week4/1-1结构化数据建模流程.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 1-1 结构化数据建模流程\n",
    "## 数据集概述\n",
    "titanic数据集的目标是根据乘客信息预测他们再Titanic号撞击沉没后是否能生存\n",
    "## 结构化数据处理的方式\n",
    "结构化的数据一般使用Pandas进行处理"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 163,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np \n",
    "import pandas as pd \n",
    "import matplotlib.pyplot as plt\n",
    "import tensorflow as tf\n",
    "from tensorflow.keras import models,layers\n",
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 读入数据"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 字段说明\n",
    "* Survied：0代表死亡, 1代表存活 【y标签】\n",
    "* Pclass: 乘客所持的票类\n",
    "* Name: 乘客姓名【与预测无关】\n",
    "* Sex: 性别\n",
    "* SibSp:兄弟姐妹的个数\n",
    "* Parch:父母和孩子的个数\n",
    "* Ticket: 票号【与预测无关】\n",
    "* Fare: 票价\n",
    "* Cabin: 所在船舱\n",
    "* Embarked: 登船港口"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 164,
   "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>PassengerId</th>\n",
       "      <th>Survived</th>\n",
       "      <th>Pclass</th>\n",
       "      <th>Name</th>\n",
       "      <th>Sex</th>\n",
       "      <th>Age</th>\n",
       "      <th>SibSp</th>\n",
       "      <th>Parch</th>\n",
       "      <th>Ticket</th>\n",
       "      <th>Fare</th>\n",
       "      <th>Cabin</th>\n",
       "      <th>Embarked</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>493</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>Molson, Mr. Harry Markland</td>\n",
       "      <td>male</td>\n",
       "      <td>55.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>113787</td>\n",
       "      <td>30.5000</td>\n",
       "      <td>C30</td>\n",
       "      <td>S</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>53</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>Harper, Mrs. Henry Sleeper (Myna Haxtun)</td>\n",
       "      <td>female</td>\n",
       "      <td>49.0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>PC 17572</td>\n",
       "      <td>76.7292</td>\n",
       "      <td>D33</td>\n",
       "      <td>C</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>388</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>Buss, Miss. Kate</td>\n",
       "      <td>female</td>\n",
       "      <td>36.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>27849</td>\n",
       "      <td>13.0000</td>\n",
       "      <td>NaN</td>\n",
       "      <td>S</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>192</td>\n",
       "      <td>0</td>\n",
       "      <td>2</td>\n",
       "      <td>Carbines, Mr. William</td>\n",
       "      <td>male</td>\n",
       "      <td>19.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>28424</td>\n",
       "      <td>13.0000</td>\n",
       "      <td>NaN</td>\n",
       "      <td>S</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>687</td>\n",
       "      <td>0</td>\n",
       "      <td>3</td>\n",
       "      <td>Panula, Mr. Jaako Arnold</td>\n",
       "      <td>male</td>\n",
       "      <td>14.0</td>\n",
       "      <td>4</td>\n",
       "      <td>1</td>\n",
       "      <td>3101295</td>\n",
       "      <td>39.6875</td>\n",
       "      <td>NaN</td>\n",
       "      <td>S</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   PassengerId  Survived  Pclass                                      Name  \\\n",
       "0          493         0       1                Molson, Mr. Harry Markland   \n",
       "1           53         1       1  Harper, Mrs. Henry Sleeper (Myna Haxtun)   \n",
       "2          388         1       2                          Buss, Miss. Kate   \n",
       "3          192         0       2                     Carbines, Mr. William   \n",
