FireAI_013_ClassifyIncome.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt\n",
    "%matplotlib inline\n",
    "np.random.seed(37) # 使得每次运行得到的随机数都一样"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "RangeIndex: 32561 entries, 0 to 32560\n",
      "Data columns (total 15 columns):\n",
      "0     32561 non-null int64\n",
      "1     32561 non-null object\n",
      "2     32561 non-null int64\n",
      "3     32561 non-null object\n",
      "4     32561 non-null int64\n",
      "5     32561 non-null object\n",
      "6     32561 non-null object\n",
      "7     32561 non-null object\n",
      "8     32561 non-null object\n",
      "9     32561 non-null object\n",
      "10    32561 non-null int64\n",
      "11    32561 non-null int64\n",
      "12    32561 non-null int64\n",
      "13    32561 non-null object\n",
      "14    32561 non-null object\n",
      "dtypes: int64(6), object(9)\n",
      "memory usage: 3.7+ MB\n",
      "None\n"
     ]
    }
   ],
   "source": [
    "# 准备数据集\n",
    "dataset_path='E:\\PyProjects\\DataSet\\CensusIncome/adult.data'\n",
    "df=pd.read_csv(dataset_path,header=None)\n",
    "print(df.info()) # 加载没有问题\n",
    "# 原数据集包含有32561个样本，每一个样本含有14个features, 一个label\n",
    "# print(df.head())\n",
    "raw_set=df.values"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "col-0 info: \n",
      "values: [17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40\n",
      " 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64\n",
      " 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88\n",
      " 90]\n",
      "values num: 73\n",
      "\n",
      "col-1 info: \n",
      "values: [' ?' ' Federal-gov' ' Local-gov' ' Never-worked' ' Private'\n",
      " ' Self-emp-inc' ' Self-emp-not-inc' ' State-gov' ' Without-pay']\n",
      "values num: 9\n",
      "\n",
      "col-2 info: \n",
      "values: [  12285   13769   14878 ... 1366120 1455435 1484705]\n",
      "values num: 21648\n",
      "\n",
      "col-3 info: \n",
      "values: [' 10th' ' 11th' ' 12th' ' 1st-4th' ' 5th-6th' ' 7th-8th' ' 9th'\n",
      " ' Assoc-acdm' ' Assoc-voc' ' Bachelors' ' Doctorate' ' HS-grad'\n",
      " ' Masters' ' Preschool' ' Prof-school' ' Some-college']\n",
      "values num: 16\n",
      "\n",
      "col-4 info: \n",
      "values: [ 1  2  3  4  5  6  7  8  9 10 11 12 13 14 15 16]\n",
      "values num: 16\n",
      "\n",
      "col-5 info: \n",
      "values: [' Divorced' ' Married-AF-spouse' ' Married-civ-spouse'\n",
      " ' Married-spouse-absent' ' Never-married' ' Separated' ' Widowed']\n",
      "values num: 7\n",
      "\n",
      "col-6 info: \n",
      "values: [' ?' ' Adm-clerical' ' Armed-Forces' ' Craft-repair' ' Exec-managerial'\n",
      " ' Farming-fishing' ' Handlers-cleaners' ' Machine-op-inspct'\n",
      " ' Other-service' ' Priv-house-serv' ' Prof-specialty' ' Protective-serv'\n",
      " ' Sales' ' Tech-support' ' Transport-moving']\n",
      "values num: 15\n",
      "\n",
      "col-7 info: \n",
      "values: [' Husband' ' Not-in-family' ' Other-relative' ' Own-child' ' Unmarried'\n",
      " ' Wife']\n",
      "values num: 6\n",
      "\n",
      "col-8 info: \n",
      "values: [' Amer-Indian-Eskimo' ' Asian-Pac-Islander' ' Black' ' Other' ' White']\n",
      "values num: 5\n",
      "\n",
      "col-9 info: \n",
      "values: [' Female' ' Male']\n",
      "values num: 2\n",
      "\n",
      "col-10 info: \n",
      "values: [    0   114   401   594   914   991  1055  1086  1111  1151  1173  1409\n",
