决策树_实战_波士顿房价

import numpy as np
import sklearn.datasets as sd
import pandas as pd
import matplotlib.pyplot as plt

boston = sd.load_boston()
boston.keys()

boston.filename

print(boston.DESCR)

print(boston.feature_names)

print(boston.data.shape)#x

boston.target.shape

#划分训练集和测试集
import sklearn.model_selection as ms

x = boston.data
y = boston.target

train_x,\
test_x,\
train_y,\
test_y = ms.train_test_split(x,y,test_size=0.1,random_state=7)

res = ms.train_test_split(x,y,test_size=0.1,random_state=7)
res[0].shape#训练集输入数据
res[1].shape# 测试集输入数据
res[2].shape# 训练集输出数据
res[3].shape# 测试集输出数据

#构建决策树模型
import sklearn.tree as st 
import sklearn.metrics as sm

model = st.DecisionTreeRegressor(max_depth=6)
model.fit(train_x,train_y)
pred_test_y = model.predict(test_x)
pred_test_y

sm.r2_score(test_y,pred_test_y)

# Adaboost
import sklearn.tree as st
import sklearn.ensemble as se
model = st.DecisionTreeRegressor(max_depth=4)
model = se.AdaBoostRegressor(model,n_estimators=400,random_state=7)

model.fit(train_x,train_y)
pred_test_y = model.predict(test_x)
sm.r2_score(test_y,pred_test_y)
#获取特征重要性
fi = model.feature_importances_
fi

nzp = pd.Series(fi,index=boston.feature_names)
nzp.sort_values(ascending=False).plot.bar()

#GBDT
model = se.GradientBoostingRegressor(max_depth=6,
                                     n_estimators=400,
                                     min_samples_split=7)
model.fit(train_x,train_y)
pred_test_y = model.predict(test_x)
pred_test_y 

sm.r2_score(test_y,pred_test_y)

params_depth = np.arange(4,14)
params_n = np.arange(400,1000,100)
params_split = np.arange(5,11)
for i in params_depth:
    for j in params_n:
        for k in params_split:
            model = se.GradientBoostingRegressor(max_depth=i,
                                                 n_estimators=j,
                                                min_samples_split=k)
            model.fit(train_x,train_y)
            pred_test_y = model.predict(test_x)
            print(sm.r2_score(test_y,pred_test_y))
#作业：写代码：找出最优的参数组合（最大值索引）