1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
# -*- coding: utf-8 -*
import pandas as pd
import numpy as np
from sklearn.feature_extraction import DictVectorizer
from sklearn.tree import DecisionTreeClassifier
from sklearn.model_selection import cross_val_score
from sklearn.tree import export_graphviz
import matplotlib.pyplot as plt
# 数据加载
train_data = pd.read_csv('train.csv')
test_data = pd.read_csv('test.csv')
# 数据探索
# print(train_data.info())
# print(train_data.describe())
# print(train_data.describe(include=['O']))
# print(train_data.head())
# print(train_data.tail())
# 数据清洗
# 使用平均年龄来填充年龄中的Nan值
train_data['Age'].fillna(train_data['Age'].mean(), inplace=True)
test_data['Age'].fillna(test_data['Age'].mean(),inplace=True)
# 使用票价的均值填充票价中的Nan值
train_data['Fare'].fillna(train_data['Fare'].mean(), inplace=True)
test_data['Fare'].fillna(test_data['Fare'].mean(),inplace=True)
# 使用登录最多的港口来填充登录港口的nan值
# print(train_data['Embarked'].value_counts())
train_data['Embarked'].fillna('S', inplace=True)
test_data['Embarked'].fillna('S',inplace=True)
# 特征选择
features = ['Pclass', 'Sex', 'Age', 'SibSp', 'Parch', 'Fare', 'Embarked']
train_features = train_data[features]
train_labels = train_data['Survived']
test_features = test_data[features]
test_labels = test_data['Survived']
dvec = DictVectorizer(sparse=False)
train_features = dvec.fit_transform(train_features.to_dict(orient='record'))
test_features = dvec.fit_transform(test_features.to_dict(orient='record'))
# 构造ID3决策树
clf = DecisionTreeClassifier(criterion='entropy')
# 构造CART决策树
#clf = DecisionTreeClassifier(criterion='gini')
# 决策树训练
clf.fit(train_features, train_labels)
# # 模型预测 & 评估
# test_features=dvec.transform(test_features.to_dict(orient='record'))
# # 决策树预测
# pred_labels = clf.predict(test_features)
# 得到决策树准确率
acc_decision_tree = round(clf.score(test_features, test_labels), 6)
print(u'score准确率为 %.4lf' % acc_decision_tree)
# 使用K折交叉验证 统计决策树准确率
print(u'cross_val_score准确率为 %.4lf' % np.mean(cross_val_score(clf, test_features, test_labels, cv=10)))
#用学习曲线确定最优剪枝参数取值
test = []
for i in range(10):
clf = DecisionTreeClassifier(max_depth=i+1,criterion="gini")
clf.fit(train_features, train_labels)
score = np.mean(cross_val_score(clf, test_features, test_labels, cv=10))
test.append(score)
plt.plot(range(1,11),test,color="red",label="max_depth")
plt.legend()
plt.show()
#可视化决策树
export_graphviz(clf, out_file="titanic_tree.dot" , feature_names=dvec.get_feature_names())

1
2
3
4
5
6
7
8
9
10
11
#用学习曲线确定最优剪枝参数取值
test = []
for i in range(10):
clf = DecisionTreeClassifier(max_depth=i+1,criterion="gini")
clf.fit(train_features, train_labels)
score = np.mean(cross_val_score(clf, test_features, test_labels, cv=10))
test.append(score)
plt.plot(range(1,11),test,color="red",label="max_depth")
plt.legend()
plt.show()

1
2
3
features = ['Pclass', 'Sex', 'Age', 'SibSp', 'Parch', 'Fare', 'Embarked']
train_features = train_data[features]

1
2
3
4
5
6
7
8
9
10
11
12
13
14
>>> df
col1
col2
0
1
3
1
2
4
>>> [df.to_dict(orient='index')]
[{0: {'col1': 1, 'col2': 3}, 1: {'col1': 2, 'col2': 4}}]
>>> df.to_dict(orient='records')
[{'col1': 1, 'col2': 3}, {'col1': 2, 'col2': 4}]

1
2
3
4
5
from sklearn.feature_extraction import DictVectorizer
dvec=DictVectorizer(sparse=False) # 不创建稀疏矩阵
# fit_transform 这个函数，它可以将特征向量转化为特征值矩阵
train_features=dvec.fit_transform(train_features.to_dict(orient='record'))

1
2
3
4
from sklearn.model_selection import cross_val_score
# 使用 K 折交叉验证 统计决策树准确率
print(u'cross_val_score 准确率为 %.4lf' % np.mean(cross_val_score(clf, train_features, train_labels, cv=10)))