# 导包
import pandas as pd
import matplotlib.pyplot as plt
import seaborn as sns
import numpy as np
from sklearn.model_selection import train_test_split
from sklearn.ensemble import RandomForestClassifier
from sklearn.tree import DecisionTreeClassifier
from sklearn.ensemble import GradientBoostingClassifier
from sklearn.metrics import confusion_matrix,accuracy_score
import warnings
warnings.filterwarnings('ignore')
plt.rcParams['font.sans-serif'] = ['SimHei'] #解决中文显示
plt.rcParams['axes.unicode_minus'] = False   #解决符号无法显示
# 加载数据
data = pd.read_csv('Life Expectancy Data.csv')
pd.set_option('display.max_columns',None)
data.head(10)

# 查看数据信息
data.info()

# 查看数值型数据的描述
data.describe()

# 查看非数值型数据的描述
data.describe(include=np.object)

# 查看数据大小
data.shape

# 查看缺失值
data.isnull().sum()

# 删除缺失值
data.dropna(inplace=True)
data.isnull().sum()

# 查看是否有重复值,返回结果为True则存在重复值,为False则说明不存在重复值
any(data.duplicated())

# 预期寿命值低于（<65）的国家是否应该增加其医疗保健支出以改善其平均寿命？
data1 = data[data['Life expectancy ']<65]
plt.figure(figsize=(12,8))
plt.scatter(x=data1['percentage expenditure'],y=data1['Life expectancy '])
plt.title('支出百分比与预期寿命值低于（<65）的国家关系')
plt.show()

#相关性
"""
使用heatmap绘制相关性热力图
vmax设定热力图色块的最大区分值
square设定图片是否为正方形
annot 设定是否显示每个色块的系数值
cbar 是否显示图例
cmap 颜色主题
"""
fig = plt.figure(figsize=(18,18))
sns.heatmap(data.corr(),vmax=1,annot=True,linewidths=0.5,cbar=False,cmap='YlGnBu',annot_kws={'fontsize':18})
plt.xticks(fontsize=20)
plt.yticks(fontsize=20)
plt.title('各个因素之间的相关系数',fontsize=20)
plt.show()

# 是否接受教育对人类寿命有何影响？
plt.figure(figsize=(12,8))
plt.subplot(2,2,1)
plt.scatter(x=data['Schooling'],y=data['Life expectancy '])
plt.title('教育与预期寿命值的关系')
plt.subplot(2,2,2)
plt.scatter(x=data['Schooling'],y=data['Adult Mortality'])
plt.title('教育与成年死亡数的关系')
plt.subplot(2,2,3)
plt.scatter(x=data['Schooling'],y=data[' thinness  1-19 years'])
plt.title('教育与虚弱1-19年的关系')
plt.subplot(2,2,4)
plt.scatter(x=data['Schooling'],y=data[' thinness 5-9 years'])
plt.title('教育与虚弱5-9年的关系')
plt.show()

# 婴儿和成人死亡率如何影响预期寿命？
plt.figure(figsize=(12,8))
plt.subplot(1,2,1)
plt.scatter(x=data['Adult Mortality'],y=data['Life expectancy '])
plt.title('成人死亡率与预期寿命的关系')
plt.subplot(1,2,2)
plt.scatter(x=data['infant deaths'],y=data['Life expectancy '])
plt.title('婴儿死亡率与预期寿命的关系')
plt.show()

# 预期寿命与饮酒是正相关还是负相关？
plt.figure(figsize=(12,8))
plt.scatter(x=data['Alcohol'],y=data['Life expectancy '])
plt.title('饮酒与预期寿命的关系')
plt.show()

# 人口稠密的国家的预期寿命是否有降低的趋势？
plt.figure(figsize=(12,8))
plt.subplot(1,2,1)
plt.scatter(x=data['Population'],y=data['Life expectancy '])
plt.title('人口数量与预期寿命的关系')
plt.subplot(1,2,2)
plt.plot(data['Population'],data['Life expectancy '])
plt.title('人口数量与预期寿命的折线图')
plt.show()

