",
      "1  0.02731   0.0   7.07   0.0  0.469  6.421  78.9  4.9671  2.0  242.0   \n",
      "2  0.02729   0.0   7.07   0.0  0.469  7.185  61.1  4.9671  2.0  242.0   \n",
      "3  0.03237   0.0   2.18   0.0  0.458  6.998  45.8  6.0622  3.0  222.0   \n",
      "4  0.06905   0.0   2.18   0.0  0.458  7.147  54.2  6.0622  3.0  222.0   \n",
      "\n",
      "   PTRATIO       B  LSTAT  \n",
      "0     15.3  396.90   4.98  \n",
      "1     17.8  396.90   9.14  \n",
      "2     17.8  392.83   4.03  \n",
      "3     18.7  394.63   2.94  \n",
      "4     18.7  396.90   5.33  \n"
     ]
    }
   ],
   "source": [
    "%matplotlib inline\n",
    "from sklearn import datasets\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "from sklearn.datasets import load_boston\n",
    "import seaborn as sns\n",
    "\n",
    "boston = load_boston()\n",
    "print(\"資料包內含資料的key值：\", boston.keys())\n",
    "print(\"資料欄位名稱(資料特徵參數)：\",boston.feature_names)\n",
    "print(\"檔案位置:\",boston['filename'])\n",
    "print(\"第一筆資料:\",boston.data[:1])\n",
    "print(\"資料型別:\",type(boston.data))\n",
    "\n",
    "\n",
    "target_names = boston.target\n",
    "print(\"資料size:\",boston['data'].size)\n",
    "print(\"前10筆target_names(房價): \", boston['target'][:10])\n",
    "#print(boston.target)\n",
    "\n",
    "\n",
    "#組成 pandas DataFrame (資料表的方式顯示)\n",
    "x = pd.DataFrame(boston['data'], columns=boston['feature_names'])\n",
    "print(\"pandas DataFrame\\n\",x.head(5),sep='')\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 實戰2\n",
    "### 切割數據集3:1比例分割為(訓練集:測試集)<br>\n",
    "印出如下資料<br>\n",
    "x_train的shape：(379, 13)<br>\n",
    "y_train的shape：(379, 1)<br>\n",
    "x_test的shape：(127, 13)<br>\n",
    "y_test 的shape：(127, 1)<br>\n",
    "\n",
    "##### train_test_split 模組內 random_state功能"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "x_train的shape : (379, 13)\n",
      "y_train的shape : (379,)\n",
      "x_test的shape : (127, 13)\n",
      "y_test的shape : (127,)\n",
      "=======no assign random_state=========\n",
      "        CRIM   ZN  INDUS  CHAS    NOX     RM   AGE     DIS  RAD    TAX  \\\n",
      "174  0.08447  0.0   4.05   0.0  0.510  5.859  68.7  2.7019  5.0  296.0   \n",
      "315  0.25356  0.0   9.90   0.0  0.544  5.705  77.7  3.9450  4.0  304.0   \n",
      "\n",
      "     PTRATIO       B  LSTAT  \n",
      "174     16.6  393.23   9.64  \n",
      "315     18.4  396.42  11.50  \n",
      "==========assign random_state=1============\n",
      "        CRIM   ZN  INDUS  CHAS    NOX     RM   AGE     DIS  RAD    TAX  \\\n",
      "502  0.04527  0.0  11.93   0.0  0.573  6.120  76.7  2.2875  1.0  273.0   \n",
      "172  0.13914  0.0   4.05   0.0  0.510  5.572  88.5  2.5961  5.0  296.0   \n",
      "\n",
      "     PTRATIO      B  LSTAT  \n",
      "502     21.0  396.9   9.08  \n",
      "172     16.6  396.9  14.69  \n"
     ]
    }
   ],
   "source": [
    "#講數據集1:3比例分割為 測試集：訓練集\n",
    "from sklearn.model_selection import train_test_split\n",
