08. Tensorflow使用feature_column实现银行营销二分类.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "rNdWfPXCjTjY"
   },
   "source": [
    "## Tensorflow使用feature_column实现银行营销二分类"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 演示流程：\n",
    "\n",
    "1. 用Pandas导入 CSV 文件。\n",
    "2. 用tf.dataset读取数据，进行分批（batch）、随机排序（shuffle）处理\n",
    "3. 用feature_column将 CSV 中的列映射到用于训练模型的特征\n",
    "4. 用 Keras 构建，训练并评估模型"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "K1y4OHpGgss7"
   },
   "source": [
    "### 银行营销数据集\n",
    "\n",
    "数据来自葡萄牙银行机构电话营销活动的记录。\n",
    "\n",
    "每行是同一个客户的电话沟通记录，以及最后一列是这客户最后明确购买还是不购买产品，是一个二分类问题。\n",
    "\n",
    "数据列如下：\n",
    "01. 年龄，数字\n",
    "02. 工作类型：分类，比如管理员、企业家\n",
    "03. 婚姻状况：分类，比如已婚、离婚、单身\n",
    "04. 教育：分类，小学、中学等\n",
    "05. 是否有信用：分类、是/否\n",
    "06. 余额：数字、年均余额\n",
    "07. 是否有住房贷款：分类、是/否\n",
    "08. 是否有贷款：分类，是/否\n",
    "09. 联系方式：分类、位置、手机号码、固定号码\n",
    "10. 天：数字、最后一次联系日\n",
    "11. 月：数字、最后一次联系的月份\n",
    "12. 持续时间：数字、上次联系时间\n",
    "13. 广告系列：数字，在这个广告和用户联系的次数，包括其他人\n",
    "14. pdays：数字，和这个客户上次联系的间隔天数\n",
    "15. 以前：数字，这个广告此前和此客户的联系次数\n",
    "16. 以前的结果：分类、成功、失败\n",
    "17. 最终是否订阅了定期存款：二分类、是、否"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "VxyBFc_kKazA"
   },
   "source": [
    "### 1. 导入库"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "9dEreb4QKizj"
   },
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "\n",
    "import tensorflow as tf\n",
    "\n",
    "from tensorflow import feature_column\n",
    "from tensorflow.keras import layers\n",
    "from sklearn.model_selection import train_test_split"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "KCEhSZcULZ9n"
   },
   "source": [
    "### 2. 使用 Pandas读取CSV"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "REZ57BXCLdfG"
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>age</th>\n",
       "      <th>job</th>\n",
       "      <th>marital</th>\n",
       "      <th>education</th>\n",
       "      <th>default</th>\n",
       "      <th>balance</th>\n",
       "      <th>housing</th>\n",
       "      <th>loan</th>\n",
       "      <th>contact</th>\n",
       "      <th>day</th>\n",
       "      <th>month</th>\n",
       "      <th>duration</th>\n",
       "      <th>campaign</th>\n",
       "      <th>pdays</th>\n",
       "      <th>previous</th>\n",
       "      <th>poutcome</th>\n",
       "      <th>y</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>58</td>\n",
       "      <td>management</td>\n",
       "      <td>married</td>\n",
       "      <td>tertiary</td>\n",
       "      <td>no</td>\n",
       "      <td>2143</td>\n",
       "      <td>yes</td>\n",
       "      <td>no</td>\n",
       "      <td>unknown</td>\n",
       "      <td>5</td>\n",
       "      <td>may</td>\n",
       "      <td>261</td>\n",
       "      <td>1</td>\n",
       "      <td>-1</td>\n",
       "      <td>0</td>\n",
       "      <td>unknown</td>\n",
       "      <td>no</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>44</td>\n",
       "      <td>technician</td>\n",
       "      <td>single</td>\n",
       "      <td>secondary</td>\n",
       "      <td>no</td>\n",
       "      <td>29</td>\n",
       "      <td>yes</td>\n",
       "      <td>no</td>\n",
       "      <td>unknown</td>\n",
       "      <td>5</td>\n",
       "      <td>may</td>\n",
       "      <td>151</td>\n",
       "      <td>1</td>\n",
       "      <td>-1</td>\n",
       "      <td>0</td>\n",
       "      <td>unknown</td>\n",
       "      <td>no</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>33</td>\n",
       "      <td>entrepreneur</td>\n",
