notebook/ML_diabetes_prediction/diabetes_prediciton_logistic_regression.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 使用逻辑回归实现病人用药辅助\n",
    "\n",
    "## 实验内容\n",
    "\n",
    "在本案例中，我们会通过分析患者的若干项医学检验指标，构建一个逻辑回归模型，预测一位患者是否有糖尿病，从而辅助医生的用药。\n",
    "\n",
    "我们使用的数据集是患者的各项检验指数据，数据集共有9个字段，解释如下：\n",
    "* BloodPressure：血压值\n",
    "* SkinThickness：肱三头肌皮肤褶皱厚度\n",
    "* Insulin：2小时血清胰岛素含量\n",
    "* BMI：体重指数\n",
    "* DiabetesPedigreeFunction：糖尿病谱系功能\n",
    "* Age: 年龄\n",
    "* Outcome：结果，是否有糖尿病\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 实验步骤"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 进入ModelArts界面\n",
    "\n",
    "点击如下链接：https://www.huaweicloud.com/product/modelarts.html ，进入ModelArts主页。点击“立即使用”按钮，输入用户名和密码登录，进入ModelArts使用页面。\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 创建ModelArts Notebook\n",
    "\n",
    "下面，我们在ModelArts中创建一个notebook开发环境，ModelArts notebook提供网页版的Python开发环境，可以方便的编写、运行代码，并查看运行结果。\n",
    "\n",
    "第一步：在ModelArts服务主界面依次点击“开发环境”、“创建”\n",
    "\n",
    "![create_nb_create_button](./img/create_nb_create_button.png)\n",
    "\n",
    "第二步：填写notebook所需的参数：\n",
    "\n",
    "| 参数 | 说明 |\n",
    "| - - - - - | - - - - - |\n",
    "| 计费方式 | 按需计费  |\n",
    "| 名称 | Notebook实例名称，如 diabetes_prediciton_logistic_regression |\n",
    "| 工作环境 | Python3 |\n",
    "| 资源池 | 选择\"公共资源池\"即可 |\n",
    "| 类型 | CPU |\n",
    "| 规格 | 2核8GiB |\n",
    "| 存储配置 | 选择EVS，磁盘规格5GB |\n",
    "\n",
    "第三步：配置好notebook参数后，点击下一步，进入notebook信息预览。确认无误后，点击“立即创建”\n",
    "\n",
    "![create_nb_creation_summary](./img/create_nb_creation_summary.png)\n",
    "\n",
    "第四步：创建完成后，返回开发环境主界面，等待Notebook创建完毕后，打开Notebook，进行下一步操作。\n",
    "![modelarts_notebook_index](./img/modelarts_notebook_index.png)\n",
    "\n",
    "### 在ModelArts中创建开发环境\n",
    "\n",
    "接下来，我们创建一个实际的开发环境，用于后续的实验步骤。\n",
    "\n",
    "第一步：点击下图所示的“打开”按钮，进入刚刚创建的Notebook\n",
    "![inter_dev_env](img/enter_dev_env.png)\n",
    "\n",
    "第二步：创建一个Python3环境的的Notebook。点击右上角的\"New\"，然后创建 XGBoost-Sklearn 开发环境。\n",
    "\n",
    "第三步：点击左上方的文件名\"Untitled\"，并输入一个与本实验相关的名称，如\"diabetes_prediciton_logistic_regression\"\n",
    "![notebook_untitled_filename](./img/notebook_untitled_filename.png)\n",
    "![notebook_name_the_ipynb](./img/notebook_name_the_ipynb.png)\n",
    "\n",
    "\n",
    "### 在Notebook中编写并执行代码\n",
    "\n",
    "在Notebook中，我们输入一个简单的打印语句，然后点击上方的运行按钮，可以查看语句执行的结果：\n",
    "![run_helloworld](./img/run_helloworld.png)\n",
    "\n",
    "\n",
    "开发环境准备好啦，接下来可以愉快地写代码啦！\n",
    "\n",
    "\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 初始化ModelArts SDK"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "from modelarts.session import Session\n",
    "import os\n",
    "session = Session()\n",
    "\n",
    "if session.region_name == \"cn-north-1\":\n",
    "    bucket_path = \"ai-course-common-20/diabetes_prediciton/diabetes_prediciton.tar.gz\"\n",
    "elif session.region_name == \"cn-north-4\":\n",
    "    bucket_path = \"ai-course-common-20-bj4/diabetes_prediciton/diabetes_prediciton.tar.gz\"\n",
    "else:\n",
    "    print(\"请更换地区到北京一或北京四\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 准备源代码和数据\n",
    "\n",
    "这一步准备案例所需的源代码和数据，相关资源已经保存在OBS（华为云对象存储服务）中，我们通过ModelArts SDK将资源下载到本地，并解压到当前目录下。解压后，当前目录包含src和data两个目录，分别存有源代码和数据集。\n",
    "\n",
    "会显示下载成功的日志。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Successfully download file ai-course-common-20/diabetes_prediciton/diabetes_prediciton.tar.gz from OBS to local ./diabetes_prediciton.tar.gz\n"