    "y =  boston['target']\n",
    "x_train, x_test, y_train, y_test = train_test_split(x, y)\n",
    "\n",
    "print(\"x_train的shape :\", x_train.shape)\n",
    "print(\"y_train的shape :\", y_train.shape)\n",
    "print(\"x_test的shape :\", x_test.shape)\n",
    "print(\"y_test的shape :\", y_test.shape)\n",
    "\n",
    "print(\"=======no assign random_state=========\")\n",
    "print(x_train[:2])\n",
    "\n",
    "print(\"==========assign random_state=1============\")\n",
    "x_train, x_test, y_train, y_test = train_test_split(x, y, random_state=1)\n",
    "print(x_train[:2])\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 實戰3進階實戰範\n",
    "## 線性回歸分析: Property value prediction\n",
    "### 此檔案使用scikit-learn 機器學習套件裡的linear regression演算法，來達成波士頓房地產價錢預測\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 步驟流程\n",
    "\n",
    "<br>* 切割數據集3:1比例分割為(訓練集:測試集)<br>\n",
    "x_train的shape：(379, 13)<br>\n",
    "y_train的shape：(379, 1)<br>\n",
    "x_verify的shape：(127, 13)<br>\n",
    "y_verify 的shape：(127, 1)<br>\n",
    "\n",
    "### 使用實戰2切割的數據完成以下\n",
    "\n",
    "<br>* 定義線性回歸模型，進行訓練、預測<br>\n",
    "from sklearn import linear_model<br>\n",
    "linear_model.LinearRegression()<br>\n",
    "\n",
    "<br>* 輸出模型參數、評價模型<br>\n",
    "此題主要使用線性迴歸<br>\n",
    "要解決的問題是如何量化地找出這些特徵值和目標值（即房價）的線性關係<br>\n",
    "MEDV = k1*CRIM + k2*ZN + … + k13*LITAT + b\n",
    "\n",
    "<br>* 獲取測試集房價的預測結果，歸納出最終結論\n",
    "xxx.predict(test_X)\n",
    "![預測模型](lr_out.png)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "y_test\n",
      " [[28.2]\n",
      " [23.9]\n",
      " [16.6]\n",
      " [22. ]\n",
      " [20.8]\n",
      " [23. ]\n",
      " [27.9]\n",
      " [14.5]\n",
      " [21.5]\n",
      " [22.6]\n",
      " [23.7]\n",
      " [31.2]\n",
      " [19.3]\n",
      " [19.4]\n",
      " [19.4]\n",
      " [27.9]\n",
      " [13.9]\n",
      " [50. ]\n",
      " [24.1]\n",
      " [14.6]\n",
      " [16.2]\n",
      " [15.6]\n",
      " [23.8]\n",
      " [25. ]\n",
      " [23.5]\n",
      " [ 8.3]\n",
      " [13.5]\n",
      " [17.5]\n",
      " [43.1]\n",
      " [11.5]\n",
      " [24.1]\n",
      " [18.5]\n",
      " [50. ]\n",
      " [12.6]\n",
      " [19.8]\n",
      " [24.5]\n",
      " [14.9]\n",
      " [36.2]\n",
      " [11.9]\n",
      " [19.1]\n",
      " [22.6]\n",
      " [20.7]\n",
      " [30.1]\n",
      " [13.3]\n",
      " [14.6]\n",
      " [ 8.4]\n",
      " [50. ]\n",
      " [12.7]\n",
      " [25. ]\n",
      " [18.6]\n",
      " [29.8]\n",
      " [22.2]\n",
      " [28.7]\n",
      " [23.8]\n",
      " [ 8.1]\n",
      " [22.2]\n",
      " [ 6.3]\n",
      " [22.1]\n",
      " [17.5]\n",
      " [48.3]\n",
      " [16.7]\n",
      " [26.6]\n",
      " [ 8.5]\n",
      " [14.5]\n",
      " [23.7]\n",
      " [37.2]\n",
      " [41.7]\n",
      " [16.5]\n",
      " [21.7]\n",
      " [22.7]\n",
      " [23. ]\n",
      " [10.5]\n",
      " [21.9]\n",
      " [21. ]\n",
      " [20.4]\n",
      " [21.8]\n",
      " [50. ]\n",
      " [22. ]\n",