       "      <td>married</td>\n",
       "      <td>secondary</td>\n",
       "      <td>no</td>\n",
       "      <td>2</td>\n",
       "      <td>yes</td>\n",
       "      <td>yes</td>\n",
       "      <td>unknown</td>\n",
       "      <td>5</td>\n",
       "      <td>may</td>\n",
       "      <td>76</td>\n",
       "      <td>1</td>\n",
       "      <td>-1</td>\n",
       "      <td>0</td>\n",
       "      <td>unknown</td>\n",
       "      <td>no</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>47</td>\n",
       "      <td>blue-collar</td>\n",
       "      <td>married</td>\n",
       "      <td>unknown</td>\n",
       "      <td>no</td>\n",
       "      <td>1506</td>\n",
       "      <td>yes</td>\n",
       "      <td>no</td>\n",
       "      <td>unknown</td>\n",
       "      <td>5</td>\n",
       "      <td>may</td>\n",
       "      <td>92</td>\n",
       "      <td>1</td>\n",
       "      <td>-1</td>\n",
       "      <td>0</td>\n",
       "      <td>unknown</td>\n",
       "      <td>no</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>33</td>\n",
       "      <td>unknown</td>\n",
       "      <td>single</td>\n",
       "      <td>unknown</td>\n",
       "      <td>no</td>\n",
       "      <td>1</td>\n",
       "      <td>no</td>\n",
       "      <td>no</td>\n",
       "      <td>unknown</td>\n",
       "      <td>5</td>\n",
       "      <td>may</td>\n",
       "      <td>198</td>\n",
       "      <td>1</td>\n",
       "      <td>-1</td>\n",
       "      <td>0</td>\n",
       "      <td>unknown</td>\n",
       "      <td>no</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   age           job  marital  education default  balance housing loan  \\\n",
       "0   58    management  married   tertiary      no     2143     yes   no   \n",
       "1   44    technician   single  secondary      no       29     yes   no   \n",
       "2   33  entrepreneur  married  secondary      no        2     yes  yes   \n",
       "3   47   blue-collar  married    unknown      no     1506     yes   no   \n",
       "4   33       unknown   single    unknown      no        1      no   no   \n",
       "\n",
       "   contact  day month  duration  campaign  pdays  previous poutcome   y  \n",
       "0  unknown    5   may       261         1     -1         0  unknown  no  \n",
       "1  unknown    5   may       151         1     -1         0  unknown  no  \n",
       "2  unknown    5   may        76         1     -1         0  unknown  no  \n",
       "3  unknown    5   may        92         1     -1         0  unknown  no  \n",
       "4  unknown    5   may       198         1     -1         0  unknown  no  "
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df = pd.read_csv(\"./datas/bank/bank-full.csv\", sep=\";\")\n",
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(45211, 17)"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "RangeIndex: 45211 entries, 0 to 45210\n",
      "Data columns (total 17 columns):\n",
      " #   Column     Non-Null Count  Dtype \n",
      "---  ------     --------------  ----- \n",
      " 0   age        45211 non-null  int64 \n",
      " 1   job        45211 non-null  object\n",
      " 2   marital    45211 non-null  object\n",
      " 3   education  45211 non-null  object\n",
      " 4   default    45211 non-null  object\n",
      " 5   balance    45211 non-null  int64 \n",
      " 6   housing    45211 non-null  object\n",
      " 7   loan       45211 non-null  object\n",
      " 8   contact    45211 non-null  object\n",
      " 9   day        45211 non-null  int64 \n",
