ExaScience / smurff

Bayesian Factorization with Side Information in C++ with Python wrapper
MIT License
70 stars 14 forks source link

prediction empty values #129

Closed jordantkohn closed 4 years ago

jordantkohn commented 4 years ago

Hello,

What's the best way to make predictions for only the empty values in the original dataframe? I've found that predict_some() only returns predictions for cells that aren't empty.

My current method is to use predict_all() but then only extract the predictions in positions that are empty in the original input matrix. Is there a better way to do this?

Thank you

tvandera commented 4 years ago

Hello Jordan,

I think predict_all is currently your best option. If this is too slow, we can look at improving this.

Tom