       "4          687         0       3                  Panula, Mr. Jaako Arnold   \n",
       "\n",
       "      Sex   Age  SibSp  Parch    Ticket     Fare Cabin Embarked  \n",
       "0    male  55.0      0      0    113787  30.5000   C30        S  \n",
       "1  female  49.0      1      0  PC 17572  76.7292   D33        C  \n",
       "2  female  36.0      0      0     27849  13.0000   NaN        S  \n",
       "3    male  19.0      0      0     28424  13.0000   NaN        S  \n",
       "4    male  14.0      4      1   3101295  39.6875   NaN        S  "
      ]
     },
     "execution_count": 164,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "dftrain_raw = pd.read_csv('./data/titanic/train.csv')\n",
    "dftest_raw = pd.read_csv('./data/titanic/test.csv')\n",
    "dftrain_raw.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### label分布情况"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "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": [
    "ax = dftrain_raw['Survived'].value_counts().plot(kind='bar', fontsize=15)\n",
    "ax.set_ylabel('Counts')\n",
    "ax.set_xlabel('Survived')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Help on method value_counts in module pandas.core.base:\n",
      "\n",
      "value_counts(normalize=False, sort=True, ascending=False, bins=None, dropna=True) method of pandas.core.series.Series instance\n",
      "    Return a Series containing counts of unique values.\n",
      "    \n",
      "    The resulting object will be in descending order so that the\n",
      "    first element is the most frequently-occurring element.\n",
      "    Excludes NA values by default.\n",
      "    \n",
      "    Parameters\n",
      "    ----------\n",
      "    normalize : bool, default False\n",
      "        If True then the object returned will contain the relative\n",
      "        frequencies of the unique values.\n",
      "    sort : bool, default True\n",
      "        Sort by frequencies.\n",
      "    ascending : bool, default False\n",
      "        Sort in ascending order.\n",
      "    bins : int, optional\n",
      "        Rather than count values, group them into half-open bins,\n",
      "        a convenience for ``pd.cut``, only works with numeric data.\n",
      "    dropna : bool, default True\n",
      "        Don't include counts of NaN.\n",
      "    \n",
      "    Returns\n",
      "    -------\n",
      "    Series\n",
      "    \n",
      "    See Also\n",
      "    --------\n",
      "    Series.count: Number of non-NA elements in a Series.\n",
      "    DataFrame.count: Number of non-NA elements in a DataFrame.\n",
      "    \n",
      "    Examples\n",
      "    --------\n",
      "    >>> index = pd.Index([3, 1, 2, 3, 4, np.nan])\n",
      "    >>> index.value_counts()\n",
      "    3.0    2\n",
      "    4.0    1\n",
      "    2.0    1\n",
      "    1.0    1\n",
      "    dtype: int64\n",
      "    \n",
      "    With `normalize` set to `True`, returns the relative frequency by\n",
      "    dividing all values by the sum of values.\n",
      "    \n",
      "    >>> s = pd.Series([3, 1, 2, 3, 4, np.nan])\n",
      "    >>> s.value_counts(normalize=True)\n",
      "    3.0    0.4\n",
      "    4.0    0.2\n",
      "    2.0    0.2\n",
      "    1.0    0.2\n",
      "    dtype: float64\n",
      "    \n",
      "    **bins**\n",
      "    \n",
      "    Bins can be useful for going from a continuous variable to a\n",
      "    categorical variable; instead of counting unique\n",
      "    apparitions of values, divide the index in the specified\n",
      "    number of half-open bins.\n",
      "    \n",
      "    >>> s.value_counts(bins=3)\n",
      "    (2.0, 3.0]      2\n",
      "    (0.996, 2.0]    2\n",
