1
2
3
4
5
6
7

import pandas as pd
import numpy as np
from sklearn.tree import DecisionTreeClassifier  #分类器 只能分类数字
import matplotlib.pyplot as plt
from sklearn.model_selection import GridSearchCV
from sklearn.model_selection import train_test_split
from sklearn.model_selection import cross_val_score

1
2

data = pd.read_csv("train.csv")
data.head()

1

data.info()

1
2
3
4
5
6

#筛选特征
data.drop(["Name","Ticket","Cabin"],inplace=True,axis=1)
#处理缺失值
data["Age"] = data["Age"].fillna(data["Age"].mean())
#删除缺失值少的行
data = data.dropna()

1

data.info()

1

labels = data["Embarked"].unique().tolist()

1
2
3

#转换多分类为数值
data["Embarked"] = data["Embarked"].apply(lambda x: labels.index(x))
#data["Sex"] = data["Sex"].apply(lambda x: labels.index(x))

1
2
3

#转换二分类为01变量
data["Sex"] = (data["Sex"] == "male").astype("int")
(data["Sex"] == "male").astype("int")

1

data

1
2

x = data.iloc[:,data.columns != "Survived"]
x

1
2

y = data.iloc[:,data.columns == "Survived"]
y

1

Xtrain,Xtest,ytrain,ytest = train_test_split(x,y,test_size=0.3)

1
2
3
4

Xtrain.index = range(Xtrain.shape[0])
#重构index
#Xtrain.reset_index(drop=True,inplace=True)
#会多一列index列，不好用

1
2

for i in [Xtrain,Xtest,ytrain,ytest]:
    i.index = range(i.shape[0])

1
2
3
4

clf = DecisionTreeClassifier(random_state=25)
clf = clf.fit(Xtrain,ytrain)
score = clf.score(Xtest,ytest)
score

1
2
3
4

clf = DecisionTreeClassifier(random_state=25)
#别划分测试和训练集，自动划分
score = cross_val_score(clf,x,y,cv=10).mean()
score

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14

# 交叉验证
tr = []#训练集分数
te = []#交叉验证分数
for i in range(10):
    clf = DecisionTreeClassifier(random_state=25
                                ,max_depth=i+1
                                ,criterion="entropy"#熵,欠拟合才用
                                )
    clf = clf.fit(Xtrain,ytrain)
    score_tr = clf.score(Xtrain,ytrain)
    score_te = cross_val_score(clf,x,y,cv=10).mean()
    tr.append(score_tr)
    te.append(score_te)
print(max(te))

1
2
3
4
5

plt.plot(range(1,11),tr,color="red",label="train")
plt.plot(range(1,11),te,color="blue",label="test")
plt.xticks(range(1,11))
plt.legend()
plt.show()

1

[*range(0,50,5)]

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16

#min_impurity_decrease 
gini_thresholds = np.linspace(0,0.5,20)#线性插值取20个数在0~0.5中

#entropy_threholds = np.linspace(0,1,50)#线性插值取50个数在0~1中

#parameters 本质是一串参数和这些参数对应的，希望网格搜索来搜索的参数的取值范围
parameters = {"criterion":("gini","entropy")#基尼 信息熵
              ,"splitter":("best","random")
              ,"max_depth":[*range(1,10)]#1~更多
              ,"min_samples_leaf":[*range(1,50,5)]#1~更多
              ,"min_impurity_decrease":[*np.linspace(0,0.5,20)]#看选基尼还是信息熵，区间变化
}

clf = DecisionTreeClassifier(random_state=25)#决策树分类器 随机稳定25
GS = GridSearchCV(clf,parameters,cv=10)
GS = GS.fit(Xtrain,ytrain)

1

GS.best_params_ #从我们输入的参数和参数取值的列表中，返回最佳组合

1

GS.best_score_ #网格搜索后的模型的评判标准