The official scikit-learn document says "Number of alpha long the regularization path, used for each l1_ratio." Could you tell me what Elasticnet's "n_alpha" argument is?
python machine-learning scikit-learn
In the case of a normal ElasticNet, cross-validation is not performed internally, so you specify a single alpha, but in the case of CV, cross-validation processes the parameters to be grid-searched, so you can specify multiple alpha.If n_alphas is specified and alphas=None is automatically configured as many as n_alphas.ElasticNetCV scores on each of these alpha-trained models and finds the best alpha.
© 2024 OneMinuteCode. All rights reserved.