# 以Status进行分组求出个因素的平均值
group1 = data.groupby('Status').mean()
group1
# 分析不同Status的国家的预期寿命、婴儿死亡数、麻疹、5岁以下死亡数、小儿麻痹、B已型肝炎的差异
plt.figure(figsize=(20,15))
plt.subplot(3,2,1)
group1['Life expectancy '].plot(kind='bar')
plt.subplot(3,2,2)
group1['infant deaths'].plot(kind='bar')
plt.subplot(3,2,3)
group1['Measles '].plot(kind='bar')
plt.subplot(3,2,4)
group1['under-five deaths '].plot(kind='bar')
plt.subplot(3,2,5)
group1['Polio'].plot(kind='bar')
plt.subplot(3,2,6)
group1['Hepatitis B'].plot(kind='bar')

# 将预期寿命按阶段划分为等级1-6
data1 = data.copy()
data1['grade'] = pd.cut(data1['Life expectancy '],bins=[35,45,55,65,75,85,90],labels=['1','2','3','4','5','6'],right=False)
data1.head()

# 将Status这一列值用0和1表示
data1['Status'] = data1['Status'].apply(lambda x:0 if x == 'Developing' else 1)
data1.head()

# 准备数据
X = data1.drop(['Life expectancy ','grade','Country'],axis=1)
y = data1['grade']
# 划分数据集
x_train,x_test,y_train,y_test = train_test_split(X,y,test_size=0.2,random_state=42)

# 构建决策树模型
from sklearn import tree
from six import StringIO
import pydotplus
tree_clf = DecisionTreeClassifier(max_depth=3)
tree_clf.fit(x_train,y_train)
y_pred = tree_clf.predict(x_test)
print('模型准确率:',accuracy_score(y_pred=y_pred,y_true=y_test))
print('混淆矩阵:',confusion_matrix(y_test,y_pred))
dot_data = StringIO()
tree.export_graphviz(
    tree_clf,
    out_file=dot_data,
    feature_names=X.columns[:],
    class_names=['1','2','3','4','5','6'],
    rounded=True,
    filled=True
)
graph = pydotplus.graph_from_dot_data(dot_data.getvalue())
graph.write_png('tree1.png')

模型准确率: 0.7515151515151515
混淆矩阵: [[  0   0   1   0   0   0]
 [  0  10  13   1   0   0]
 [  0   4  48   3   0   0]
 [  0   0  17 141   0   0]
 [  0   0   0  39  49   0]
 [  0   0   0   0   4   0]]

# 构建GBDT模型
gbst = GradientBoostingClassifier()
gbst.fit(x_train,y_train)
y_pred = gbst.predict(x_test)
print('模型准确率',accuracy_score(y_pred,y_test))
print(confusion_matrix(y_test,y_pred))

模型准确率 0.8939393939393939
[[  0   1   0   0   0   0]
 [  0  17   7   0   0   0]
 [  0   1  51   3   0   0]
 [  0   0   3 153   2   0]
 [  0   0   0  17  71   0]
 [  0   0   0   0   1   3]]

# 构建随机森林模型
rfc = RandomForestClassifier()
rfc.fit(x_train,y_train)
y_pred = rfc.predict(x_test)
print('模型准确率:',accuracy_score(y_pred=y_pred,y_true=y_test))
print('混淆矩阵:',confusion_matrix(y_test,y_pred))
#打印特征重要性评分
feat_labels = x_train.columns[0:]
importances = rfc.feature_importances_
indices = np.argsort(importances)[::-1]
for f,j in zip(range(x_train.shape[1]-1),indices):
    print(f + 1, feat_labels[j], importances[j])

模型准确率: 0.8878787878787879
混淆矩阵: [[  0   1   0   0   0   0]
 [  0  18   6   0   0   0]
 [  0   2  51   2   0   0]
 [  0   0   3 155   0   0]
 [  0   0   0  20  68   0]
 [  0   0   0   0   3   1]]
1 Adult Mortality 0.2072291122189003
2 Income composition of resources 0.15119291250758393
3  HIV/AIDS 0.08028919893081636
4  thinness 5-9 years 0.07166453731213923
5 Schooling 0.06714860274427781
6  thinness  1-19 years 0.052012459499437905
7 percentage expenditure 0.049963594335953106
8 Alcohol 0.04191949699555478
9 GDP 0.04132651013754628
10  BMI  0.04056282954843009
11 Total expenditure 0.03431848502937779
12 under-five deaths  0.023643667064835658
13 infant deaths 0.022416087868431765
14 Population 0.020849863426314803
15 Year 0.019126681509410066
16 Hepatitis B 0.0189391761817397
17 Diphtheria  0.018302310953593165
18 Polio 0.01756546236252764
19 Measles  0.015518045355663208