From what I've read, F1-score is a commonly used metric to assess the performance of a multilabel classification problem.

However, I recently came across mAP@K and mAR@K as metrics used for recommender systems/information retrieval systems. These values help in judging the model based on the top K items recommended by the system and change as the value of K changes.

I am interested in understanding if mAP@K and mAR@K are valid metrics for a simple multilabel classification problem. I'm aware that in multilabel classification, there is no concept of "retrieving the top K" predictions but I would like to know if my understanding is correct.


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.