Closed youngahn closed 7 years ago
This is related to how data are normalized (i.e., z-scored).
With the following setting, the results are replicated. Currently the default of 'preprocess' is NULL and I think its default should be set to 'preprocess_scaler'.
preprocess = preprocess_scaler,
The glmnet results using easyml are very different from what's reported in Ahn et al (2016). Need to verify.