      "    (3.0, 4.0]      1\n",
      "    dtype: int64\n",
      "    \n",
      "    **dropna**\n",
      "    \n",
      "    With `dropna` set to `False` we can also see NaN index values.\n",
      "    \n",
      "    >>> s.value_counts(dropna=False)\n",
      "    3.0    2\n",
      "    NaN    1\n",
      "    4.0    1\n",
      "    2.0    1\n",
      "    1.0    1\n",
      "    dtype: int64\n",
      "\n"
     ]
    }
   ],
   "source": [
    "help(dftrain_raw['Survived'].value_counts)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 年龄分布情况"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "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": [
    "ax = dftrain_raw['Age'].plot(kind='hist', bins=20)\n",
    "ax.set_ylabel('Frequency')\n",
    "ax.set_xlabel('Age')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 死亡和幸存的核密度折线图"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "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": [
    "ax = dftrain_raw.query('Survived == 0')['Age'].plot(kind='density')  # 死亡\n",
    "dftrain_raw.query('Survived == 1')['Age'].plot(kind='density') # 幸存\n",
    "ax.legend(['Die', 'Survive'])\n",
    "ax.set_ylabel('Density')\n",
    "ax.set_xlabel('Age')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "> 从上图中可以看出来的是, 死亡和幸存较多的人都集中在30岁左右"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Help on function get_dummies in module pandas.core.reshape.reshape:\n",
      "\n",
      "get_dummies(data, prefix=None, prefix_sep='_', dummy_na=False, columns=None, sparse=False, drop_first=False, dtype=None) -> 'DataFrame'\n",
      "    Convert categorical variable into dummy/indicator variables.\n",
      "    \n",
      "    Parameters\n",
      "    ----------\n",
      "    data : array-like, Series, or DataFrame\n",
      "        Data of which to get dummy indicators.\n",
      "    prefix : str, list of str, or dict of str, default None\n",
      "        String to append DataFrame column names.\n",
      "        Pass a list with length equal to the number of columns\n",
      "        when calling get_dummies on a DataFrame. Alternatively, `prefix`\n",
      "        can be a dictionary mapping column names to prefixes.\n",
      "    prefix_sep : str, default '_'\n",
      "        If appending prefix, separator/delimiter to use. Or pass a\n",
      "        list or dictionary as with `prefix`.\n",
      "    dummy_na : bool, default False\n",
      "        Add a column to indicate NaNs, if False NaNs are ignored.\n",
      "    columns : list-like, default None\n",
      "        Column names in the DataFrame to be encoded.\n",
      "        If `columns` is None then all the columns with\n",
      "        `object` or `category` dtype will be converted.\n",
      "    sparse : bool, default False\n",
      "        Whether the dummy-encoded columns should be backed by\n",
      "        a :class:`SparseArray` (True) or a regular NumPy array (False).\n",
      "    drop_first : bool, default False\n",
      "        Whether to get k-1 dummies out of k categorical levels by removing the\n",
      "        first level.\n",
      "    dtype : dtype, default np.uint8\n",
      "        Data type for new columns. Only a single dtype is allowed.\n",
      "    \n",
      "        .. versionadded:: 0.23.0\n",
      "    \n",
      "    Returns\n",
      "    -------\n",
      "    DataFrame\n",
      "        Dummy-coded data.\n",
      "    \n",
      "    See Also\n",
      "    --------\n",
      "    Series.str.get_dummies : Convert Series to dummy codes.\n",
      "    \n",
      "    Examples\n",
      "    --------\n",
      "    >>> s = pd.Series(list('abca'))\n",
      "    \n",
      "    >>> pd.get_dummies(s)\n",
