I need to train an LSTM on some light curves, in order to find a signal (there are 2 classes signal and background). However the signal (data points corresponding to signal) is around 100 times less frequent than the background so I have a huge class imbalance and in the end all points are labelled as background. I tried to use focal loss, but it doesn't help. Is there a way to make it work?


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.