      "  1424  1455  1471  1506  1639  1797  1831  1848  2009  2036  2050  2062\n",
      "  2105  2174  2176  2202  2228  2290  2329  2346  2354  2387  2407  2414\n",
      "  2463  2538  2580  2597  2635  2653  2829  2885  2907  2936  2961  2964\n",
      "  2977  2993  3103  3137  3273  3325  3411  3418  3432  3456  3464  3471\n",
      "  3674  3781  3818  3887  3908  3942  4064  4101  4386  4416  4508  4650\n",
      "  4687  4787  4865  4931  4934  5013  5060  5178  5455  5556  5721  6097\n",
      "  6360  6418  6497  6514  6723  6767  6849  7298  7430  7443  7688  7896\n",
      "  7978  8614  9386  9562 10520 10566 10605 11678 13550 14084 14344 15020\n",
      " 15024 15831 18481 20051 22040 25124 25236 27828 34095 41310 99999]\n",
      "values num: 119\n",
      "\n",
      "col-11 info: \n",
      "values: [   0  155  213  323  419  625  653  810  880  974 1092 1138 1258 1340\n",
      " 1380 1408 1411 1485 1504 1539 1564 1573 1579 1590 1594 1602 1617 1628\n",
      " 1648 1651 1668 1669 1672 1719 1721 1726 1735 1740 1741 1755 1762 1816\n",
      " 1825 1844 1848 1876 1887 1902 1944 1974 1977 1980 2001 2002 2042 2051\n",
      " 2057 2080 2129 2149 2163 2174 2179 2201 2205 2206 2231 2238 2246 2258\n",
      " 2267 2282 2339 2352 2377 2392 2415 2444 2457 2467 2472 2489 2547 2559\n",
      " 2603 2754 2824 3004 3683 3770 3900 4356]\n",
      "values num: 92\n",
      "\n",
      "col-12 info: \n",
      "values: [ 1  2  3  4  5  6  7  8  9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24\n",
      " 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48\n",
      " 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 70 72 73 74\n",
      " 75 76 77 78 80 81 82 84 85 86 87 88 89 90 91 92 94 95 96 97 98 99]\n",
      "values num: 94\n",
      "\n",
      "col-13 info: \n",
      "values: [' ?' ' Cambodia' ' Canada' ' China' ' Columbia' ' Cuba'\n",
      " ' Dominican-Republic' ' Ecuador' ' El-Salvador' ' England' ' France'\n",
      " ' Germany' ' Greece' ' Guatemala' ' Haiti' ' Holand-Netherlands'\n",
      " ' Honduras' ' Hong' ' Hungary' ' India' ' Iran' ' Ireland' ' Italy'\n",
      " ' Jamaica' ' Japan' ' Laos' ' Mexico' ' Nicaragua'\n",
      " ' Outlying-US(Guam-USVI-etc)' ' Peru' ' Philippines' ' Poland'\n",
      " ' Portugal' ' Puerto-Rico' ' Scotland' ' South' ' Taiwan' ' Thailand'\n",
      " ' Trinadad&Tobago' ' United-States' ' Vietnam' ' Yugoslavia']\n",
      "values num: 42\n",
      "\n",
      "col-14 info: \n",
      "values: [' <=50K' ' >50K']\n",
      "values num: 2\n"
     ]
    }
   ],
   "source": [
    "# 打印数据集各列的基本信息\n",
    "def print_col_info(dataset):\n",
    "    '''print info of every column in dataset:\n",
    "    detailed info includes:\n",
    "    1, values\n",
    "    2, value type num'''\n",
    "    col_num=dataset.shape[1]\n",
    "    for i in range(col_num):\n",
    "        print('\\ncol-{} info: '.format(i))\n",
    "        temp=np.sort(list(set(dataset[:,i])))\n",
    "        print('values: {}'.format(temp))\n",
    "        print('values num: {}'.format(temp.shape[0]))\n",
    "\n",
    "print_col_info(raw_set)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "# 数据处理一：去除字符串数值前面的空格\n",
    "str_cols=[1,3,5,6,7,8,9,13,14]\n",
    "for col in str_cols:\n",
    "    df.iloc[:,col]=df.iloc[:,col].map(lambda x: x.strip())\n",
    "\n",
    "# print_col_info(df.values) # 检查发现所有的字符串列都已经掉了空格"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "# 数据处理二： 删除缺失值样本\n",
