Python 决策树实战（数据挖掘原理、算法及应用例子）

一、代码

import pandas as pd
f = open('weather.csv')
data = pd.read_csv(f)
data.drop(['属性'], axis = 1, inplace = True)  # axis = 1纵向删除
print(data)


# 将数值型的Pclass转换为类别型，否则无法对其哑变量处理
# Titanic.Pclass = Titanic.Pclass.astype('category')
# 哑变量处理
dummy = pd.get_dummies(data[['Outlook','Temperature','Humidity','Windy']])
# 水平合并Titanic数据集和哑变量的数据集
data = pd.concat([data,dummy], axis = 1)
# 删除原始的'Outlook','Temperature','Humidity','Windy'变量
data.drop(['Outlook','Temperature','Humidity','Windy'], inplace=True, axis = 1)

for i in range(len(data.Class)):
    if data.Class[i] =='No':
        data.Class[i] = 0
    elif data.Class[i] =='Yes':
        data.Class[i] = 1
    else:
        data.Class[i] = 2 
        
data.head()


# 导入第三方模块
from sklearn.model_selection import GridSearchCV
from sklearn import tree
# 预设各参数的不同选项值
max_depth = [2,3,6,8]
min_samples_split = [2,3,4,6]
min_samples_leaf = [2,3,4,6]
# 将各参数值以字典形式组织起来
parameters = {'max_depth':max_depth, 'min_samples_split':min_samples_split, 'min_samples_leaf':min_samples_leaf}
# 网格搜索法，测试不同的参数值
grid_dtcateg = GridSearchCV(estimator = tree.DecisionTreeClassifier(),scoring='accuracy', param_grid = parameters, cv=2)
# 模型拟合
grid_dtcateg.fit(X_train, y_train)
# 返回最佳组合的参数值
grid_dtcateg.best_params_


def creat_tree(data):
    # 需要在电脑中安装Graphviz
    # https://graphviz.gitlab.io/_pages/Download/Download_windows.html
    # 然后将解压文件中的bin设置到环境变量中
    # 导入第三方模块
    from sklearn.tree import export_graphviz
    from IPython.display import Image
    import pydotplus
    from sklearn.externals.six import StringIO
    from sklearn import model_selection
    from sklearn import metrics
    from sklearn.tree import DecisionTreeClassifier
  
    #调用dataset()函数获取数据
    predictors = data.columns[1:]
    # 将数据集拆分为训练集和测试集，且测试集的比例为25%
    X_train, y_train = data[predictors], data['Class'].astype('int')
    # 构建分类决策树
    CART_Class = DecisionTreeClassifier(max_depth=3, min_samples_leaf = 2, min_samples_split=2,criterion='entropy')
    # 模型拟合
    decision_tree = CART_Class.fit(X_train, y_train)

    # 绘制决策树
    dot_data = StringIO()
    export_graphviz(
        decision_tree,
        out_file=dot_data,  
        feature_names=predictors,
        class_names=['No','Yes'],  
        # filled=True,
        rounded=True,  
        special_characters=True
    )
    # 决策树展现
    graph = pydotplus.graph_from_dot_data(dot_data.getvalue())
    
    graph.write_pdf('data1.pdf') #保存为决策树
    return Image(graph.create_png()) 
结果：