I would like to found out what's happening if a customized weights list (not all equalling to 1 as default) is passed to glmnet()
. I have looked into the glmnet
package @github: glmnet but could not figure out what's going underneath. Is it the same as up-sampling by duplicating the minority observations?
Asked
Active
Viewed 930 times
0

CcMango
- 377
- 1
- 4
- 15
-
1Something like this? https://stats.stackexchange.com/questions/196614/what-are-the-weights-for-in-predict-function-in-r – Roman Luštrik Nov 27 '17 at 19:57
-
Thank you for your reply. But I don't think that's the answer. I know how to assign `weights` to `glmnet()` or `cv.glmnet()`. What I want to do is to understand how the weights will be used and what are the effects of using weights. – CcMango Nov 28 '17 at 07:08
1 Answers
0
In glmnet vignette by Trevor Hasties et al, "weights" is explained to be used to reconstruct deviance during convergency process, therefore impacting the final fitted model.

CcMango
- 377
- 1
- 4
- 15