    "# 将?字符串替换为NaN缺失值标志\n",
    "df.replace(\"?\",np.nan,inplace=True)\n",
    "# 此处直接删除缺失值样本\n",
    "df.dropna(inplace=True)\n",
    "# print(df2.shape) # (30162, 15)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "col-0 info: \n",
      "values: [17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34.\n",
      " 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49. 50. 51. 52.\n",
      " 53. 54. 55. 56. 57. 58. 59. 60. 61. 62. 63. 64. 65. 66. 67. 68. 69. 70.\n",
      " 71. 72. 73. 74. 75. 76. 77. 78. 79. 80. 81. 82. 83. 84. 85. 86. 88. 90.]\n",
      "values num: 72\n",
      "\n",
      "col-1 info: \n",
      "values: [0. 1. 2. 3. 4. 5. 6.]\n",
      "values num: 7\n",
      "\n",
      "col-2 info: \n",
      "values: [  13769.   14878.   18827. ... 1366120. 1455435. 1484705.]\n",
      "values num: 20263\n",
      "\n",
      "col-3 info: \n",
      "values: [ 0.  1.  2.  3.  4.  5.  6.  7.  8.  9. 10. 11. 12. 13. 14. 15.]\n",
      "values num: 16\n",
      "\n",
      "col-4 info: \n",
      "values: [ 1.  2.  3.  4.  5.  6.  7.  8.  9. 10. 11. 12. 13. 14. 15. 16.]\n",
      "values num: 16\n",
      "\n",
      "col-5 info: \n",
      "values: [0. 1. 2. 3. 4. 5. 6.]\n",
      "values num: 7\n",
      "\n",
      "col-6 info: \n",
      "values: [ 0.  1.  2.  3.  4.  5.  6.  7.  8.  9. 10. 11. 12. 13.]\n",
      "values num: 14\n",
      "\n",
      "col-7 info: \n",
      "values: [0. 1. 2. 3. 4. 5.]\n",
      "values num: 6\n",
      "\n",
      "col-8 info: \n",
      "values: [0. 1. 2. 3. 4.]\n",
      "values num: 5\n",
      "\n",
      "col-9 info: \n",
      "values: [0. 1.]\n",
      "values num: 2\n",
      "\n",
      "col-10 info: \n",
      "values: [    0.   114.   401.   594.   914.   991.  1055.  1086.  1151.  1173.\n",
      "  1409.  1424.  1455.  1471.  1506.  1639.  1797.  1831.  1848.  2009.\n",
      "  2036.  2050.  2062.  2105.  2174.  2176.  2202.  2228.  2290.  2329.\n",
      "  2346.  2354.  2387.  2407.  2414.  2463.  2538.  2580.  2597.  2635.\n",
      "  2653.  2829.  2885.  2907.  2936.  2961.  2964.  2977.  2993.  3103.\n",
      "  3137.  3273.  3325.  3411.  3418.  3432.  3456.  3464.  3471.  3674.\n",
      "  3781.  3818.  3887.  3908.  3942.  4064.  4101.  4386.  4416.  4508.\n",
      "  4650.  4687.  4787.  4865.  4931.  4934.  5013.  5060.  5178.  5455.\n",
      "  5556.  5721.  6097.  6360.  6418.  6497.  6514.  6723.  6767.  6849.\n",
      "  7298.  7430.  7443.  7688.  7896.  7978.  8614.  9386.  9562. 10520.\n",
      " 10566. 10605. 11678. 13550. 14084. 14344. 15020. 15024. 15831. 18481.\n",
      " 20051. 22040. 25124. 25236. 27828. 34095. 41310. 99999.]\n",
      "values num: 118\n",
      "\n",
      "col-11 info: \n",
      "values: [   0.  155.  213.  323.  419.  625.  653.  810.  880.  974. 1092. 1138.\n",
      " 1258. 1340. 1380. 1408. 1411. 1485. 1504. 1539. 1564. 1573. 1579. 1590.\n",
      " 1594. 1602. 1617. 1628. 1648. 1651. 1668. 1669. 1672. 1719. 1721. 1726.\n",
      " 1735. 1740. 1741. 1755. 1762. 1816. 1825. 1844. 1848. 1876. 1887. 1902.\n",
      " 1944. 1974. 1977. 1980. 2001. 2002. 2042. 2051. 2057. 2080. 2129. 2149.\n",
      " 2174. 2179. 2201. 2205. 2206. 2231. 2238. 2246. 2258. 2267. 2282. 2339.\n",
      " 2352. 2377. 2392. 2415. 2444. 2457. 2467. 2472. 2547. 2559. 2603. 2754.\n",
      " 2824. 3004. 3683. 3770. 3900. 4356.]\n",
      "values num: 90\n",
      "\n",
      "col-12 info: \n",