      "       a  b  c\n",
      "    0  1  0  0\n",
      "    1  0  1  0\n",
      "    2  0  0  1\n",
      "    3  1  0  0\n",
      "    \n",
      "    >>> s1 = ['a', 'b', np.nan]\n",
      "    \n",
      "    >>> pd.get_dummies(s1)\n",
      "       a  b\n",
      "    0  1  0\n",
      "    1  0  1\n",
      "    2  0  0\n",
      "    \n",
      "    >>> pd.get_dummies(s1, dummy_na=True)\n",
      "       a  b  NaN\n",
      "    0  1  0    0\n",
      "    1  0  1    0\n",
      "    2  0  0    1\n",
      "    \n",
      "    >>> df = pd.DataFrame({'A': ['a', 'b', 'a'], 'B': ['b', 'a', 'c'],\n",
      "    ...                    'C': [1, 2, 3]})\n",
      "    \n",
      "    >>> pd.get_dummies(df, prefix=['col1', 'col2'])\n",
      "       C  col1_a  col1_b  col2_a  col2_b  col2_c\n",
      "    0  1       1       0       0       1       0\n",
      "    1  2       0       1       1       0       0\n",
      "    2  3       1       0       0       0       1\n",
      "    \n",
      "    >>> pd.get_dummies(pd.Series(list('abcaa')))\n",
      "       a  b  c\n",
      "    0  1  0  0\n",
      "    1  0  1  0\n",
      "    2  0  0  1\n",
      "    3  1  0  0\n",
      "    4  1  0  0\n",
      "    \n",
      "    >>> pd.get_dummies(pd.Series(list('abcaa')), drop_first=True)\n",
      "       b  c\n",
      "    0  0  0\n",
      "    1  1  0\n",
      "    2  0  1\n",
      "    3  0  0\n",
      "    4  0  0\n",
      "    \n",
      "    >>> pd.get_dummies(pd.Series(list('abc')), dtype=float)\n",
      "         a    b    c\n",
      "    0  1.0  0.0  0.0\n",
      "    1  0.0  1.0  0.0\n",
      "    2  0.0  0.0  1.0\n",
      "\n"
     ]
    }
   ],
   "source": [
    "help(pd.get_dummies)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {},
   "outputs": [],
   "source": [
    "def preprocessing(df):\n",
    "    dfresult = pd.DataFrame()\n",
    "    # Pclass, 变成one-hot编码\n",
    "    dfPlass = pd.get_dummies(df['Pclass'], prefix = 'Pclass')\n",
    "    dfresult = pd.concat([dfresult, dfPlass], axis = 1)\n",
    "    \n",
    "    # Sex\n",
    "    dfSex = pd.get_dummies(df['Sex'], prefix = 'Sex')\n",
    "    dfresult = pd.concat([dfresult, dfSex], axis = 1)\n",
    "    \n",
    "    # Age\n",
    "    dfresult['Age'] = df['Age'].fillna(0)  # 如果为nan的话就用0填充\n",
    "    dfresult['Age_null'] = pd.isna(df['Age']).astype('int32')  # 虽然用0填充了，但还需要一个标记，标识这里是否为null\n",
    "    \n",
    "    # SibSp, Parch, Fare\n",
    "    dfresult['SibSp'] = df['SibSp']\n",
    "    dfresult['Parch'] = df['Parch']\n",
    "    dfresult['Fare'] = df['Fare']\n",
    "    \n",
    "    # Carbin\n",
    "    dfresult['Cabin_null'] = pd.isna(df['Cabin']).astype('int32')  # 标记是否有船舱\n",
    "    \n",
    "    # Embarked\n",
    "    dfEmbarked = pd.get_dummies(df['Embarked'], dummy_na=True, prefix='Embarked')\n",
    "    \n",
    "    dfresult = pd.concat([dfresult, dfEmbarked], axis=1)\n",
    "    return dfresult"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "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>Pclass_1</th>\n",
       "      <th>Pclass_2</th>\n",
       "      <th>Pclass_3</th>\n",
       "      <th>Sex_female</th>\n",
       "      <th>Sex_male</th>\n",
       "      <th>Age</th>\n",
       "      <th>Age_null</th>\n",
       "      <th>SibSp</th>\n",
       "      <th>Parch</th>\n",
       "      <th>Fare</th>\n",
       "      <th>Cabin_null</th>\n",
       "      <th>Embarked_C</th>\n",
       "      <th>Embarked_Q</th>\n",
       "      <th>Embarked_S</th>\n",