     ]
    }
   ],
   "source": [
    "if not os.path.exists('./data'):\n",
    "    session.download_data(bucket_path= bucket_path, path=\"./diabetes_prediciton.tar.gz\")\n",
    "\n",
    "    #  使用tar命令解压资源包\n",
    "    !tar xf ./diabetes_prediciton.tar.gz\n",
    "    \n",
    "    # 清理压缩包\n",
    "    !rm -f ./diabetes_prediciton.tar.gz"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 导入工具库\n",
    "\n",
    "matplotlib和seaborn是Python绘图工具，pandas和numpy是矩阵运算工具。\n",
    "\n",
    "此段代码只是引入Python包，无回显（代码执行输出）。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np \n",
    "import pandas as pd \n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 数据读取\n",
    "使用pandas.read_excel(filepath)方法读取notebook中的数据文件。\n",
    "* filepath：如果是在ModelArts notebook中。\n",
    "\n",
    "\n",
    "此段代码无回显。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "df = pd.read_csv('data/diabetes.csv')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 展示样本数据\n",
    "\n",
    "执行这段代码可以看到数据集的5个样本数据"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "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>Pregnancies</th>\n",
       "      <th>Glucose</th>\n",
       "      <th>BloodPressure</th>\n",
       "      <th>SkinThickness</th>\n",
       "      <th>Insulin</th>\n",
       "      <th>BMI</th>\n",
       "      <th>DiabetesPedigreeFunction</th>\n",
       "      <th>Age</th>\n",
       "      <th>Outcome</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>6</td>\n",
       "      <td>148</td>\n",
       "      <td>72</td>\n",
       "      <td>35</td>\n",
       "      <td>0</td>\n",
       "      <td>33.6</td>\n",
       "      <td>0.627</td>\n",
       "      <td>50</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>85</td>\n",
       "      <td>66</td>\n",
       "      <td>29</td>\n",
       "      <td>0</td>\n",
       "      <td>26.6</td>\n",
       "      <td>0.351</td>\n",
       "      <td>31</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>8</td>\n",
       "      <td>183</td>\n",
       "      <td>64</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>23.3</td>\n",
       "      <td>0.672</td>\n",
       "      <td>32</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>1</td>\n",
       "      <td>89</td>\n",
       "      <td>66</td>\n",
       "      <td>23</td>\n",
       "      <td>94</td>\n",
       "      <td>28.1</td>\n",
       "      <td>0.167</td>\n",
       "      <td>21</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0</td>\n",
       "      <td>137</td>\n",
       "      <td>40</td>\n",
       "      <td>35</td>\n",
       "      <td>168</td>\n",
       "      <td>43.1</td>\n",
       "      <td>2.288</td>\n",
       "      <td>33</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   Pregnancies  Glucose  BloodPressure  SkinThickness  Insulin   BMI  \\\n",
       "0            6      148             72             35        0  33.6   \n",
       "1            1       85             66             29        0  26.6   \n",
       "2            8      183             64              0        0  23.3   \n",
       "3            1       89             66             23       94  28.1   \n",
       "4            0      137             40             35      168  43.1   \n",
       "\n",
       "   DiabetesPedigreeFunction  Age  Outcome  \n",
       "0                     0.627   50        1  \n",
       "1                     0.351   31        0  \n",