      "values: [ 1.  2.  3.  4.  5.  6.  7.  8.  9. 10. 11. 12. 13. 14. 15. 16. 17. 18.\n",
      " 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36.\n",
      " 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49. 50. 51. 52. 53. 54.\n",
      " 55. 56. 57. 58. 59. 60. 61. 62. 63. 64. 65. 66. 67. 68. 70. 72. 73. 74.\n",
      " 75. 76. 77. 78. 80. 81. 82. 84. 85. 86. 87. 88. 89. 90. 91. 92. 94. 95.\n",
      " 96. 97. 98. 99.]\n",
      "values num: 94\n",
      "\n",
      "col-13 info: \n",
      "values: [ 0.  1.  2.  3.  4.  5.  6.  7.  8.  9. 10. 11. 12. 13. 14. 15. 16. 17.\n",
      " 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35.\n",
      " 36. 37. 38. 39. 40.]\n",
      "values num: 41\n",
      "\n",
      "col-14 info: \n",
      "values: [0. 1.]\n",
      "values num: 2\n"
     ]
    }
   ],
   "source": [
    "# 数据处理三：对字符数据进行编码\n",
    "from sklearn import preprocessing\n",
    "label_encoder=[] # 放置每一列的encoder\n",
    "encoded_set = np.empty(df.shape)\n",
    "for col in range(df.shape[1]):\n",
    "    encoder=None\n",
    "    if df.iloc[:,col].dtype==object: # 字符型数据\n",
    "        encoder=preprocessing.LabelEncoder()\n",
    "        encoded_set[:,col]=encoder.fit_transform(df.iloc[:,col])\n",
    "    else:  # 数值型数据\n",
    "        encoded_set[:,col]=df.iloc[:,col]\n",
    "    label_encoder.append(encoder)\n",
    "\n",
    "print_col_info(encoded_set) # 全都是数字，没有问题"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "# 数据处理四：对某些列进行范围缩放\n",
    "# print(encoded_set.dtype) # float64 没问题\n",
    "\n",
    "cols=[2,10,11]\n",
    "data_scalers=[] # 专门用来放置scaler\n",
    "for col in cols:\n",
    "    data_scaler=preprocessing.MinMaxScaler(feature_range=(-1,1)) \n",
    "    encoded_set[:,col]=np.ravel(data_scaler.fit_transform(encoded_set[:,col].reshape(-1,1)))\n",
    "    data_scalers.append(data_scaler)\n",
    "    \n",
    "# print_col_info(encoded_set) # 已经发生了改变，没问题"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "dataset_X,dataset_y=encoded_set[:,:-1],encoded_set[:,-1]\n",
    "# 数据处理五：拆分数据集为train set和test set\n",
    "from sklearn.model_selection import train_test_split\n",
    "train_X, test_X, train_y, test_y=train_test_split(dataset_X,dataset_y,\n",
    "                                                  test_size=0.3,random_state=42)\n",
    "\n",
    "# print(dataset_X.shape) # (30162, 14)\n",
    "# print(dataset_y.shape) # (30162,)\n",
    "# print(train_X.shape) # (21113, 14)\n",
    "# print(train_y.shape) # (21113,)\n",
    "# print(test_X.shape) # (9049, 14)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "准确率：79.84%\n",
      "精确度：78.45%\n",
      "召回率：79.84%\n",
      "F1  值：77.21%\n",
      "[[6420  347]\n",
      " [1518  764]]\n",
      "             precision    recall  f1-score   support\n",
      "\n",
      "      <=50K       0.81      0.95      0.87      6767\n",
      "       >50K       0.69      0.33      0.45      2282\n",
      "\n",
      "avg / total       0.78      0.79      0.77      9049\n",
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "e:\\programsoft\\python\\lib\\site-packages\\sklearn\\cross_validation.py:41: DeprecationWarning: This module was deprecated in version 0.18 in favor of the model_selection module into which all the refactored classes and functions are moved. Also note that the interface of the new CV iterators are different from that of this module. This module will be removed in 0.20.\n",
