【决策树-泰坦尼克数据案例】

---

前言

基于泰坦尼克数据进行决策树分析，考过特征值、目标值提取，样本数据分割，特征工程，模型训练等过程。

---

一、泰坦尼克数据介绍

数据主要根据年龄、性别、舱位等级等特征值预测是否生还。

PassengerId	Survived	Pclass	…	Fare	Cabin	Embarked
1	0	3st	…	7.2500	NaN	S
2	1	1st	…	71.2833	C85	C
3	1	3st	…	7.9250	NaN	S
4	1	1st	…	53.1000	C123	S
5	0	3st	…	8.0500	NaN	S

二、使用步骤

1.引入库

代码如下（示例）：

import pandas as pd
import numpy as np
from sklearn.model_selection import train_test_split
from sklearn.feature_extraction import DictVectorizer
from sklearn.tree import DecisionTreeClassifier

2.读入数据

代码如下（示例）：

data = pd.read_csv('E://titanic//train.csv')
data['Pclass'] = data['Pclass'].astype(str) + 'st'

3.提起特征值和目标值

x = data[['Pclass','Age','Sex']]
y = data['Survived']

4.处理异常值

x['Age'].fillna(value=data['Age'].mean(),inplace=True)

5.数据集分割

x_train,x_test,y_train,y_test = train_test_split(x,y,random_state=22,test_size=0.2)
x_train = x_train.to_dict(orient='records')
x_test = x_test.to_dict(orient='records')

6.特征工程，将类型值转化为ont-hot矩阵

transfer = DictVectorizer(sparse=False)
x_train = transfer.fit_transform(x_train)
x_test = transfer.fit_transform(x_test)

7.特征训练与评分

estimator = DecisionTreeClassifier(max_depth=4)
estimator.fit(x_train,y_train)
y_pre = estimator.predict(x_test)
print(y_pre)
print(estimator.score(x_test,y_test))