I am working on multilabel binary classification task. Dataset have 5 labels each have value 0 or 1. The dataset is highly imbalanced with 94% and 6% ratio. Working with Keras, I have used binary cross entropy loss function but I want to assign higher weights to class having lower frequency. I have searched Keras: weighted binary crossentropy but it is for binary classification problem. Can I use this https://www.tensorflow.org/api_docs/python/tf/nn/weighted_cross_entropy_with_logits for this kind of problem?
Asked
Active
Viewed 592 times