       "      <th>Embarked_nan</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>55.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>30.5000</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>49.0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>76.7292</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>36.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>13.0000</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>19.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>13.0000</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>14.0</td>\n",
       "      <td>0</td>\n",
       "      <td>4</td>\n",
       "      <td>1</td>\n",
       "      <td>39.6875</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   Pclass_1  Pclass_2  Pclass_3  Sex_female  Sex_male   Age  Age_null  SibSp  \\\n",
       "0         1         0         0           0         1  55.0         0      0   \n",
       "1         1         0         0           1         0  49.0         0      1   \n",
       "2         0         1         0           1         0  36.0         0      0   \n",
       "3         0         1         0           0         1  19.0         0      0   \n",
       "4         0         0         1           0         1  14.0         0      4   \n",
       "\n",
       "   Parch     Fare  Cabin_null  Embarked_C  Embarked_Q  Embarked_S  \\\n",
       "0      0  30.5000           0           0           0           1   \n",
       "1      0  76.7292           0           1           0           0   \n",
       "2      0  13.0000           1           0           0           1   \n",
       "3      0  13.0000           1           0           0           1   \n",
       "4      1  39.6875           1           0           0           1   \n",
       "\n",
       "   Embarked_nan  \n",
       "0             0  \n",
       "1             0  \n",
       "2             0  \n",
       "3             0  \n",
       "4             0  "
      ]
     },
     "execution_count": 45,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "x_train = preprocessing(dftrain_raw)\n",
    "x_train.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "x_train shape: (712, 15)\n",
      "x_test shape: (179, 15)\n"
     ]
    }
   ],
   "source": [
    "y_train = dftrain_raw['Survived']\n",
    "x_test = preprocessing(dftest_raw)\n",
    "y_test = dftest_raw['Survived']\n",
    "print('x_train shape:', x_train.shape)\n",
    "print('x_test shape:', x_test.shape)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 定义模型"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Keras构建模型"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"model\"\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "input_1 (InputLayer)         [(None, 15)]              0         \n",
      "_________________________________________________________________\n",
      "dense (Dense)                (None, 20)                320       \n",
      "_________________________________________________________________\n",
      "dense_1 (Dense)              (None, 10)                210       \n",
      "_________________________________________________________________\n",
      "dense_2 (Dense)              (None, 1)                 11        \n",
      "=================================================================\n",
      "Total params: 541\n",
      "Trainable params: 541\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "tf.keras.backend.clear_session()\n",
    "x_input = layers.Input(shape=(15,))\n",
    "dense1 = layers.Dense(20, activation = 'relu')(x_input)\n",
    "dense2 = layers.Dense(10, activation = 'relu')(dense1)\n",
    "x_output = layers.Dense(1, activation = 'sigmoid')(dense2)\n",
    "model = tf.keras.Model(inputs = x_input, outputs = x_output)\n",
    "model.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/30\n",
      "18/18 [==============================] - 0s 25ms/step - loss: 0.7278 - auc: 0.6326 - val_loss: 0.7086 - val_auc: 0.6565\n",
