@MehrbodEstaki You can block your cross-validation so that dependent samples (samples from the same subject) are always kept together in the training/test folds. Some background is here: https://t.co/neye48U8p4 And we included an example in our latest p
645 followers
10,172 followers
RT @JakobWirbel: @tpq__ Especially information leakage from the test set is something that we see again and again when reviewing microbiome…
166 followers
RT @JakobWirbel: @tpq__ Especially information leakage from the test set is something that we see again and again when reviewing microbiome…
645 followers
@tpq__ Especially information leakage from the test set is something that we see again and again when reviewing microbiome papers with a machine learning analysis. That's why we included Figure 3 in our recent paper (sorry, personal plug here 😇) to point