ef98d58511a0750fc050d1af55f377e9dd4770ff is the major change of this pull request.
I tracked this error down after failing to align t-test results in subpanels b, d, and f to the results shown in a, c, and e (respectively).
A t-test with equal variance assumption will fail to identify cases found in b.
Another minor change is plotting panels b and d with the test-set features only. Previously, panels a and b were calculated using the test set and panels b and d were generated using the training set. Now, b and d use the test set.
I reran the pipeline after these changes and all files here are updated.
ef98d58511a0750fc050d1af55f377e9dd4770ff is the major change of this pull request.
I tracked this error down after failing to align t-test results in subpanels b, d, and f to the results shown in a, c, and e (respectively).
A t-test with equal variance assumption will fail to identify cases found in b.
Another minor change is plotting panels b and d with the test-set features only. Previously, panels a and b were calculated using the test set and panels b and d were generated using the training set. Now, b and d use the test set.
I reran the pipeline after these changes and all files here are updated.
Updated Figure