I want to use the regularized regression technique such as partial least squares (PLS), Lasso, and the elastic-net. However, as far as I know, there is no "standard" way of calculating the prediction interval for those methods. So can I use bootstrap (resample the training data, calculate the prediction) as a way to estimate the prediction interval? Will this method result in accurate coverage of the true mean response?

  • $\begingroup$ stats.stackexchange.com/questions/241082/… $\endgroup$ Nov 30, 2018 at 7:52
  • $\begingroup$ Are you asking about a confidence interval (on the coefficients), or prediction interval (on the output)? $\endgroup$
    – user20160
    Nov 30, 2018 at 11:28
  • $\begingroup$ @user20160 I am asking about the prediction interval $\endgroup$
    – Vickyyy
    Nov 30, 2018 at 19:42


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.