Igor Kuivjogi Fernandes

Hi! I think I can try to contribute too.

No problem! I just though would be helpful to tell you.

I think every time you have a new item, you should use `model.fit_partial()` on the pre-trained model, to add this new item to the existing model.

You should pass the index of user_id and the index of item_id inside the `predict()` method. You can see these indexes (they are mapped) doing: `dataset.mapping()`. This will return a...

I would like to know it too: should I pass the entire user_features when predicting new data, or just the correspondence with the user_id?

Do you have an example of doing predictions using matrix multiplication, without having to use `predict()` method? In my case I have user_features in my model.

I think you should ùse an iterator, like this one: ```[x['item_id'], [x['category']]) for x in df.values]``` Note that the second value of the tuple above is a list of list...so...

I think you can try to add user features or item features to the model, trying to avoid cold start problem.

I got this warning when fitting a `XGBSEDebiasedBCE` model: `Warning: Only one class found in a time bucket` Can you explain what it means, please?

This would be nice! Currently I figured out that for my dataset BPR performs much better than ALS. I'm using NDCG@10 to evaluate.