      " 10  month      45211 non-null  object\n",
      " 11  duration   45211 non-null  int64 \n",
      " 12  campaign   45211 non-null  int64 \n",
      " 13  pdays      45211 non-null  int64 \n",
      " 14  previous   45211 non-null  int64 \n",
      " 15  poutcome   45211 non-null  object\n",
      " 16  y          45211 non-null  object\n",
      "dtypes: int64(7), object(10)\n",
      "memory usage: 5.9+ MB\n"
     ]
    }
   ],
   "source": [
    "df.info()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array(['no', 'yes'], dtype=object)"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df[\"y\"].unique()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>age</th>\n",
       "      <th>job</th>\n",
       "      <th>marital</th>\n",
       "      <th>education</th>\n",
       "      <th>default</th>\n",
       "      <th>balance</th>\n",
       "      <th>housing</th>\n",
       "      <th>loan</th>\n",
       "      <th>contact</th>\n",
       "      <th>day</th>\n",
       "      <th>month</th>\n",
       "      <th>duration</th>\n",
       "      <th>campaign</th>\n",
       "      <th>pdays</th>\n",
       "      <th>previous</th>\n",
       "      <th>poutcome</th>\n",
       "      <th>y</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>58</td>\n",
       "      <td>management</td>\n",
       "      <td>married</td>\n",
       "      <td>tertiary</td>\n",
       "      <td>no</td>\n",
       "      <td>2143</td>\n",
       "      <td>yes</td>\n",
       "      <td>no</td>\n",
       "      <td>unknown</td>\n",
       "      <td>5</td>\n",
       "      <td>may</td>\n",
       "      <td>261</td>\n",
       "      <td>1</td>\n",
       "      <td>-1</td>\n",
       "      <td>0</td>\n",
       "      <td>unknown</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>44</td>\n",
       "      <td>technician</td>\n",
       "      <td>single</td>\n",
       "      <td>secondary</td>\n",
       "      <td>no</td>\n",
       "      <td>29</td>\n",
       "      <td>yes</td>\n",
       "      <td>no</td>\n",
       "      <td>unknown</td>\n",
       "      <td>5</td>\n",
       "      <td>may</td>\n",
       "      <td>151</td>\n",
       "      <td>1</td>\n",
       "      <td>-1</td>\n",
       "      <td>0</td>\n",
       "      <td>unknown</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>33</td>\n",
       "      <td>entrepreneur</td>\n",
       "      <td>married</td>\n",
       "      <td>secondary</td>\n",
       "      <td>no</td>\n",
       "      <td>2</td>\n",
       "      <td>yes</td>\n",
       "      <td>yes</td>\n",
       "      <td>unknown</td>\n",
       "      <td>5</td>\n",
       "      <td>may</td>\n",
       "      <td>76</td>\n",
       "      <td>1</td>\n",
       "      <td>-1</td>\n",
       "      <td>0</td>\n",
       "      <td>unknown</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>47</td>\n",
       "      <td>blue-collar</td>\n",
       "      <td>married</td>\n",
       "      <td>unknown</td>\n",
       "      <td>no</td>\n",
       "      <td>1506</td>\n",
       "      <td>yes</td>\n",
       "      <td>no</td>\n",
       "      <td>unknown</td>\n",
       "      <td>5</td>\n",
       "      <td>may</td>\n",
       "      <td>92</td>\n",
       "      <td>1</td>\n",
       "      <td>-1</td>\n",
       "      <td>0</td>\n",
       "      <td>unknown</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>33</td>\n",
       "      <td>unknown</td>\n",
       "      <td>single</td>\n",
       "      <td>unknown</td>\n",
       "      <td>no</td>\n",
       "      <td>1</td>\n",
       "      <td>no</td>\n",
