Is there a way in R to apply bootstraps across various imbalance ratios (using downsampling)? What I am interested in doing is testing various downsampling combinations, e.g., 50/50, 60/40, 70/30, etc., to see effect on model performance. Thinking bootstrapping may be a way to test the different combinations?
Asked
Active
Viewed 22 times
0
-
Your question is a little bit hard to answer, because it is not reproducible, but I think you might be looking for caret::downSample(). Have a look at https://stackoverflow.com/questions/66933618/how-to-repeat-downsample-in-r, or https://www.r-bloggers.com/2019/04/methods-for-dealing-with-imbalanced-data/ – Otto Kässi May 17 '23 at 08:02