import statsmodels.api as sm
from sklearn import datasets ## 从 scikit-learn 导入数据集
data = datasets.load_boston() ## 从数据集库加载波士顿数据集
import numpy as np
import pandas as pd
df = pd.DataFrame(data.data, columns=data.feature_names)
target = pd.DataFrame(data.target, columns=["MEDV"])
X = df[['CRIM', 'ZN', 'INDUS']] ## X 通常表示我们的输入变量 (或自变量)
y = target["MEDV"] ## Y 通常表示输出/因变量
X = sm.add_constant(X) ## 我们添加一个截距（beta_0）到我们的模型
model = sm.OLS(y, X).fit() ## sm.OLS(输出, 输入)
predictions = model.predict(X)
model.summary() ## 打印出统计模型

3. 岭回归模型

X_train,X_test,Y_train,Y_test = train_test_split(df2,df1,train_size=0.8)
model = Ridge(alpha=0.5,fit_intercept=True)
model = RidgeCV(alphas=[0.01,0.1,0.2, 0.5, 1],normalize = True,cv=10)
model.fit(X_train,Y_train)
ridge_best_alpha = model.alpha_ #得到最佳lambda值
print(f"岭回归关键正则参数={ridge_best_alpha}")
计算可决系数
a=model.intercept_
b=model.coef_
y_train_pred =model.predict(X_train)
score=model.score(X_test, Y_test)
print(score)

4. 基于最佳lambda值建模

ridge = Ridge(alpha = ridge_best_alpha,normalize = True)
ridge.fit(X_train,Y_train)
ridge_predict = ridge.predict(X_test)
计算损失函数
rmse = np.sqrt(mean_squared_error(Y_test,ridge_predict))

5.  LASSO回归模型：

lasso_cv = LassoCV(alphas = alphas, normalize=True, cv = 10, max_iter=10000)
lasso_cv.fit(x_tr,y_tr)
lasso_best_alpha = lasso_cv.alpha_
lasso_best_alpha
lasso = Lasso(alpha = lasso_best_alpha, normalize=True, max_iter=10000)
lasso.fit(x_tr, y_tr)
lasso_predict = lasso.predict(x_te) #预测
RMSE = np.sqrt(mean_squared_error(y_te,lasso_predict))

本次任务额外知识点：

seed = 7
np.random.seed(seed)
10折交叉验证
kfold = StratifiedKFold(n_splits=10, shuffle=False, random_state=seed)
固定random_state后，每次构建的模型是相同的、生成的数据集是相同的、每次的拆分结果也是相同的

y代表输出答案，y_代表标准答案
mse=tf.reduce_mean(tf.square(Y_test-yy_train_pred))

题目