Python数据科学_9_决策树算法基础

泰坦尼克号生还者预测

读取数据集

import pandas as pd

data = pd.read_csv('titanic_data.csv', index_col='PassengerId')

数据探索

data.head()

	Survived	Pclass	Sex	Age
PassengerId
1	0	3	male	22.0
2	1	1	female	38.0
3	1	3	female	26.0
4	1	1	female	35.0
5	0	3	male	35.0

重复值处理

# 查看去除重复值前数据框的尺寸
data.shape

(891, 4)

data.drop_duplicates(inplace=True)

# 查看去除重复值后数据框的尺寸
data.shape

(349, 4)

查看数据框的缺失值情况

data.isna().sum()

Survived     0
Pclass       0
Sex          0
Age         10
dtype: int64

可以发现原始数据框Age列存在缺失值

查看数据框的摘要信息

data.info()

<class 'pandas.core.frame.DataFrame'>
Int64Index: 349 entries, 1 to 890
Data columns (total 4 columns):
 #   Column    Non-Null Count  Dtype  
---  ------    --------------  -----  
 0   Survived  349 non-null    int64  
 1   Pclass    349 non-null    int64  
 2   Sex       349 non-null    object 
 3   Age       339 non-null    float64
dtypes: float64(1), int64(2), object(1)
memory usage: 13.6+ KB

发现性别列为定性数据，后续需要将其转化为定量数据

数据预处理

处理缺失值

当缺失值数量与样本总数相比时，数量不多，可以选择将缺失值删除处理；当缺失值数量与样本总数相比，数量较多时，不能删除缺失值。

# 删除缺失值
data.dropna(inplace=True)

data.isna().sum()

Survived    0
Pclass      0
Sex         0
Age         0
dtype: int64

现在可以看出已经不存在缺失值

将定性数据转化为定量数据

将类别型数据转化为数值型数据

def Sex2num(str_):
    if str_ == 'male':
        return 0
    return 1

data['Sex'] = data['Sex'].apply(Sex2num)

data.head()

	Survived	Pclass	Sex	Age
PassengerId
1	0	3	0	22.0
2	1	1	1	38.0
3	1	3	1	26.0
4	1	1	1	35.0
5	0	3	0	35.0

分离特征和标签

x = data.iloc[:, -3:]
y = data.iloc[:, 0]

print(x.shape)
print(y.shape)

(339, 3)
(339,)

将数据集切分为训练集和测试集

# 数据集的切分
# 使用留出法将数据集切分为训练集和测试集
from sklearn.model_selection import train_test_split

x_train, x_test, y_train, y_test = train_test_split(x, y, test_size=0.2)
print(x_train.shape)
print(x_test.shape)

(271, 3)
(68, 3)

模型的搭建和训练

# 导入分类树模型
from sklearn.tree import DecisionTreeClassifier

decision_tree_model = DecisionTreeClassifier()

decision_tree_model.fit(x_train, y_train)

DecisionTreeClassifier()

模型测试

y_pred = decision_tree_model.predict(x_test)

打印分类报告

from sklearn.metrics import classification_report

print(classification_report(y_test, y_pred))

              precision    recall  f1-score   support

           0       0.58      0.61      0.59        36
           1       0.53      0.50      0.52        32

    accuracy                           0.56        68
   macro avg       0.56      0.56      0.56        68
weighted avg       0.56      0.56      0.56        68