       "      <td>no</td>\n",
       "      <td>unknown</td>\n",
       "      <td>5</td>\n",
       "      <td>may</td>\n",
       "      <td>198</td>\n",
       "      <td>1</td>\n",
       "      <td>-1</td>\n",
       "      <td>0</td>\n",
       "      <td>unknown</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   age           job  marital  education default  balance housing loan  \\\n",
       "0   58    management  married   tertiary      no     2143     yes   no   \n",
       "1   44    technician   single  secondary      no       29     yes   no   \n",
       "2   33  entrepreneur  married  secondary      no        2     yes  yes   \n",
       "3   47   blue-collar  married    unknown      no     1506     yes   no   \n",
       "4   33       unknown   single    unknown      no        1      no   no   \n",
       "\n",
       "   contact  day month  duration  campaign  pdays  previous poutcome  y  \n",
       "0  unknown    5   may       261         1     -1         0  unknown  0  \n",
       "1  unknown    5   may       151         1     -1         0  unknown  0  \n",
       "2  unknown    5   may        76         1     -1         0  unknown  0  \n",
       "3  unknown    5   may        92         1     -1         0  unknown  0  \n",
       "4  unknown    5   may       198         1     -1         0  unknown  0  "
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.loc[df[\"y\"]==\"yes\", \"y\"] = 1\n",
    "df.loc[df[\"y\"]==\"no\", \"y\"] = 0\n",
    "df[\"y\"] = df[\"y\"].astype(int)\n",
    "df.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "u0zhLtQqMPem"
   },
   "source": [
    "### 3. 将 dataframe 拆分为训练、验证和测试集"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "YEOpw7LhMYsI"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "28934 训练集数目\n",
      "7234 验证集数目\n",
      "9043 测试集数目\n"
     ]
    }
   ],
   "source": [
    "train_all, test = train_test_split(df, test_size=0.2)\n",
    "train, val = train_test_split(train_all, test_size=0.2)\n",
    "print(len(train), '训练集数目')\n",
    "print(len(val), '验证集数目')\n",
    "print(len(test), '测试集数目')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "84ef46LXMfvu"
   },
   "source": [
    "### 4. 用tf.data.Dataset读取数据"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "NkcaMYP-MsRe"
   },
   "outputs": [],
   "source": [
    "def df_to_dataset(df, shuffle=True, batch_size=32):\n",
    "    \"\"\"便捷函数，将pandas的df转换成dataset\"\"\"\n",
    "    labels = df.pop('y')\n",
    "    ds = tf.data.Dataset.from_tensor_slices((dict(df), labels))\n",
    "    if shuffle:\n",
    "        ds = ds.shuffle(buffer_size=len(df))\n",
    "    ds = ds.batch(batch_size)\n",
    "    return ds"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "CXbbXkJvMy34"
   },
   "outputs": [],
   "source": [
    "batch_size = 32\n",
    "train_ds = df_to_dataset(train, shuffle=True, batch_size=batch_size)\n",
    "val_ds = df_to_dataset(val, shuffle=False, batch_size=batch_size)\n",
    "test_ds = df_to_dataset(test, shuffle=False, batch_size=batch_size)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "ypkI9zx6Rj1q"
   },
   "source": [
    "### 5. feature_column特征处理\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "# 分类列名称列表\n",
    "category_names = [\"job\", \"marital\", \"education\", \"default\", \"housing\", \"loan\", \"contact\", \"month\", \"poutcome\"]\n",
    "\n",
    "# 数值列名称列表\n",
