决策树分类

决策树分类燕尾花，因为练习的数据不太好找，就直接用燕尾花的数据来练习下·

from sklearn.datasets import load_iris
from sklearn.model_selection import train_test_split
from sklearn.tree import DecisionTreeClassifier
import pandas as pd
data = load_iris()
data1 = pd.DataFrame(data=data.data,columns=data.feature_names)
data2 = pd.DataFrame(data=data.target,columns=['target_name'])
data3 = pd.concat([data1,data2],axis=1)
data3
x = data3[['sepal length (cm)','sepal width (cm)','petal length (cm)','petal width (cm)']]
y = data3[['target_name']]
# 进行数据集划分
x_train,x_test,y_tarin,y_test = train_test_split(x,y)
# 构建模型
classifier = DecisionTreeClassifier()
classifier.fit(x_train,y_tarin)
# 预测值
perdict = classifier.predict(x_test)
# 看下评估的的准确率
y_test.values == perdict
# 评估分数
classifier.score(x_test,y_test)
################################使用随机森林优化决策树#############
# 导入随机森林
from sklearn.ensemble import RandomForestClassifier
# 构建模型
classifier = RandomForestClassifier()
classifier.fit(x_train,y_tarin)
perdict = classifier.predict(x_test)
predict == y_test.values
classifier.score(x_test,y_test)
# 添加网络交叉
from sklearn.model_selection import GridSearchCV
param_dict = {
    "n_estimators":[10,50,100,200,1200],
    "max_depth":[10,20,30,]
}
classifier = GridSearchCV(classifier,param_grid=param_dict,cv=3)
classifier.fit(x_train,y_tarin)
classifier.score(x_test,y_test)

总结，感觉决策树的还是挺nice的，速度又快，准确率也挺高的