       "2                     0.672   32        1  \n",
       "3                     0.167   21        0  \n",
       "4                     2.288   33        1  "
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.head(5)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 展示各个字段的统计值信息\n",
    "\n",
    "调用pandas.DataFrame.describe方法，可以看到各个特征的统计信息，包括样本数、均值、标准差、最小值、1/4分位数、1/2分位数、3/4分位数和最大值。transpose是矩阵的转置运算。"
   ]
  },
  {
   "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>count</th>\n",
       "      <th>mean</th>\n",
       "      <th>std</th>\n",
       "      <th>min</th>\n",
       "      <th>25%</th>\n",
       "      <th>50%</th>\n",
       "      <th>75%</th>\n",
       "      <th>max</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>Pregnancies</th>\n",
       "      <td>768.0</td>\n",
       "      <td>3.845052</td>\n",
       "      <td>3.369578</td>\n",
       "      <td>0.000</td>\n",
       "      <td>1.00000</td>\n",
       "      <td>3.0000</td>\n",
       "      <td>6.00000</td>\n",
       "      <td>17.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Glucose</th>\n",
       "      <td>768.0</td>\n",
       "      <td>120.894531</td>\n",
       "      <td>31.972618</td>\n",
       "      <td>0.000</td>\n",
       "      <td>99.00000</td>\n",
       "      <td>117.0000</td>\n",
       "      <td>140.25000</td>\n",
       "      <td>199.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>BloodPressure</th>\n",
       "      <td>768.0</td>\n",
       "      <td>69.105469</td>\n",
       "      <td>19.355807</td>\n",
       "      <td>0.000</td>\n",
       "      <td>62.00000</td>\n",
       "      <td>72.0000</td>\n",
       "      <td>80.00000</td>\n",
       "      <td>122.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>SkinThickness</th>\n",
       "      <td>768.0</td>\n",
       "      <td>20.536458</td>\n",
       "      <td>15.952218</td>\n",
       "      <td>0.000</td>\n",
       "      <td>0.00000</td>\n",
       "      <td>23.0000</td>\n",
       "      <td>32.00000</td>\n",
       "      <td>99.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Insulin</th>\n",
       "      <td>768.0</td>\n",
       "      <td>79.799479</td>\n",
       "      <td>115.244002</td>\n",
       "      <td>0.000</td>\n",
       "      <td>0.00000</td>\n",
       "      <td>30.5000</td>\n",
       "      <td>127.25000</td>\n",
       "      <td>846.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>BMI</th>\n",
       "      <td>768.0</td>\n",
       "      <td>31.992578</td>\n",
       "      <td>7.884160</td>\n",
       "      <td>0.000</td>\n",
       "      <td>27.30000</td>\n",
       "      <td>32.0000</td>\n",
       "      <td>36.60000</td>\n",
       "      <td>67.10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>DiabetesPedigreeFunction</th>\n",
       "      <td>768.0</td>\n",
       "      <td>0.471876</td>\n",
       "      <td>0.331329</td>\n",
       "      <td>0.078</td>\n",
       "      <td>0.24375</td>\n",
       "      <td>0.3725</td>\n",
       "      <td>0.62625</td>\n",
       "      <td>2.42</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Age</th>\n",
       "      <td>768.0</td>\n",
       "      <td>33.240885</td>\n",
       "      <td>11.760232</td>\n",
       "      <td>21.000</td>\n",
       "      <td>24.00000</td>\n",
       "      <td>29.0000</td>\n",
       "      <td>41.00000</td>\n",
