When trying to choose the value of the regularisation parameter(s) in lasso, ridge regression, or elasticnet, one generally computes the cross validated error or negative log-likelihood as explained by this answer. Further down the page, another answer shows an example of a plot of the cross validated error against the regularisation parameter $\lambda$, reproduced here:

enter image description here

My question is whether the expected value of this error (or NLL as relevant), in any regression setting, has one and only one minimum over the possible values of the regularisation parameters?


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.