    "numeric_names = [\"balance\", \"duration\", \"campaign\", \"pdays\", \"previous\"]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "{'job': ['management', 'technician', 'entrepreneur', 'blue-collar', 'unknown', 'retired', 'admin.', 'services', 'self-employed', 'unemployed', 'housemaid', 'student'], 'marital': ['married', 'single', 'divorced'], 'education': ['tertiary', 'secondary', 'unknown', 'primary'], 'default': ['no', 'yes'], 'housing': ['yes', 'no'], 'loan': ['no', 'yes'], 'contact': ['unknown', 'cellular', 'telephone'], 'month': ['may', 'jun', 'jul', 'aug', 'oct', 'nov', 'dec', 'jan', 'feb', 'mar', 'apr', 'sep'], 'poutcome': ['unknown', 'failure', 'other', 'success']}\n"
     ]
    }
   ],
   "source": [
    "# 数据统计，分类列的枚举列表\n",
    "category_vocabulary_list = {}\n",
    "\n",
    "for feature_name in category_names:\n",
    "    category_vocabulary_list[feature_name] = list(df[feature_name].unique())\n",
    "\n",
    "print(category_vocabulary_list)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "4PlLY7fORuzA"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[IndicatorColumn(categorical_column=VocabularyListCategoricalColumn(key='job', vocabulary_list=('management', 'technician', 'entrepreneur', 'blue-collar', 'unknown', 'retired', 'admin.', 'services', 'self-employed', 'unemployed', 'housemaid', 'student'), dtype=tf.string, default_value=-1, num_oov_buckets=0)),\n",
       " IndicatorColumn(categorical_column=VocabularyListCategoricalColumn(key='marital', vocabulary_list=('married', 'single', 'divorced'), dtype=tf.string, default_value=-1, num_oov_buckets=0)),\n",
       " IndicatorColumn(categorical_column=VocabularyListCategoricalColumn(key='education', vocabulary_list=('tertiary', 'secondary', 'unknown', 'primary'), dtype=tf.string, default_value=-1, num_oov_buckets=0)),\n",
       " IndicatorColumn(categorical_column=VocabularyListCategoricalColumn(key='default', vocabulary_list=('no', 'yes'), dtype=tf.string, default_value=-1, num_oov_buckets=0)),\n",
       " IndicatorColumn(categorical_column=VocabularyListCategoricalColumn(key='housing', vocabulary_list=('yes', 'no'), dtype=tf.string, default_value=-1, num_oov_buckets=0)),\n",
       " IndicatorColumn(categorical_column=VocabularyListCategoricalColumn(key='loan', vocabulary_list=('no', 'yes'), dtype=tf.string, default_value=-1, num_oov_buckets=0)),\n",
       " IndicatorColumn(categorical_column=VocabularyListCategoricalColumn(key='contact', vocabulary_list=('unknown', 'cellular', 'telephone'), dtype=tf.string, default_value=-1, num_oov_buckets=0)),\n",
       " IndicatorColumn(categorical_column=VocabularyListCategoricalColumn(key='month', vocabulary_list=('may', 'jun', 'jul', 'aug', 'oct', 'nov', 'dec', 'jan', 'feb', 'mar', 'apr', 'sep'), dtype=tf.string, default_value=-1, num_oov_buckets=0)),\n",
       " IndicatorColumn(categorical_column=VocabularyListCategoricalColumn(key='poutcome', vocabulary_list=('unknown', 'failure', 'other', 'success'), dtype=tf.string, default_value=-1, num_oov_buckets=0)),\n",
       " NumericColumn(key='balance', shape=(1,), default_value=None, dtype=tf.float32, normalizer_fn=None),\n",
       " NumericColumn(key='duration', shape=(1,), default_value=None, dtype=tf.float32, normalizer_fn=None),\n",
       " NumericColumn(key='campaign', shape=(1,), default_value=None, dtype=tf.float32, normalizer_fn=None),\n",
       " NumericColumn(key='pdays', shape=(1,), default_value=None, dtype=tf.float32, normalizer_fn=None),\n",
       " NumericColumn(key='previous', shape=(1,), default_value=None, dtype=tf.float32, normalizer_fn=None)]"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "feature_columns = []\n",
    "\n",
    "for feature_name in category_names:\n",
    "    feature_columns.append(\n",
    "        feature_column.indicator_column(\n",