       "      <td>81.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Outcome</th>\n",
       "      <td>768.0</td>\n",
       "      <td>0.348958</td>\n",
       "      <td>0.476951</td>\n",
       "      <td>0.000</td>\n",
       "      <td>0.00000</td>\n",
       "      <td>0.0000</td>\n",
       "      <td>1.00000</td>\n",
       "      <td>1.00</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                          count        mean         std     min       25%  \\\n",
       "Pregnancies               768.0    3.845052    3.369578   0.000   1.00000   \n",
       "Glucose                   768.0  120.894531   31.972618   0.000  99.00000   \n",
       "BloodPressure             768.0   69.105469   19.355807   0.000  62.00000   \n",
       "SkinThickness             768.0   20.536458   15.952218   0.000   0.00000   \n",
       "Insulin                   768.0   79.799479  115.244002   0.000   0.00000   \n",
       "BMI                       768.0   31.992578    7.884160   0.000  27.30000   \n",
       "DiabetesPedigreeFunction  768.0    0.471876    0.331329   0.078   0.24375   \n",
       "Age                       768.0   33.240885   11.760232  21.000  24.00000   \n",
       "Outcome                   768.0    0.348958    0.476951   0.000   0.00000   \n",
       "\n",
       "                               50%        75%     max  \n",
       "Pregnancies                 3.0000    6.00000   17.00  \n",
       "Glucose                   117.0000  140.25000  199.00  \n",
       "BloodPressure              72.0000   80.00000  122.00  \n",
       "SkinThickness              23.0000   32.00000   99.00  \n",
       "Insulin                    30.5000  127.25000  846.00  \n",
       "BMI                        32.0000   36.60000   67.10  \n",
       "DiabetesPedigreeFunction    0.3725    0.62625    2.42  \n",
       "Age                        29.0000   41.00000   81.00  \n",
       "Outcome                     0.0000    1.00000    1.00  "
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.describe().transpose()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 展示各个字段的数据类型\n",
    "\n",
    "pandas.DataFrame.info()方法可以展示各个字段的类型信息。\n",
    "\n",
    "可以看到每个字段的类型信息。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "RangeIndex: 768 entries, 0 to 767\n",
      "Data columns (total 9 columns):\n",
      "Pregnancies                 768 non-null int64\n",
      "Glucose                     768 non-null int64\n",
      "BloodPressure               768 non-null int64\n",
      "SkinThickness               768 non-null int64\n",
      "Insulin                     768 non-null int64\n",
      "BMI                         768 non-null float64\n",
      "DiabetesPedigreeFunction    768 non-null float64\n",
      "Age                         768 non-null int64\n",
      "Outcome                     768 non-null int64\n",
      "dtypes: float64(2), int64(7)\n",
      "memory usage: 54.1 KB\n"
     ]
    }
   ],
   "source": [
    "df.info()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 统计正、负样本数量\n",
    "\n",
    "pandas.DataFrame.groupby()方法可以分组统计每个值的数量。\n",
    "\n",
    "可以看到0的数量是500,1的数量是268。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Outcome\n",
       "0    500\n",
       "1    268\n",
       "dtype: int64"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.groupby('Outcome').size()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 展示正、负样本数量的分布\n",
    "\n",
    "这段代码使用matplotlib绘制条状图和饼图，展示正、负样本数量的分布。\n",
