Temporary fix for convex combination method negative log-likelihood loss Relative to existing implementation, we reduce the tolerance at which we declare predictions from a given algorithm the same as another. Note that because of the way SuperLearner is structure, one needs to install the optimization software separately.

Description

Temporary fix for convex combination method negative log-likelihood loss Relative to existing implementation, we reduce the tolerance at which we declare predictions from a given algorithm the same as another. Note that because of the way SuperLearner is structure, one needs to install the optimization software separately.

Usage

tmp_method.CC_nloglik()