jselma

> See this [issue](https://github.com/massquantity/LibRecommender/issues/20). Possible cause is that your system is Windows... > You can clone all the code and cd into the `LibRecommender` folder. Then maybe you can use...

Could you correct the "setup.py" file to make it compatible with Windows?  

Hi @massquantity, I could use the library by copying it into the python environment folder, however it does not load ALS Cython. The other algorithms like SVD ++ work without...

> Alright, I've released a new version, so you can give it a try. > > > pip install LibRecommender==0.2.2 The installation works perfect !!, **thank you very much !!!!!!**

> ALS is implemented by Cython, and SVD++ is implemented by TensorFlow. When you're coping the library, the ALS Cython version doesn't build. To use ALS, you have to install...

Hi @massquantity. The value of p@t seems very strange to me, understanding that I am providing 300 factors and an alpha of 40. It indicates a precision of less than...

> > My dataset has around 1,500,000 observations. In the column "label" there are values ​​between 1 and 20 that represent the number of reproductions per item, in addition to...

> Someone has encountered the same problem... See this [issue](https://github.com/massquantity/LibRecommender/issues/12) > > > How do you order users and trained items for predictions? I imagine the function split_by_ratio or DatasetPure.build_trainset...

> > The value of p@t seems very strange to me, understanding that I am providing 300 factors and an alpha of 40. It indicates a precision of less than...

> I'm gonna tell you, man, you've touched a very subtle point in this library:) > As you may have already known, ALS is quite different from other algorithms such...