    "\n",
    "可以看到一张条状图和一张饼图。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fcbc1aeb208>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "f, ax = plt.subplots(1, 2, figsize = (15, 7))  # figsize用来设置画布大小\n",
    "f.suptitle(\"Diabetes?\", fontsize = 18.)        # 设置图像的标题\n",
    "_ = df.Outcome.value_counts().plot.bar(ax = ax[0], rot = 0, color = (sns.color_palette()[2], sns.color_palette()[3])).set(xticklabels = [\"No\", \"Yes\"])\n",
    "_ = df.Outcome.value_counts().plot.pie(labels = (\"No\", \"Yes\"), autopct = \"%.2f%%\", label = \"\", fontsize = 13., ax = ax[1],\\\n",
    "colors = (sns.color_palette()[2], sns.color_palette()[3]), wedgeprops = {\"linewidth\": 1.5, \"edgecolor\": \"white\"}), ax[1].texts[1].set_color(\"white\"), ax[1].texts[3].set_color(\"white\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 展示每个特征的直方图分布\n",
    "\n",
    "直方图是数值数据分布的精确图形表示，可以用来估计连续变量的概率分布。\n",
    "\n",
    "可以看到每个特征的直方图分布图。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[<matplotlib.axes._subplots.AxesSubplot object at 0x7fcbc1aeb0f0>,\n",
       "        <matplotlib.axes._subplots.AxesSubplot object at 0x7fcbb89b4048>,\n",
       "        <matplotlib.axes._subplots.AxesSubplot object at 0x7fcbb896c588>],\n",
       "       [<matplotlib.axes._subplots.AxesSubplot object at 0x7fcbb8925588>,\n",
       "        <matplotlib.axes._subplots.AxesSubplot object at 0x7fcbb8960518>,\n",
       "        <matplotlib.axes._subplots.AxesSubplot object at 0x7fcbb8960550>],\n",
       "       [<matplotlib.axes._subplots.AxesSubplot object at 0x7fcbb88c2b38>,\n",
       "        <matplotlib.axes._subplots.AxesSubplot object at 0x7fcbb88860b8>,\n",
       "        <matplotlib.axes._subplots.AxesSubplot object at 0x7fcbb883f5f8>]],\n",
       "      dtype=object)"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fcbc1ac7a90>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "df.hist(figsize=(16,15))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 观察特征之间的相关性\n",
    "\n",
    "此段代码通过seaborn绘制热力图，展示任意两个特征之间的相关性。数值的绝对值越高说明相关性越高。\n",
    "\n",
    "观察特征相关性的目的是，如果发现两个特征相关性很高，那么在特征选择时，可以丢弃其中的一个特征。\n",
    "\n",
    "此段代码执行后，会有一张特征相关性热力图输出。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0.5,1,'Correlation between features')"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fcbb87f7c18>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "corr=df.corr() # 生产相关性矩阵\n",
    "sns.set(font_scale=1.15) # 设置字体大小\n",
    "plt.figure(figsize=(14, 10)) # 设置图像大小\n",
    "\n",
    "# 绘制热力图\n",
    "sns.heatmap(corr, vmax=.8, linewidths=0.01, square=True,annot=True,cmap=\"Greens\",linecolor=\"black\")\n",
    "plt.title('Correlation between features')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 特征选择\n",
    "\n",
    "设置Outcome为预测目标。在数据分析小节中，发现各个特征并没有特别强的相关性，而且各项特征和预测目标Outcome之间的相关性也都不低，所以保留所有的特征。\n",
    "\n",
    "此段代码无回显。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "feature = df.drop(\"Outcome\",axis=1) # 特征\n",
    "target = df[\"Outcome\"] # 预测目标"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 训练模型"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 导入训练和评估工具\n",
    "\n",
    "在糖尿病预测实验中，我们使用开源机器学习框架sklearn提供的准确率计算函数accuracy_score和数据集切分函数train_test_split。这两个方法常用于机器学习分类器的构建过程中。\n",
    "\n",
    "此段代码无回显。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.metrics import accuracy_score # 用于计算模型准确率\n",
    "from sklearn.model_selection import train_test_split # 用于数据集切分"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 切分数据集\n",