      " [23.3]\n",
      " [37.3]\n",
      " [18. ]\n",
      " [19.2]\n",
      " [34.9]\n",
      " [13.4]\n",
      " [22.9]\n",
      " [22.5]\n",
      " [13. ]\n",
      " [24.6]\n",
      " [18.3]\n",
      " [18.1]\n",
      " [23.9]\n",
      " [50. ]\n",
      " [13.6]\n",
      " [22.9]\n",
      " [10.9]\n",
      " [18.9]\n",
      " [22.4]\n",
      " [22.9]\n",
      " [44.8]\n",
      " [21.7]\n",
      " [10.2]\n",
      " [15.4]\n",
      " [25.3]\n",
      " [23.3]\n",
      " [ 7.2]\n",
      " [21.2]\n",
      " [11.7]\n",
      " [27. ]\n",
      " [29.6]\n",
      " [26.5]\n",
      " [43.5]\n",
      " [23.6]\n",
      " [11. ]\n",
      " [33.4]\n",
      " [36. ]\n",
      " [36.4]\n",
      " [19. ]\n",
      " [20.2]\n",
      " [34.9]\n",
      " [50. ]\n",
      " [19.3]\n",
      " [14.9]\n",
      " [26.6]\n",
      " [19.9]\n",
      " [24.8]\n",
      " [21.2]\n",
      " [23.9]]\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "斜率： [[-1.14428903e-01  5.71299780e-02  3.83002824e-02  2.42854641e+00\n",
      "  -2.12326236e+01  2.87723416e+00  6.91118094e-03 -1.47158266e+00\n",
      "   3.05784197e-01 -1.06750361e-02 -9.96138270e-01  6.27746234e-03\n",
      "  -5.57414427e-01]]\n",
      "截距： [45.19251539]\n",
      "MAE: 3.5748681261275412\n",
      "MSE: 21.897765396049422\n",
      "RMSE: 4.679504823808757\n",
      "0.7168057552393374\n",
      "0.7789410172622865\n"
     ]
    }
   ],
   "source": [
    "from sklearn import datasets\n",
    "from sklearn import linear_model\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn import metrics\n",
    "import os\n",
    "import matplotlib.pyplot as plt\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "\n",
    "\n",
    "'''----------load 數據集-----------'''\n",
    "dataset = datasets.load_boston()\n",
    "\n",
    "# x 訓練特征：['CRIM', 'ZN', 'INDUS', 'CHAS', 'NOX', 'RM', 'AGE', 'DIS','RAD', 'TAX', 'PTRATIO', 'B', 'LSTAT']\n",
    "x = dataset.data\n",
    "\n",
    "target = dataset.target\n",
    "#把label變為(?, 1)維度，為了使用下面的數據集合分割\n",
    "y = np.reshape(target,(len(target), 1))\n",
    "\n",
    "#將數據集1:3比例分割為 測試集：訓練集  設定random_state=1\n",
    "x_train, x_test, y_train, y_test = train_test_split(x, y, random_state=1)\n",
    "\n",
    "\n",
    "'''----------定義線性回歸模型，進行訓練、預測-----------'''\n",
    "lr = linear_model.LinearRegression()\n",
    "lr.fit(x_train,y_train)\n",
    "y_pred = lr.predict(x_test)\n",
    "\n",
    "print(\"y_test\\n\",y_test)\n",
    "\n",
    "'''----------圖形化預測結果-----------'''\n",
    "#只顯示前50個預測結果，太多的話看起來不直觀\n",
    "plt.xlim([0,50])\n",
    "plt.plot( range(len(y_test)), y_test, 'r', label='y_test')\n",
    "plt.plot( range(len(y_pred)), y_pred, 'g--', label='y_predict' )\n",
    "plt.title('sklearn: Linear Regression')\n",
    "plt.legend()\n",
    "plt.savefig('lr_out.png')\n",
    "plt.show()\n",
    "\n",
    "\n",
    "'''----------輸出模型參數、評價模型-----------'''\n",
    "# 使用metrics.XXX()函式、()內參數放入預測值與目標值 \n",
    "print(\"斜率：\",lr.coef_)\n",