      "Epoch 2/30\n",
      "18/18 [==============================] - 0s 4ms/step - loss: 0.6806 - auc: 0.7129 - val_loss: 0.6932 - val_auc: 0.6558\n",
      "Epoch 3/30\n",
      "18/18 [==============================] - 0s 4ms/step - loss: 0.6741 - auc: 0.7163 - val_loss: 0.6841 - val_auc: 0.6543\n",
      "Epoch 4/30\n",
      "18/18 [==============================] - 0s 4ms/step - loss: 0.6608 - auc: 0.7743 - val_loss: 0.6708 - val_auc: 0.6968\n",
      "Epoch 5/30\n",
      "18/18 [==============================] - 0s 4ms/step - loss: 0.6395 - auc: 0.7644 - val_loss: 0.6587 - val_auc: 0.6788\n",
      "Epoch 6/30\n",
      "18/18 [==============================] - 0s 4ms/step - loss: 0.6161 - auc: 0.7614 - val_loss: 0.6514 - val_auc: 0.6705\n",
      "Epoch 7/30\n",
      "18/18 [==============================] - 0s 4ms/step - loss: 0.5987 - auc: 0.7529 - val_loss: 0.6476 - val_auc: 0.6822\n",
      "Epoch 8/30\n",
      "18/18 [==============================] - 0s 4ms/step - loss: 0.5890 - auc: 0.7569 - val_loss: 0.6509 - val_auc: 0.6906\n",
      "Epoch 9/30\n",
      "18/18 [==============================] - 0s 4ms/step - loss: 0.5837 - auc: 0.7533 - val_loss: 0.6503 - val_auc: 0.6943\n",
      "Epoch 10/30\n",
      "18/18 [==============================] - 0s 4ms/step - loss: 0.5749 - auc: 0.7719 - val_loss: 0.6314 - val_auc: 0.6972\n",
      "Epoch 11/30\n",
      "18/18 [==============================] - 0s 4ms/step - loss: 0.5690 - auc: 0.7759 - val_loss: 0.6243 - val_auc: 0.7039\n",
      "Epoch 12/30\n",
      "18/18 [==============================] - 0s 4ms/step - loss: 0.5575 - auc: 0.7846 - val_loss: 0.6148 - val_auc: 0.7149\n",
      "Epoch 13/30\n",
      "18/18 [==============================] - 0s 4ms/step - loss: 0.5497 - auc: 0.7929 - val_loss: 0.6027 - val_auc: 0.7312\n",
      "Epoch 14/30\n",
      "18/18 [==============================] - 0s 4ms/step - loss: 0.5411 - auc: 0.8079 - val_loss: 0.6008 - val_auc: 0.7372\n",
      "Epoch 15/30\n",
      "18/18 [==============================] - 0s 4ms/step - loss: 0.5325 - auc: 0.8087 - val_loss: 0.5961 - val_auc: 0.7423\n",
      "Epoch 16/30\n",
      "18/18 [==============================] - 0s 4ms/step - loss: 0.5241 - auc: 0.8142 - val_loss: 0.5848 - val_auc: 0.7538\n",
      "Epoch 17/30\n",
      "18/18 [==============================] - 0s 4ms/step - loss: 0.5133 - auc: 0.8343 - val_loss: 0.5778 - val_auc: 0.7615\n",
      "Epoch 18/30\n",
      "18/18 [==============================] - 0s 4ms/step - loss: 0.5022 - auc: 0.8347 - val_loss: 0.5775 - val_auc: 0.7581\n",
      "Epoch 19/30\n",
      "18/18 [==============================] - 0s 4ms/step - loss: 0.4949 - auc: 0.8432 - val_loss: 0.5658 - val_auc: 0.7675\n",
      "Epoch 20/30\n",
      "18/18 [==============================] - 0s 4ms/step - loss: 0.4837 - auc: 0.8510 - val_loss: 0.5686 - val_auc: 0.7643\n",
      "Epoch 21/30\n",
      "18/18 [==============================] - 0s 4ms/step - loss: 0.4782 - auc: 0.8490 - val_loss: 0.5676 - val_auc: 0.7649\n",
      "Epoch 22/30\n",
      "18/18 [==============================] - 0s 4ms/step - loss: 0.4688 - auc: 0.8578 - val_loss: 0.5657 - val_auc: 0.7645\n",
      "Epoch 23/30\n",
      "18/18 [==============================] - 0s 4ms/step - loss: 0.4630 - auc: 0.8560 - val_loss: 0.5633 - val_auc: 0.7657\n",
      "Epoch 24/30\n",
      "18/18 [==============================] - 0s 4ms/step - loss: 0.4591 - auc: 0.8598 - val_loss: 0.5728 - val_auc: 0.7644\n",
      "Epoch 25/30\n",
      "18/18 [==============================] - 0s 4ms/step - loss: 0.4606 - auc: 0.8563 - val_loss: 0.5604 - val_auc: 0.7709\n",
      "Epoch 26/30\n",