      "  \"This module will be removed in 0.20.\", DeprecationWarning)\n"
     ]
    }
   ],
   "source": [
    "# 建立朴素贝叶斯分类器模型\n",
    "from sklearn.naive_bayes import GaussianNB\n",
    "gaussianNB=GaussianNB()\n",
    "gaussianNB.fit(train_X,train_y)\n",
    "\n",
    "# 2 用交叉验证来检验模型的准确性，只是在test set上验证准确性\n",
    "from sklearn.cross_validation import cross_val_score\n",
    "num_validations=5\n",
    "accuracy=cross_val_score(gaussianNB,test_X,test_y,\n",
    "                         scoring='accuracy',cv=num_validations)\n",
    "print('准确率：{:.2f}%'.format(accuracy.mean()*100))\n",
    "precision=cross_val_score(gaussianNB,test_X,test_y,\n",
    "                         scoring='precision_weighted',cv=num_validations)\n",
    "print('精确度：{:.2f}%'.format(precision.mean()*100))\n",
    "recall=cross_val_score(gaussianNB,test_X,test_y,\n",
    "                         scoring='recall_weighted',cv=num_validations)\n",
    "print('召回率：{:.2f}%'.format(recall.mean()*100))\n",
    "f1=cross_val_score(gaussianNB,test_X,test_y,\n",
    "                         scoring='f1_weighted',cv=num_validations)\n",
    "print('F1  值：{:.2f}%'.format(f1.mean()*100))\n",
    "                   \n",
    "# 3 打印性能报告\n",
    "from sklearn.metrics import confusion_matrix\n",
    "y_pred=gaussianNB.predict(test_X)\n",
    "confusion_mat = confusion_matrix(test_y, y_pred)\n",
    "print(confusion_mat) #看看混淆矩阵长啥样\n",
    "\n",
    "from sklearn.metrics import classification_report\n",
    "# 直接使用sklearn打印精度，召回率和F1值\n",
    "target_names = ['<=50K', '>50K']\n",
    "print(classification_report(test_y, y_pred, target_names=target_names))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[39.          5.         -0.91332458  9.         13.          4.\n",
      "  0.          1.          4.          1.         -0.95651957 -1.\n",
      " 40.         38.        ]\n",
      "output: [0.], class: <=50K\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "e:\\programsoft\\python\\lib\\site-packages\\sklearn\\preprocessing\\label.py:151: DeprecationWarning: The truth value of an empty array is ambiguous. Returning False, but in future this will result in an error. Use `array.size > 0` to check that an array is not empty.\n",
      "  if diff:\n"
     ]
    }
   ],
   "source": [
    "### 2.2 用训练好的模型来预测新样本\n",
    "new_sample=[39, 'State-gov', 77516, 'Bachelors', 13, \n",
    "            'Never-married', 'Adm-clerical', 'Not-in-family', \n",
    "            'White', 'Male', 2174, 0, 40, 'United-States'] \n",
    "# 先对新样本中的字符型数据编码\n",
    "encoded_sample=np.empty(np.array(new_sample).shape)\n",
    "for i,item in enumerate(new_sample):\n",
    "    if label_encoder[i] is not None:\n",
    "        encoded_sample[i]=float(label_encoder[i].transform([item]))\n",
    "    else:\n",
    "        encoded_sample[i]=item\n",
    "\n",
    "# print(encoded_sample) # 貌似没有错\n",
    "\n",
    "# 再对数值较大列进行范围缩放\n",
    "cols=[2,10,11]\n",
    "for i,col in enumerate(cols):\n",
    "    encoded_sample[col]=np.ravel(data_scalers[i].transform(encoded_sample[col].reshape(-1,1)))\n",
    "print(encoded_sample) # 检查没有错\n",
    "\n",
    "# 将新数据放入模型中进行预测\n",
    "output=gaussianNB.predict(encoded_sample.reshape(1,-1))\n",
    "print('output: {}, class: {}'.format(output,\n",
    "       label_encoder[-1].inverse_transform(int(output))))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "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": "3.6.6"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}