    "print(\"截距：\",lr.intercept_)\n",
    "print(\"MAE:\",metrics.mean_absolute_error(y_pred,y_test))\n",
    "print(\"MSE:\",metrics.mean_squared_error(y_test,y_pred))\n",
    "print(\"RMSE:\",np.sqrt(metrics.mean_squared_error(y_test,y_pred)))\n",
    "\n",
    "#輸出模型對應R-Square\n",
    "print(lr.score(x_train,y_train))\n",
    "print(lr.score(x_test,y_test))\n",
    "\n",
    "\n",
    " "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# other\n",
    "* 機器學習方法：線性迴歸\n",
    "* 探討重點：10 等分的交叉驗証(10-fold Cross-Validation)來實際測試資料以及預測值的關係\n",
    "* 關鍵函式： sklearn.cross_validation.cross_val_predict\n",
    "\n",
    "X軸為回歸目標，Y軸為預測結果。\n",
    "### 並劃出一條斜率=1的理想曲線(用虛線標示)\n",
    "https://machine-learning-python.kspax.io/general_examplesmd/ex1_plotting_cross-validated_predictions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[[30.04900564]\n",
      " [24.73268691]\n",
      " [30.36234996]\n",
      " [28.32103823]\n",
      " [27.54679474]\n",
      " [24.95005201]\n",
      " [22.76828807]\n",
      " [18.96826752]\n",
      " [10.69099639]\n",
      " [18.38869103]\n",
      " [18.32600667]\n",
      " [21.2151136 ]\n",
      " [20.67482233]\n",
      " [19.40202976]\n",
      " [19.02252621]\n",
      " [19.19247283]\n",
      " [20.53108732]\n",
      " [16.56719839]\n",
      " [16.14556027]\n",
      " [18.27871067]\n",
      " [12.06183163]\n",
      " [17.36395791]\n",
      " [15.35782636]\n",
      " [13.30827225]\n",
      " [15.25002158]\n",
      " [13.00072901]\n",
      " [15.07032347]\n",
      " [14.23138876]\n",
      " [19.13481689]\n",
      " [20.50406407]\n",
      " [10.88711375]\n",
      " [17.71425893]\n",
      " [ 8.10328694]\n",
      " [13.88702754]\n",
      " [13.18400457]\n",
      " [23.76248399]\n",
      " [22.26475553]\n",
      " [23.09795323]\n",
      " [22.88603574]\n",
      " [31.66165856]\n",
      " [34.54324123]\n",
      " [27.90889033]\n",
      " [25.13715879]\n",
      " [24.49149475]\n",
      " [22.69695788]\n",
      " [21.9712375 ]\n",
      " [20.16998059]\n",
      " [17.4184761 ]\n",
      " [ 8.1817061 ]\n",
      " [16.73776047]\n",
      " [20.93195787]\n",
      " [24.23669423]\n",
      " [27.85360352]\n",
      " [24.36110652]\n",
      " [15.43299525]\n",
      " [31.19077366]\n",
      " [24.97280986]\n",
      " [33.54851394]\n",
      " [21.77252458]\n",
      " [21.24357737]\n",
      " [17.97092977]\n",
      " [18.56991958]\n",
      " [23.97640077]\n",
      " [22.23963354]\n",
      " [22.80526009]\n",
      " [31.05850231]\n",
      " [26.28564522]\n",
      " [21.38462876]\n",
      " [17.6842577 ]\n",
      " [21.04192909]\n",
      " [25.53176898]\n",
      " [22.17169278]\n",
      " [25.04701676]\n",
      " [24.41776731]\n",
      " [25.97565456]\n",
      " [24.36353678]\n",
      " [23.35292065]\n",
      " [23.83613546]\n",
      " [21.52460215]\n",
      " [22.96803958]\n",
      " [28.63544312]\n",
      " [27.23544792]\n",
      " [26.38802581]\n",