      "18/18 [==============================] - 0s 4ms/step - loss: 0.4463 - auc: 0.8679 - val_loss: 0.5594 - val_auc: 0.7742\n",
      "Epoch 27/30\n",
      "18/18 [==============================] - 0s 4ms/step - loss: 0.4425 - auc: 0.8669 - val_loss: 0.5558 - val_auc: 0.7723\n",
      "Epoch 28/30\n",
      "18/18 [==============================] - 0s 4ms/step - loss: 0.4399 - auc: 0.8698 - val_loss: 0.5562 - val_auc: 0.7737\n",
      "Epoch 29/30\n",
      "18/18 [==============================] - 0s 4ms/step - loss: 0.4360 - auc: 0.8712 - val_loss: 0.5589 - val_auc: 0.7763\n",
      "Epoch 30/30\n",
      "18/18 [==============================] - 0s 4ms/step - loss: 0.4376 - auc: 0.8676 - val_loss: 0.5626 - val_auc: 0.7749\n"
     ]
    }
   ],
   "source": [
    "model.compile(optimizer='adam', loss='binary_crossentropy', metrics=['AUC'])\n",
    "history = model.fit(x_train, y_train, validation_split=0.2, epochs=30)  # 分割一部分用于数据验证"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "metadata": {},
   "outputs": [
    {
     "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": [
    "len(history.history['loss'])\n",
    "# 绘制loss\n",
    "plt.plot(history.history['loss'])\n",
    "plt.plot(history.history['val_loss'])\n",
    "plt.legend(['train_loss', 'val_loss'])\n",
    "plt.xlabel('Epochs')\n",
    "plt.ylabel('loss')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "6/6 [==============================] - 0s 2ms/step - loss: 0.4671 - auc: 0.8198\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "[0.46710455417633057, 0.819786548614502]"
      ]
     },
     "execution_count": 60,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 预测类别\n",
    "model.evaluate(x_test, y_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "6/6 [==============================] - 0s 2ms/step - loss: 0.4671 - auc: 0.8198\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "[0.46710455417633057, 0.819786548614502]"
      ]
     },
     "execution_count": 61,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 使用Keras方式保存\n",
    "model.save('./keras_1-1_model.h5')\n",
    "del model\n",
    "model = models.load_model('./keras_1-1_model.h5')\n",
    "model.evaluate(x_test, y_test)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Tensorflow构建动态计算图"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 136,
   "metadata": {},
   "outputs": [],
   "source": [
    "def loss_func(y_pred, y):\n",
    "    epsilon = 1e-7\n",
    "    loss = -tf.reduce_mean(y * tf.math.log(y_pred + epsilon) + (1 - y) * tf.math.log(1 - y_pred + epsilon))\n",
    "    return loss"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 160,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "tf.Tensor(7.3237233, shape=(), dtype=float32)\n"
     ]
    }
   ],
   "source": [
    "x_train_tensor = x_train.values\n",
    "y_train_tensor = y_train.values\n",
    "class MyModel:\n",
    "    def __init__(self):\n",
    "        self.w1 = tf.Variable(tf.random.truncated_normal([15, 20], dtype='float32'))\n",
    "        self.b1 = tf.Variable(tf.random.truncated_normal([1, 20], dtype='float32'))\n",
    "        self.w2 = tf.Variable(tf.random.truncated_normal([20, 10], dtype='float32'))\n",
    "        self.b2 = tf.Variable(tf.random.truncated_normal([1, 10], dtype='float32'))\n",
    "        self.w3 = tf.Variable(tf.random.truncated_normal([10, 1], dtype='float32'))\n",
    "        self.b3 = tf.Variable(tf.random.truncated_normal([1, 1], dtype='float32'))\n",
    "        \n",
    "    def __call__(self, x):\n",
    "        # InputLayer -> Dense1\n",
    "        densor1 = x @ self.w1 + self.b1\n",
    "        # Dense1 -> Dense2\n",
    "        densor2 = densor1 @ self.w2 + self.b2\n",
    "        # Dense2 -> sigmoid\n",
