LinearRegression模型与SGDRegressor模型的性能比较

本文是以《Python机器学习及实践 从零开始通往kaggle竞赛之路》为参考书籍进行的实践

1  LinearRegression模型与SGDRegressor模型的性能比较

  利用对数据处理的结果进行性能的比较；数据源为：美国波士顿地区房价数据

2 实验代码及结果截图

#coding:utf-8

#导入数据读取器
from sklearn.datasets import load_boston
boston=load_boston()
#输出数据描述
print boston.DESCR

#数据分割
from sklearn.cross_validation import train_test_split
import numpy as np
X=boston.data
y=boston.target

#随机25%的测试样本数据，其他为训练样本数据
X_train,X_test,y_train,y_test=train_test_split(X,y,random_state=33,test_size=0.25)
#分析回归目标值的差异
print '差异'
print 'The max target value is:',np.max(boston.target)
print 'The min target value is:',np.min(boston.target)
print 'The avg target value is:',np.mean(boston.target)

#数据的标准化处理
from sklearn.preprocessing import StandardScaler
ss_X=StandardScaler()
ss_y=StandardScaler()

X_train=ss_X.fit_transform(X_train)
X_test=ss_X.fit_transform(X_test)
y_train=ss_y.fit_transform(y_train.reshape(-1,1))
y_test=ss_y.fit_transform(y_test.reshape(-1,1))
(此处与书上的代码不同，书上的代码在运行会出现错误！）
#LR模型的预测分析
from sklearn.linear_model import LinearRegression
lr=LinearRegression()
#参数估计
lr.fit(X_train, y_train)
#预测分析
lr_y_predict=lr.predict(X_test)


#SGD模型
from sklearn.linear_model import SGDRegressor
sgdr=SGDRegressor(max_iter=5)
sgdr.fit(X_train, y_train.ravel())
sgdr_y_predict=sgdr.predict(X_test)

#性能评估
print 'LR:',lr.score(X_test,y_test)
from sklearn.metrics import r2_score,mean_squared_error,mean_absolute_error
print 'R-squared:',r2_score(y_test, lr_y_predict)
print 'mean_squared:',mean_squared_error(ss_y.inverse_transform(y_test), ss_y.inverse_transform(lr_y_predict))
print 'mean_absolute:',mean_absolute_error(ss_y.inverse_transform(y_test), ss_y.inverse_transform(lr_y_predict))

print 'SGDC:',sgdr.score(X_test,y_test)
print 'R-squared:',r2_score(y_test, sgdr_y_predict)
print 'mean_squared:',mean_squared_error(ss_y.inverse_transform(y_test), ss_y.inverse_transform(sgdr_y_predict))
print 'mean_absolute:',mean_absolute_error(ss_y.inverse_transform(y_test), ss_y.inverse_transform(sgdr_y_predict))