有人尝试过通过在Python中实现ElasticNetCV和在R中实现cvglmnet来获得相同的结果吗？我已经找到了如何在Python中的ElasticNet和R中的glmnet上实现的方法，但无法用交叉验证方法重现…

在Python中重现的步骤:

预处理:

from sklearn.datasets import make_regressionfrom sklearn.linear_model import ElasticNet, ElasticNetCVfrom sklearn.metrics import mean_squared_errorfrom sklearn.model_selection import train_test_splitimport pandas as pddata = make_regression(    n_samples=100000,    random_state=0)X, y = data[0], data[1]X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=.25)pd.DataFrame(X_train).to_csv('X_train.csv', index=None)pd.DataFrame(X_test).to_csv('X_test.csv', index=None)pd.DataFrame(y_train).to_csv('y_train.csv', index=None)pd.DataFrame(y_test).to_csv('y_test.csv', index=None)

模型:

model = ElasticNet(    alpha=1.0,    l1_ratio=0.5,    fit_intercept=True,    normalize=True,    precompute=False,    max_iter=100000,    copy_X=True,    tol=0.0000001,    warm_start=False,    positive=False,    random_state=0,    selection='cyclic')model.fit(    X=X_train,    y=y_train)y_pred = model.predict(    X=X_test)print(    mean_squared_error(        y_true=y_test,        y_pred=y_pred    ))

输出: 42399.49815189786

model = ElasticNetCV(    l1_ratio=0.5,    eps=0.001,    n_alphas=100,    alphas=None,    fit_intercept=True,    normalize=True,    precompute=False,    max_iter=100000,    tol=0.0000001,    cv=10,    copy_X=True,    verbose=0,    n_jobs=-1,    positive=False,    random_state=0,    selection='cyclic')model.fit(    X=X_train,    y=y_train)y_pred = model.predict(    X=X_test)print(    mean_squared_error(        y_true=y_test,        y_pred=y_pred    ))

输出: 39354.729173913176

在R中重现的步骤:

预处理:

library(glmnet)X_train <- read.csv(path)X_test <- read.csv(path)y_train <- read.csv(path)y_test <- read.csv(path)fit <- glmnet(x=as.matrix(X_train), y=as.matrix(y_train))y_pred <- predict(fit, newx = as.matrix(X_test))y_error = y_test - y_predmean(as.matrix(y_error)^2)

输出: 42399.5

fit <- cv.glmnet(x=as.matrix(X_train), y=as.matrix(y_train))y_pred <- predict(fit, newx = as.matrix(X_test))y_error <- y_test - y_predmean(as.matrix(y_error)^2)

输出: 37.00207

回答：

非常感谢提供的示例，我使用的是笔记本电脑，所以不得不将样本数量减少到100:

from sklearn.datasets import make_regressionfrom sklearn.linear_model import ElasticNet, ElasticNetCVfrom sklearn.metrics import mean_squared_errorfrom sklearn.model_selection import train_test_splitimport pandas as pddata = make_regression(    n_samples=100,    random_state=0)X, y = data[0], data[1]X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=.25)

当你在glmnet中进行预测时，你需要指定lambda，否则它会返回所有lambda的预测结果，所以在R中:

fit <- glmnet(x=as.matrix(X_train), y=as.matrix(y_train))y_pred <- predict(fit, newx = as.matrix(X_test))dim(y_pred)[1] 25 89

当你运行cv.glmnet时，它会从cv中选择最佳的lambda，即lambda.1se，所以它只给你一组结果，这就是你想要的rmse:

fit <- cv.glmnet(x=as.matrix(X_train), y=as.matrix(y_train))y_pred <- predict(fit, newx = as.matrix(X_test))y_error <- y_test - y_predmean(as.matrix(y_error)^2)[1] 22.03504dim(y_error)[1] 25  1fit$lambda.1se[1] 1.278699

如果我们在glmnet中选择最接近cv.glmnet选择的lambda，你会得到正确范围内的结果:

fit <- glmnet(x=as.matrix(X_train), y=as.matrix(y_train))sel = which.min(fit$lambda-1.278699)y_pred <- predict(fit, newx = as.matrix(X_test))[,sel]mean((y_test - y_pred)^2)dim(y_error)mean(as.matrix((y_test - y_pred)^2))[1] 20.0775

在我们与sklearn进行比较之前，我们需要确保我们在相同的lambda范围内进行测试。

L = c(0.01,0.05,0.1,0.2,0.5,1,2)fit <- cv.glmnet(x=as.matrix(X_train), y=as.matrix(y_train),lambda=L)y_pred <- predict(fit, newx = as.matrix(X_test))y_error <- y_test - y_predmean(as.matrix(y_error)^2)[1] 0.003065869

所以我们期望结果在0.003065869的范围内。我们使用相同的lambda运行它，在ElasticNet中lambda被称为alpha。glmnet中的alpha实际上是你的l1_ratio，请参见说明文档。并且normalize选项应该设置为False，因为:

如果为True，回归变量X将在回归前通过减去均值并除以l2范数进行标准化。如果你希望标准化，请在调用带有normalize=False的估计器的fit之前使用sklearn.preprocessing.StandardScaler。

所以我们只需使用CV运行:

model = ElasticNetCV(l1_ratio=1,fit_intercept=True,alphas=[0.01,0.05,0.1,0.2,0.5,1,2])model.fit(X=X_train,y=y_train)y_pred = model.predict(X=X_test)mean_squared_error(y_true=y_test,y_pred=y_pred)0.0018007824874741929

结果与R的结果大致相同。

如果你对ElasticNet这样做，如果你指定alpha，你会得到相同的结果。