李沐基于Pytorch的深度学习笔记（8）-基础优化算法（附代码）

from sklearn.datasets import load_boston
import joblib
from sklearn.linear_model import LinearRegression, SGDRegressor, Ridge
from sklearn.metrics import mean_squared_error
from sklearn.model_selection import train_test_split
from sklearn.preprocessing import StandardScaler


"""线性回归预测房子价格"""
# 获取数据
lb = load_boston()
print(lb)
# 分割数据集 （训练集、测试集）
x_train, x_text, y_train, y_test = train_test_split(lb.data, lb.target, test_size=0.25)
# 进行标准化
# 特征值 目标值 都标准化
std_x = StandardScaler()
x_train = std_x.fit_transform(x_train)
x_text = std_x.fit_transform(x_text)

std_y = StandardScaler()
y_train = std_y.fit_transform(y_train.reshape(-1, 1))  # 传入数据必须二维
y_test = std_y.fit_transform(y_test.reshape(-1, 1))
# estimator预测
# 正规方程求解方式
lr = LinearRegression()
lr.fit(x_train, y_train)

# print(lr.coef_)

y_lr_predict = std_y.inverse_transform(lr.predict(x_text))
# print("预测：\n", y_lr_predict)
print("均方误差：", mean_squared_error(std_y.inverse_transform(y_test), y_lr_predict))
print(f"R : {lr.score(x_test, y_test)}")
# 梯度下降方式
sgd = SGDRegressor()
sgd.fit(x_train, y_train)

y_sgd_predict = std_y.inverse_transform(sgd.predict(x_text))
print(y_sgd_predict)
print("均方误差：", mean_squared_error(std_y.inverse_transform(y_test), y_sgd_predict))
print(f"R : {sgd.score(x_test, y_test)}")

本文参考的是下面这篇博文：

这篇文章里有些东西其实已经用不了了，我自己改了下

线性回归及python代码实现___zachary的博客-CSDN博客_python线性回归代码