    "            feature_column.categorical_column_with_vocabulary_list(feature_name, category_vocabulary_list[feature_name])\n",
    "        )\n",
    "    )\n",
    "    \n",
    "for feature_name in numeric_names:\n",
    "    feature_columns.append(feature_column.numeric_column(feature_name))\n",
    "\n",
    "feature_columns"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "bBx4Xu0eTXWq"
   },
   "source": [
    "### 6. 创建，编译和训练模型"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "_YJPPb3xTPeZ"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/10\n",
      "905/905 [==============================] - 3s 3ms/step - loss: 1.9766 - accuracy: 0.8191 - val_loss: 0.8286 - val_accuracy: 0.8814\n",
      "Epoch 2/10\n",
      "905/905 [==============================] - 2s 2ms/step - loss: 0.6598 - accuracy: 0.8591 - val_loss: 0.3908 - val_accuracy: 0.8839\n",
      "Epoch 3/10\n",
      "905/905 [==============================] - 2s 2ms/step - loss: 0.4846 - accuracy: 0.8709 - val_loss: 0.2884 - val_accuracy: 0.8846\n",
      "Epoch 4/10\n",
      "905/905 [==============================] - 2s 2ms/step - loss: 0.3524 - accuracy: 0.8788 - val_loss: 0.2737 - val_accuracy: 0.8850\n",
      "Epoch 5/10\n",
      "905/905 [==============================] - 2s 2ms/step - loss: 0.3455 - accuracy: 0.8800 - val_loss: 0.3289 - val_accuracy: 0.8763\n",
      "Epoch 6/10\n",
      "905/905 [==============================] - 2s 2ms/step - loss: 0.3051 - accuracy: 0.8841 - val_loss: 0.2601 - val_accuracy: 0.8853\n",
      "Epoch 7/10\n",
      "905/905 [==============================] - 2s 3ms/step - loss: 0.2886 - accuracy: 0.8865 - val_loss: 0.3332 - val_accuracy: 0.8829\n",
      "Epoch 8/10\n",
      "905/905 [==============================] - 2s 2ms/step - loss: 0.2866 - accuracy: 0.8873 - val_loss: 0.2727 - val_accuracy: 0.8818\n",
      "Epoch 9/10\n",
      "905/905 [==============================] - 2s 2ms/step - loss: 0.2735 - accuracy: 0.8885 - val_loss: 0.2752 - val_accuracy: 0.8831\n",
      "Epoch 10/10\n",
      "905/905 [==============================] - 2s 2ms/step - loss: 0.2706 - accuracy: 0.8865 - val_loss: 0.2686 - val_accuracy: 0.8831\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<tensorflow.python.keras.callbacks.History at 0x7f3a80c3a1d0>"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model = tf.keras.Sequential([\n",
    "    layers.DenseFeatures(feature_columns),\n",
    "    layers.Dense(128, activation='relu'),\n",
    "    layers.Dense(64, activation='relu'),\n",
    "    layers.Dense(32, activation='relu'),\n",
    "    layers.Dense(1, activation='sigmoid')\n",
    "])\n",
    "\n",
    "model.compile(optimizer=\"adam\",\n",
    "              loss='binary_crossentropy',\n",
    "              metrics=['accuracy'])\n",
    "\n",
    "model.fit(train_ds,\n",
    "          validation_data=val_ds,\n",
    "          epochs=10)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "GnFmMOW0Tcaa"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "283/283 [==============================] - 0s 2ms/step - loss: 0.2584 - accuracy: 0.8899\n",
      "Accuracy 0.8898595571517944\n"
     ]
    }
   ],
   "source": [
    "loss, accuracy = model.evaluate(test_ds)\n",
    "print(\"Accuracy\", accuracy)"
   ]
  }
 ],
 "metadata": {
  "colab": {
   "collapsed_sections": [],
   "name": "feature_columns.ipynb",
   "private_outputs": true,
   "provenance": [],
   "toc_visible": true,
   "version": "0.3.2"
  },
  "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": "3.7.6"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}