      " [25.44978353]\n",
      " [24.9436409 ]\n",
      " [27.9897942 ]\n",
      " [22.43863884]\n",
      " [26.33634473]\n",
      " [30.76905302]\n",
      " [30.88421162]\n",
      " [27.39210868]\n",
      " [27.70851763]\n",
      " [29.67834392]\n",
      " [29.93942338]\n",
      " [27.80805505]\n",
      " [28.92497323]\n",
      " [25.08616066]\n",
      " [35.60743988]\n",
      " [35.03056994]\n",
      " [32.26519023]\n",
      " [24.59746028]\n",
      " [25.61811391]\n",
      " [20.54498733]\n",
      " [20.57449973]\n",
      " [21.7996183 ]\n",
      " [18.77705636]\n",
      " [17.24648879]\n",
      " [20.99588646]\n",
      " [23.02607826]\n",
      " [19.93892566]\n",
      " [20.64399332]\n",
      " [26.83479438]\n",
      " [20.95771908]\n",
      " [20.82817802]\n",
      " [25.58833952]\n",
      " [20.65250441]\n",
      " [23.57885385]\n",
      " [24.10559215]\n",
      " [20.46709801]\n",
      " [20.90617611]\n",
      " [19.8006881 ]\n",
      " [20.48096767]\n",
      " [18.43437449]\n",
      " [13.86123645]\n",
      " [18.49955851]\n",
      " [20.5013552 ]\n",
      " [12.04977924]\n",
      " [14.08132943]\n",
      " [17.85188314]\n",
      " [12.8728536 ]\n",
      " [19.09933422]\n",
      " [18.49759328]\n",
      " [19.22074767]\n",
      " [14.73373173]\n",
      " [12.28778723]\n",
      " [16.07879217]\n",
      " [14.74521644]\n",
      " [18.31747433]\n",
      " [12.46813013]\n",
      " [15.22128782]\n",
      " [11.9723311 ]\n",
      " [ 2.01042089]\n",
      " [12.44330743]\n",
      " [ 9.59436745]\n",
      " [ 6.06913683]\n",
      " [ 9.55278059]\n",
      " [14.08660303]\n",
      " [ 5.81846075]\n",
      " [ 7.07510619]\n",
      " [12.5282049 ]\n",
      " [18.91076553]\n",
      " [16.5955447 ]\n",
      " [18.85580574]\n",
      " [17.835508  ]\n",
      " [22.83494276]\n",
      " [20.80440765]\n",
      " [14.70952463]\n",
      " [32.33195457]\n",
      " [28.76442348]\n",
      " [25.44022843]\n",
      " [32.6152015 ]\n",
      " [35.43857014]\n",
      " [39.24588355]\n",
      " [40.10960644]\n",
      " [24.68933831]\n",
      " [25.24646723]\n",
      " [35.5266262 ]\n",
      " [23.16696205]\n",
      " [26.01851192]\n",
      " [26.18515728]\n",
      " [22.54958068]\n",
      " [24.21369333]\n",
      " [23.26606489]\n",
      " [28.74461549]\n",
      " [26.6403928 ]\n",
      " [30.40333829]\n",
      " [25.71577427]\n",
      " [28.97724731]\n",
      " [30.81563009]\n",
      " [32.17649812]\n",
      " [33.33917192]\n",
      " [27.61827372]\n",
      " [32.97251754]\n",
      " [30.53865679]\n",
      " [22.99597472]\n",
      " [24.6762538 ]\n",
      " [34.53459219]\n",
      " [32.70058758]\n",
      " [31.97251246]\n",
      " [33.56647635]\n",
      " [30.22966208]\n",
      " [29.90244642]\n",
      " [32.20406332]\n",
      " [31.50427204]\n",
      " [31.11539078]\n",
      " [39.38159477]\n",
      " [35.23243695]\n",
      " [31.94019606]\n",
      " [33.81185049]\n",