    "        pred = tf.sigmoid(densor2 @ self.w3 + self.b3)\n",
    "        return pred\n",
    "# 计算损失\n",
    "m_model = MyModel()\n",
    "loss = loss_func(m_model(x_train_tensor), y_train_tensor)\n",
    "print(loss)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 161,
   "metadata": {},
   "outputs": [],
   "source": [
    "def train(model, inputs, outputs, learning_rate):\n",
    "    with tf.GradientTape() as t:\n",
    "        current_loss = loss_func(model(inputs), outputs)\n",
    "        dw1, db1, dw2, db2, dw3, db3 = t.gradient(current_loss, [model.w1, model.b1, model.w2, model.b2, model.w3, model.b3])\n",
    "        model.w1.assign_sub(learning_rate * dw1)\n",
    "        model.b1.assign_sub(learning_rate * db1)\n",
    "\n",
    "        model.w2.assign_sub(learning_rate * dw2)\n",
    "        model.b2.assign_sub(learning_rate * db2)\n",
    "\n",
    "        model.w3.assign_sub(learning_rate * dw3)\n",
    "        model.b3.assign_sub(learning_rate * db3)\n",
    "        return current_loss"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 162,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "epoch: 0  loss: tf.Tensor(7.3237233, shape=(), dtype=float32)\n",
      "epoch: 1  loss: tf.Tensor(6.250588, shape=(), dtype=float32)\n",
      "epoch: 2  loss: tf.Tensor(5.8712854, shape=(), dtype=float32)\n",
      "epoch: 3  loss: tf.Tensor(6.339668, shape=(), dtype=float32)\n",
      "epoch: 4  loss: tf.Tensor(6.3395123, shape=(), dtype=float32)\n",
      "epoch: 5  loss: tf.Tensor(6.339357, shape=(), dtype=float32)\n",
      "epoch: 6  loss: tf.Tensor(6.339203, shape=(), dtype=float32)\n",
      "epoch: 7  loss: tf.Tensor(6.339048, shape=(), dtype=float32)\n",
      "epoch: 8  loss: tf.Tensor(6.3388925, shape=(), dtype=float32)\n",
      "epoch: 9  loss: tf.Tensor(6.3360558, shape=(), dtype=float32)\n",
      "epoch: 10  loss: tf.Tensor(6.3330803, shape=(), dtype=float32)\n",
      "epoch: 11  loss: tf.Tensor(6.3302727, shape=(), dtype=float32)\n",
      "epoch: 12  loss: tf.Tensor(6.326312, shape=(), dtype=float32)\n",
      "epoch: 13  loss: tf.Tensor(6.3211823, shape=(), dtype=float32)\n",
      "epoch: 14  loss: tf.Tensor(6.3153396, shape=(), dtype=float32)\n",
      "epoch: 15  loss: tf.Tensor(6.3093104, shape=(), dtype=float32)\n",
      "epoch: 16  loss: tf.Tensor(6.30321, shape=(), dtype=float32)\n",
      "epoch: 17  loss: tf.Tensor(6.297239, shape=(), dtype=float32)\n",
      "epoch: 18  loss: tf.Tensor(6.291599, shape=(), dtype=float32)\n",
      "epoch: 19  loss: tf.Tensor(6.2865033, shape=(), dtype=float32)\n",
      "epoch: 20  loss: tf.Tensor(6.2820473, shape=(), dtype=float32)\n",
      "epoch: 21  loss: tf.Tensor(6.278247, shape=(), dtype=float32)\n",
      "epoch: 22  loss: tf.Tensor(6.2752357, shape=(), dtype=float32)\n",
      "epoch: 23  loss: tf.Tensor(6.273239, shape=(), dtype=float32)\n",
      "epoch: 24  loss: tf.Tensor(6.2722106, shape=(), dtype=float32)\n",
      "epoch: 25  loss: tf.Tensor(6.2717714, shape=(), dtype=float32)\n",
      "epoch: 26  loss: tf.Tensor(6.271593, shape=(), dtype=float32)\n",
      "epoch: 27  loss: tf.Tensor(6.2715154, shape=(), dtype=float32)\n",
      "epoch: 28  loss: tf.Tensor(6.2714763, shape=(), dtype=float32)\n",
      "epoch: 29  loss: tf.Tensor(6.271453, shape=(), dtype=float32)\n"
     ]
    }
   ],
   "source": [
    "for epoch in range(30):\n",
    "    current_loss = train(m_model, x_train_tensor, y_train_tensor, learning_rate = 0.1)\n",
    "    print(\"epoch:\", epoch, \" loss:\", current_loss)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "tensorflow2",
   "language": "python",
   "name": "tensorflow2"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.10"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}