search

byu-dml / metalearn

BYU's python library of useable tools for metalearning
MIT License
22 stars 6 forks source link

Cross validation #168

Closed ekvdg closed 5 years ago

ekvdg commented 5 years ago

Implemented new cross validation function and wrote test function in test_metafeatures.py to test SKLearn cross validation against new function

bjschoenfeld commented 5 years ago

@ekvdg You don't need to open a new pull request for each new commit. A pull request merges one branch (not a commit) into another. When you make a new commit, the pull request will be updated.

bjschoenfeld commented 5 years ago

The tests do not pass. Please fix those.

bjschoenfeld commented 5 years ago

We have decided to just use sklearn's cross validate functionality and set return_estimator=True. With the estimator, we can call predict twice, once to get the scores we now compute and once to get the probabilities to compute roc auc metrics.