      " [29.41395865]\n",
      " [29.85692841]\n",
      " [29.12282692]\n",
      " [35.82556053]\n",
      " [40.40077662]\n",
      " [42.46383491]\n",
      " [21.92026858]\n",
      " [23.09558492]\n",
      " [17.84442996]\n",
      " [24.89013237]\n",
      " [19.50515806]\n",
      " [24.44211471]\n",
      " [19.24352019]\n",
      " [24.22696124]\n",
      " [24.28315961]\n",
      " [ 9.69896227]\n",
      " [23.94816292]\n",
      " [28.13659582]\n",
      " [28.1787416 ]\n",
      " [26.95619142]\n",
      " [31.63819302]\n",
      " [34.84252574]\n",
      " [25.64260822]\n",
      " [33.61278817]\n",
      " [29.86289333]\n",
      " [37.11534049]\n",
      " [38.20778365]\n",
      " [36.73417889]\n",
      " [32.0821312 ]\n",
      " [33.4819358 ]\n",
      " [30.4047625 ]\n",
      " [24.73815198]\n",
      " [32.70321394]\n",
      " [36.641163  ]\n",
      " [35.72577755]\n",
      " [33.1897742 ]\n",
      " [25.32027685]\n",
      " [31.74687113]\n",
      " [32.0561082 ]\n",
      " [27.88225248]\n",
      " [28.19827488]\n",
      " [26.83282997]\n",
      " [24.19256614]\n",
      " [24.1100325 ]\n",
      " [26.73166353]\n",
      " [17.12407013]\n",
      " [13.75824466]\n",
      " [19.63212155]\n",
      " [20.22992569]\n",
      " [20.89823078]\n",
      " [22.91279729]\n",
      " [23.21671396]\n",
      " [24.13886425]\n",
      " [23.40170105]\n",
      " [27.24595855]\n",
      " [24.44514537]\n",
      " [22.81505488]\n",
      " [37.85986454]\n",
      " [41.4560564 ]\n",
      " [35.37629481]\n",
      " [34.19081432]\n",
      " [33.7907225 ]\n",
      " [35.9822857 ]\n",
      " [39.30042267]\n",
      " [33.28580714]\n",
      " [34.81656876]\n",
      " [28.23985707]\n",
      " [30.21384255]\n",
      " [39.40152275]\n",
      " [38.49206186]\n",
      " [26.42411078]\n",
      " [22.79227462]\n",
      " [27.64869735]\n",
      " [28.6504771 ]\n",
      " [35.29208637]\n",
      " [36.75870886]\n",
      " [34.05020372]\n",
      " [35.87127843]\n",
      " [35.45358822]\n",
      " [30.77427703]\n",
      " [35.23550069]\n",
      " [38.06229503]\n",
      " [34.15733682]\n",
      " [40.09106184]\n",
      " [44.91228788]\n",
      " [32.03270734]\n",
      " [27.82170278]\n",
      " [20.95693276]\n",
      " [27.77981528]\n",
      " [27.84291441]\n",
      " [27.42280453]\n",
      " [34.17431693]\n",
      " [34.96620505]\n",
      " [32.69543199]\n",
      " [26.26784473]\n",
      " [24.85513928]\n",
      " [28.58969344]\n",
      " [27.50178731]\n",
      " [19.95083866]\n",
      " [29.86025509]\n",
      " [32.26945989]\n",
      " [31.1055611 ]\n",
      " [29.16294461]\n",
      " [29.23469909]\n",
      " [32.93266015]\n",
      " [32.97646474]\n",
      " [30.82615846]\n",
      " [35.95236603]\n",
      " [33.16399768]\n",
      " [28.96035552]\n",
      " [24.00743779]\n",
      " [19.32405524]\n",
      " [27.50044098]\n",
      " [23.68150148]\n",
      " [25.88949428]\n",
      " [25.67115786]\n",
