How to determine the hyper-parameters?

[ghost]

In the original paper, the author said he used a randomly sampled one interaction for each user as the validation data, but in your code, you just used test data as the validation data, so I feel confused. Does that mean you used that test data to select the best parameters?

[hexiangnan]

The codes are just a demo to implement and run the model. You can split the validation data from training by yourself, tuning hyperparams on the validation data, and run the codes with optimal hyperparams to get final testing performance.

…

On Tue, Aug 13, 2019 at 7:07 AM Yaowei Hu ***@***.***> wrote: In the original paper, the author said he used a randomly sampled one interaction for each user as the validation data, but in your code, you just used test data as the validation data, so I feel confused. — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#55?email_source=notifications&email_token=AAI3DDWWEKOG66HUOORVVXTQEHULVA5CNFSM4ILFWRNKYY3PNVWWK3TUL52HS4DFUVEXG43VMWVGG33NNVSW45C7NFSM4HE2QQ3Q>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AAI3DDQOQRFOS3JTXWLF7TLQEHULVANCNFSM4ILFWRNA> .