岭回归案例分析
def linearmodel():
"""
线性回归对波士顿数据集处理
:return: None
"""
ld = load_boston()
x_train,x_test,y_train,y_test = train_test_split(ld.data,ld.target,test_size=0.25)
std_x = StandardScaler()
x_train = std_x.fit_transform(x_train)
x_test = std_x.transform(x_test)
std_y = StandardScaler()
y_train = std_y.fit_transform(y_train)
y_test = std_y.transform(y_test)
lr = LinearRegression()
lr.fit(x_train,y_train)
y_lr_predict = lr.predict(x_test)
y_lr_predict = std_y.inverse_transform(y_lr_predict)
print("Lr预测值:",y_lr_predict)
sgd = SGDRegressor()
sgd.fit(x_train,y_train)
y_sgd_predict = sgd.predict(x_test)
y_sgd_predict = std_y.inverse_transform(y_sgd_predict)
print("SGD预测值:",y_sgd_predict)
rd = Ridge(alpha=0.01)
rd.fit(x_train,y_train)
y_rd_predict = rd.predict(x_test)
y_rd_predict = std_y.inverse_transform(y_rd_predict)
print(rd.coef_)
print("lr的均方误差为:",mean_squared_error(std_y.inverse_transform(y_test),y_lr_predict))
print("SGD的均方误差为:",mean_squared_error(std_y.inverse_transform(y_test),y_sgd_predict))
print("Ridge的均方误差为:",mean_squared_error(std_y.inverse_transform(y_test),y_rd_predict))
return None