      " [20.79581517]\n",
      " [17.7188336 ]\n",
      " [18.56258257]\n",
      " [24.57581235]\n",
      " [21.54864965]\n",
      " [25.11376967]\n",
      " [25.10218127]\n",
      " [23.12006866]\n",
      " [19.62256689]\n",
      " [25.36093764]\n",
      " [24.855056  ]\n",
      " [23.82718617]\n",
      " [19.38893597]\n",
      " [21.53018783]\n",
      " [24.59546518]\n",
      " [21.75220184]\n",
      " [19.8930345 ]\n",
      " [23.2683596 ]\n",
      " [22.06012621]\n",
      " [21.4601922 ]\n",
      " [20.66634445]\n",
      " [20.33563742]\n",
      " [19.31982621]\n",
      " [22.44677148]\n",
      " [21.49924594]\n",
      " [21.64530966]\n",
      " [30.06709052]\n",
      " [21.95683724]\n",
      " [27.49771111]\n",
      " [28.27740675]\n",
      " [16.15895821]\n",
      " [14.37287343]\n",
      " [24.51909275]\n",
      " [26.92811236]\n",
      " [21.41789802]\n",
      " [19.76857097]\n",
      " [19.42258961]\n",
      " [15.89723391]\n",
      " [23.9995477 ]\n",
      " [13.19641823]\n",
      " [15.46206389]\n",
      " [18.21317515]\n",
      " [19.98457763]\n",
      " [18.50646989]\n",
      " [17.29949585]\n",
      " [19.68574464]\n",
      " [17.23594933]\n",
      " [11.81639529]\n",
      " [15.51605966]\n",
      " [39.93278142]\n",
      " [ 0.13375416]\n",
      " [ 8.39099051]\n",
      " [-1.0605184 ]\n",
      " [12.48012705]\n",
      " [26.30973905]\n",
      " [29.38134831]\n",
      " [20.99131102]\n",
      " [19.28141782]\n",
      " [ 7.35474037]\n",
      " [ 1.67981585]\n",
      " [29.09156412]\n",
      " [21.87801276]\n",
      " [23.25081996]\n",
      " [20.77217946]\n",
      " [19.33734747]\n",
      " [30.63145189]\n",
      " [21.64186623]\n",
      " [12.85794168]\n",
      " [12.55706888]\n",
      " [ 2.43738905]\n",
      " [10.51735703]\n",
      " [ 6.68491687]\n",
      " [ 8.89261759]\n",
      " [ 6.78849629]\n",
      " [12.03536225]\n",
      " [14.768068  ]\n",
      " [16.92251747]\n",
      " [ 8.88269525]\n",
      " [19.20043954]\n",
      " [16.98275821]\n",
      " [20.69166773]\n",
      " [19.89566262]\n",
      " [14.92666055]\n",
      " [13.22010358]\n",
      " [14.27527525]\n",
      " [16.80420557]\n",
      " [19.79219002]\n",
      " [19.51865503]\n",
      " [13.23766944]\n",
      " [13.84663581]\n",
      " [17.98891496]\n",
      " [ 6.44247051]\n",
      " [20.05335382]\n",
      " [14.24409248]\n",
      " [21.60639748]\n",
      " [14.80657728]\n",
      " [19.52413889]\n",
      " [ 2.73382579]\n",
      " [11.21343895]\n",
      " [-4.95365512]\n",
      " [12.26341373]\n",
      " [16.78987771]\n",
      " [ 7.61506865]\n",
      " [ 4.96886063]\n",
      " [18.01492017]\n",
      " [20.75083624]\n",
      " [19.25347609]\n",
      " [18.88851017]\n",
      " [16.19948849]\n",
      " [16.99888493]\n",
      " [12.40078511]\n",
      " [18.62207302]\n",
      " [15.29966661]\n",
      " [16.989741  ]\n",
      " [15.9557087 ]\n",
      " [20.55383998]\n",
