My test data are imbalanced, i tried to use the precision or the gmean as metrics for a multi-label learning model, but both metrics are not very informative. Is there any way to use for example the precision that gives information about the minor class and the accuracy in a single procedure. Any other solution for this problem is welcomed.

Thanks in advance.


Maybe you should instead try to use some techinique to rebalance your dataset and then use a classical metric for classification problems. Example of tchniques for rebalancing the dataset for binary classification is given in the following post: https://stats.stackexchange.com/questions/152823/how-to-balance-my-dataset If you still want to work on the unbalanced dataset in the following paper some s are suggested: https://eva.fing.edu.uy/pluginfile.php/69453/mod_resource/content/1/7633-10048-1-PB.pdf

| improve this answer | |
  • $\begingroup$ Thanks lorenzo for the response. If i am understanding, you suggest to balance test data, as i balanced only training data to avoid overfitting. If yes, is it recommanded to predict new data(after balancing test dataset). $\endgroup$ – Born New May 23 '19 at 12:29

Your Answer

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

Not the answer you're looking for? Browse other questions tagged or ask your own question.