    "\n",
    "使用train_test_split函数切分数据集，训练和测试的数据集比例是4：1。返回的X_train和y_train是训练数据的特征值和标签，X_test和y_test是测试数据的特征值和标签。\n",
    "\n",
    "可以看到训练集的矩阵形状是(576, 8)。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(576, 8)\n"
     ]
    }
   ],
   "source": [
    "X_train, X_test, y_train, y_test = train_test_split(feature, target, test_size=0.25, random_state=0)\n",
    "\n",
    "print(X_train.shape)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 使用逻辑回归算法训练模型\n",
    "\n",
    "使用自定义的逻辑回归算法训练模型，算法源码见src目录下的logistic_regression.py文件。大家可以阅读逻辑回归分类器的源码，学习如何使用Python语言实现逻辑回归分类器。\n",
    "\n",
    "可以看到分类器训练过程的日志打印，有训练步数，以及每步的损失值。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Iteration 0 \tcost= 0.686827383026021\n",
      "Iteration 1 \tcost= 0.6807756180442618\n",
      "Iteration 2 \tcost= 0.6749791363111561\n",
      "Iteration 3 \tcost= 0.6694256663270032\n",
      "Iteration 4 \tcost= 0.6641034265154976\n",
      "Iteration 5 \tcost= 0.6590011314279043\n",
      "Iteration 6 \tcost= 0.6541079927960338\n",
      "Iteration 7 \tcost= 0.6494137161819967\n",
      "Iteration 8 \tcost= 0.6449084939368468\n",
      "Iteration 9 \tcost= 0.6405829951305448\n",
      "Iteration 10 \tcost= 0.6364283530571244\n",
      "Iteration 11 \tcost= 0.6324361508557742\n",
      "Iteration 12 \tcost= 0.6285984057242167\n",
      "Iteration 13 \tcost= 0.6249075521376098\n",
      "Iteration 14 \tcost= 0.6213564244263141\n",
      "Iteration 15 \tcost= 0.6179382390101935\n",
      "Iteration 16 \tcost= 0.614646576536573\n",
      "Iteration 17 \tcost= 0.6114753641237805\n",
      "Iteration 18 \tcost= 0.6084188578723659\n",
      "Iteration 19 \tcost= 0.6054716257715006\n",
      "Iteration 20 \tcost= 0.602628531098368\n",
      "Iteration 21 \tcost= 0.5998847163831359\n",
      "Iteration 22 \tcost= 0.5972355879909381\n",
      "Iteration 23 \tcost= 0.5946768013547421\n",
      "Iteration 24 \tcost= 0.5922042468785533\n",
      "Iteration 25 \tcost= 0.58981403651873\n",
      "Iteration 26 \tcost= 0.5875024910418151\n",
      "Iteration 27 \tcost= 0.5852661279499721\n",
      "Iteration 28 \tcost= 0.5831016500593652\n",
      "Iteration 29 \tcost= 0.5810059347125844\n",
      "Iteration 30 \tcost= 0.5789760236030477\n",
      "Iteration 31 \tcost= 0.5770091131871474\n",
      "Iteration 32 \tcost= 0.5751025456584565\n",
      "Iteration 33 \tcost= 0.5732538004575649\n",
      "Iteration 34 \tcost= 0.5714604862907808\n",
      "Iteration 35 \tcost= 0.5697203336310365\n",
      "Iteration 36 \tcost= 0.5680311876747367\n",
      "Iteration 37 \tcost= 0.5663910017289019\n",
      "Iteration 38 \tcost= 0.5647978310037712\n",
      "Iteration 39 \tcost= 0.5632498267869185\n",
      "Iteration 40 \tcost= 0.5617452309759671\n",
      "Iteration 41 \tcost= 0.5602823709480255\n",
      "Iteration 42 \tcost= 0.5588596547450168\n",
      "Iteration 43 \tcost= 0.5574755665552353\n",
      "Iteration 44 \tcost= 0.5561286624724363\n",
      "Iteration 45 \tcost= 0.5548175665149263\n",
      "Iteration 46 \tcost= 0.5535409668880563\n",
      "Iteration 47 \tcost= 0.5522976124745863\n",
      "Iteration 48 \tcost= 0.5510863095382916\n",
      "Iteration 49 \tcost= 0.5499059186271242\n",
      "Iteration 50 \tcost= 0.548755351663075\n",
      "Iteration 51 \tcost= 0.547633569206752\n",
      "Iteration 52 \tcost= 0.5465395778854135\n",
      "Iteration 53 \tcost= 0.5454724279739684\n",
