Outline a Smaller Class With the Custom Loss Function

A short guide to you for taking the most from classification when dealing with imbalanced datasets

(Left) Photo by Mike Lawrence on Flickr | (Middle) Photo by Jernej Furman on Flickr | (Right) Photo by Leonard J Matthews on Flickr

Why use a custom loss function?

There may be situations when the accuracy metric is insufficient to get the expected results. We may want to reduce the false negative (FN) or false positive (FP) …

classification-algorithms cross-entropy function log-loss loss machine learning xgboost