Nicolas Hug

>fix_k_closest - is clear, together with a good param description **No, it's not**. The number of neighbors is not fixed **in either of the methods**. >Previously you agreed on the...

>k neighbors are absolutely fixed in the method that I advocate. I understand what you mean: they're fixed in the sense that for a given user,we will always look at...

Awesome! :) For this I think we should try to follow as close as possible what scikit learn is doing (see e.g. their [conventions](http://scikit-learn.org/stable/developers/contributing.html#estimators) for estimators). There's no need to...

Sure go ahead :)

Sure. Have you looked at #168? It seems that the bulk of the work is done already and what's left is a benchmark to assess how fast / accurate spearman...

Hmm looks like I cannot assign you to the issue because you don't have write rights. It's not important though so don't worry.

I think a good way to do it would be to pull the PR branch (look [here](https://scikit-learn.org/stable/developers/tips.html#checking-out-pull-requests-as-remote-tracking-branches)) in your own fork / clone. Then you can either directly work on...

If by column you mean additional "features" for items or users, then no, this isn't something that surprise can support

The algorithms in surprise don't have a fallback to specifically handle cold start but some (e.g. SVD) will naturally fallback to the mean of the ratings if there are no...

Thanks a lot, Once again I really appreciate the efforts with the docs and the tests. I'm definitely interested in adding FM into surprise! This is a lot of code...