      "Iteration 54 \tcost= 0.5444312111191122\n",
      "Iteration 55 \tcost= 0.5434150581974477\n",
      "Iteration 56 \tcost= 0.5424231372990223\n",
      "Iteration 57 \tcost= 0.5414546518282806\n",
      "Iteration 58 \tcost= 0.5405088387149948\n",
      "Iteration 59 \tcost= 0.5395849667281837\n",
      "Iteration 60 \tcost= 0.5386823348865445\n",
      "Iteration 61 \tcost= 0.5378002709593193\n",
      "Iteration 62 \tcost= 0.5369381300519405\n",
      "Iteration 63 \tcost= 0.5360952932711749\n",
      "Iteration 64 \tcost= 0.5352711664648307\n",
      "Iteration 65 \tcost= 0.5344651790314205\n",
      "Iteration 66 \tcost= 0.5336767827954959\n",
      "Iteration 67 \tcost= 0.5329054509446133\n",
      "Iteration 68 \tcost= 0.5321506770242127\n",
      "Iteration 69 \tcost= 0.5314119739868715\n",
      "Iteration 70 \tcost= 0.530688873292688\n",
      "Iteration 71 \tcost= 0.5299809240577118\n",
      "Iteration 72 \tcost= 0.529287692247571\n",
      "Iteration 73 \tcost= 0.5286087599136218\n",
      "Iteration 74 \tcost= 0.5279437244690975\n",
      "Iteration 75 \tcost= 0.5272921980029402\n",
      "Iteration 76 \tcost= 0.5266538066290961\n",
      "Iteration 77 \tcost= 0.5260281898692423\n",
      "Iteration 78 \tcost= 0.5254150000669955\n",
      "Iteration 79 \tcost= 0.5248139018318323\n",
      "Iteration 80 \tcost= 0.5242245715109974\n",
      "Iteration 81 \tcost= 0.5236466966878414\n",
      "Iteration 82 \tcost= 0.5230799757050835\n",
      "Iteration 83 \tcost= 0.522524117211625\n",
      "Iteration 84 \tcost= 0.5219788397315761\n",
      "Iteration 85 \tcost= 0.5214438712543087\n",
      "Iteration 86 \tcost= 0.5209189488443302\n",
      "Iteration 87 \tcost= 0.520403818269946\n",
      "Iteration 88 \tcost= 0.5198982336496543\n",
      "Iteration 89 \tcost= 0.5194019571153321\n",
      "Iteration 90 \tcost= 0.5189147584913115\n",
      "Iteration 91 \tcost= 0.5184364149884964\n",
      "Iteration 92 \tcost= 0.5179667109127174\n",
      "Iteration 93 \tcost= 0.5175054373865868\n",
      "Iteration 94 \tcost= 0.5170523920841323\n",
      "Iteration 95 \tcost= 0.5166073789775569\n",
      "Iteration 96 \tcost= 0.51617020809549\n",
      "Iteration 97 \tcost= 0.5157406952921337\n",
      "Iteration 98 \tcost= 0.5153186620267619\n",
      "Iteration 99 \tcost= 0.51490393515303\n"
     ]
    }
   ],
   "source": [
    "from src.logistic_regression import LogisticRegression # 导入自定义的逻辑回归算法\n",
    "\n",
    "# 设置参数\n",
    "learning_rate = 0.05  # 学习率，在梯度下降中，控制权重更新的幅度\n",
    "lam = 0.01            # 参数惩罚项，为了防止过拟合\n",
    "\n",
    "# 创建模型\n",
    "LR_customized = LogisticRegression(X_train, y_train, learning_rate, lam)\n",
    "\n",
    "# 训练模型，运行100步，并打印每个训练步的损失值。\n",
    "step = 100\n",
    "LR_customized.run(step, printIter=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 计算准确率\n",
    "\n",
    "会输出逻辑回归模型的准确率。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "accuracy based on logistic regression: 0.7604166666666666\n"
     ]
    }
   ],
   "source": [
    "y_pred_cutomized = LR_customized.predict(X_test)\n",
    "print(\"accuracy based on logistic regression:\", accuracy_score(y_test, y_pred_cutomized))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "可以看到，准确率并不高。接下来教大家如何调节算法训练过程中的参数，来提升模型的准确率。"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 调节步数（step）\n",
    "\n",
    "这段代码展示如何调节步数提升模型准确率，并在实验中让大家更好地理解分类器训练中步数这个概念。\n",
    "\n",
    "这段代码逐个训练每个步数，并记录每个步数训练好之后的模型准确率。最后使用折线图显示模型准确率随着步数变化的趋势。\n",