      " [21.3644639 ]\n",
      " [23.99883528]\n",
      " [20.0341374 ]\n",
      " [18.19931231]\n",
      " [16.28033669]\n",
      " [17.52332471]\n",
      " [11.75991801]\n",
      " [ 7.57784148]\n",
      " [13.50288515]\n",
      " [12.54721303]\n",
      " [18.31315948]\n",
      " [19.72269292]\n",
      " [19.07712976]\n",
      " [12.89942875]\n",
      " [15.20898649]\n",
      " [19.25945365]\n",
      " [18.9563459 ]\n",
      " [18.2052637 ]\n",
      " [18.79843626]\n",
      " [20.31082048]\n",
      " [20.96162054]\n",
      " [19.72853293]\n",
      " [24.28420152]\n",
      " [18.87207467]\n",
      " [19.30761644]\n",
      " [12.61877731]\n",
      " [12.80766548]\n",
      " [17.3222476 ]\n",
      " [18.76307482]\n",
      " [19.22014949]\n",
      " [20.44968245]\n",
      " [20.03452385]\n",
      " [22.87081258]\n",
      " [20.43190798]\n",
      " [17.75643918]\n",
      " [14.33852211]\n",
      " [17.03030978]\n",
      " [16.99159213]\n",
      " [18.56119351]\n",
      " [20.35444574]\n",
      " [23.32924524]\n",
      " [22.6055202 ]\n",
      " [25.72363254]\n",
      " [16.59087872]\n",
      " [16.11548223]\n",
      " [20.72379788]\n",
      " [11.56073693]\n",
      " [19.44044689]\n",
      " [22.16495312]\n",
      " [23.64646213]\n",
      " [27.46632887]\n",
      " [28.98710921]\n",
      " [21.04362395]\n",
      " [19.32198066]\n",
      " [22.29857058]\n",
      " [19.86251972]\n",
      " [21.3543872 ]\n",
      " [10.87258314]\n",
      " [ 7.08593616]\n",
      " [ 2.33607188]\n",
      " [12.83899018]\n",
      " [15.01627195]\n",
      " [20.33504447]\n",
      " [20.1622401 ]\n",
      " [16.18933129]\n",
      " [13.64581384]\n",
      " [18.93877025]\n",
      " [21.09426608]\n",
      " [18.26151226]\n",
      " [20.36806946]\n",
      " [23.65972148]\n",
      " [22.58655452]\n",
      " [28.08596609]\n",
      " [26.54976756]\n",
      " [22.63987499]]\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "from sklearn.model_selection import cross_val_predict\n",
    "dataset = datasets.load_boston()\n",
    "# x 訓練特征：['CRIM', 'ZN', 'INDUS', 'CHAS', 'NOX', 'RM', 'AGE', 'DIS','RAD', 'TAX', 'PTRATIO', 'B', 'LSTAT']\n",
    "x = dataset.data\n",
    "\n",
    "target = dataset.target\n",
    "#把label變為(?, 1)維度，為了使用下面的數據集合分割\n",
    "y = np.reshape(target,(len(target), 1))\n",
    "\n",
    "#將數據集1:3比例分割為 測試集：訓練集  設定random_state=1\n",
    "x_train, x_test, y_train, y_test = train_test_split(x, y, random_state=1)\n",
    "\n",
    "\n",
    "\n",
    "lr = linear_model.LinearRegression()\n",
    "predicted = cross_val_predict(lr, boston.data, y, cv=10)\n",
    "print(predicted)\n",
    "fig, ax = plt.subplots()\n",
    "ax.scatter(y, predicted)\n",
    "ax.plot([y.min(), y.max()], [y.min(), y.max()], 'k--', lw=2)\n",
    "ax.set_xlabel('Measured')\n",
    "ax.set_ylabel('Predicted')\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": "3.8.8"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}