I have a highly imabalanced dataset but one that is not sparse. In train there are 1328 positives out of 104000. In validation and test set there are 400 positives out of 103000 Which algorithm should I experiment with: Boosting (LGBM/XGBoost) or FFM (field aware factorization machines)?

  • $\begingroup$ Why not try both and pick the one that validates the best? I expect lightGBM to be superior here for what it is worth since you state your dataset is not sparse, where FFM tend to do better. $\endgroup$ – aranglol Sep 10 '19 at 14:32

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.