    "\n",
    "可以看到一张横坐标为步数，纵坐标为准确率的折线图。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fcbb146f6d8>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "steps = [50, 100, 150, 200, 250, 300, 350] # 设置步数集合\n",
    "acc = [] # 准确率集合\n",
    "\n",
    "for step in steps:\n",
    "    lam = 0.01  # 参数惩罚项\n",
    "    learning_rate = 0.1 # 学习率\n",
    "    LR_customized = LogisticRegression(X_train, y_train, learning_rate, lam)\n",
    "    LR_customized.run(step, printIter=False)\n",
    "    y_pred_cutomized = LR_customized.predict(X_test)\n",
    "    accuracy = accuracy_score(y_test, y_pred_cutomized)\n",
    "    acc.append(accuracy)\n",
    "\n",
    "# 绘制折线图\n",
    "plt.ylabel(\"accuracy\")\n",
    "plt.xlabel(\"step\")\n",
    "plt.plot(steps, acc, color='red', marker='o', linestyle='solid')\n",
    "plt.show()\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "观察上图可以发现，训练步数从50增加到200步时，准确率是一直增加的；但是当步数从200步开始继续增加的时候，准确率就不再增加了。\n",
    "\n",
    "说明在权重惩罚项为 0.01，学习率为0.1的情况下，训练200步就够了。训练步数表示训练的充分程度，即梯度下降的迭代次数。步数少了模型精度不能达到最优，步数太多精度也不会随着增加，是浪费时间和资源。所以选择恰当的步数，停止训练即可。"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 调节学习率（learning rate）\n",
    "\n",
    "这段代码展示如何调节学习率提升模型准确率，并在实验中让大家更好地理解分类器训练中学习率这个概念。\n",
    "\n",
    "这段代码逐个使用候选学习率训练300步，并记录训练完成后的模型准确率。最后使用折线图显示模型准确率随着学习率变化的趋势。\n",
    "\n",
    "可以看到一张横坐标为学习率，纵坐标为准确率的折线图。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fcbb1452a58>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "learning_rates = [1, 0.5, 0.1, 0.05, 0.01, 0.005, 0.001] # 设置学习率集合\n",
    "acc = [] # 准确率集合\n",
    "\n",
    "for learning_rate in learning_rates:\n",
    "    lam = 0.01   # 参数惩罚项\n",
    "    step = 300   # 步数\n",
    "    LR_customized = LogisticRegression(X_train, y_train, learning_rate, lam)\n",
    "    LR_customized.run(step, printIter=False)\n",
    "    y_pred_cutomized = LR_customized.predict(X_test)\n",
    "    accuracy = accuracy_score(y_test, y_pred_cutomized)\n",
    "    acc.append(accuracy)\n",
    "\n",
    "# 绘制折线图\n",
    "plt.ylabel(\"accuracy\")\n",
    "plt.xlabel(\"learning rate\")\n",
    "plt.plot(learning_rates, acc, color='red', marker='o', linestyle='solid')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "观察上图可以发现，学习率从0.001增加到0.1的过程中，准确率是一直增加的；但是当步数从0.1继续增加的时候，准确率就不再增加了。\n",
    "\n",
    "说明在权重惩罚项为0.01，训练300步的情况下，学习率选择0.1就够了。学习率又叫步长，步长和步数一起控制梯度下降中权重的更新，步长控制权重每步更新的幅度，步数控制权重的更新次数。步长不能太小，否则就需要训练更多的步数，才能达到最优准确率；步长也不能太大，否则可能会错过梯度下降中的最优解。\n",
    "\n",
    "一般学习率的经验值取值范围是0.0001到0.1之间。最好的学习率设置策略是，刚开始训练时，选择稍大的值；一定步数过后，逐渐减小；接近训练结束的时候，学习率最好缩减至刚开始的100倍以上。\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 保存模型文件\n",
    "\n",
    "我们将逻辑回归算法训练得到的模型文件保存下来，方便以后再使用。"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 导出模型文件\n",
    "\n",
    "训练得到模型可以导出至本地。会打印出保存的本地路径。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['output/train_model_lr.m']"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.externals import joblib\n",
    "import os\n",
    "\n",
    "if not os.path.exists('output'):\n",
    "    os.mkdir('output') # 创建本地保存路径\n",
    "\n",
    "joblib.dump(LR_customized, \"output/train_model_lr.m\") # 保存"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 小结\n",
    "\n",
    "通过调节步长和步数这两个参数，逻辑回归分类器在糖尿病预测这个应用中，准确率能达到0.82。在糖尿病预测应用中，逻辑回归分类器不一定是最合适的，大家可以尝试使用其他的分类器，并应用调参技术，使分类器达到更高的准确率。"
   ]
  },
  {
   "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": "3.7.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}