dask / dask-ml

Scalable Machine Learning with Dask
http://ml.dask.org
BSD 3-Clause "New" or "Revised" License
903 stars 256 forks source link

Alternating least squares #49

Open mrocklin opened 7 years ago

mrocklin commented 7 years ago

Someone mentioned wanting to implement ALS (I think @daniel-severo ?)

I thought I'd raise an issue to solicit and discuss algorithms.

mrocklin commented 7 years ago

A long while ago I implemented a naive implementation for tensor factorizations here: https://nbviewer.jupyter.org/urls/gist.githubusercontent.com/mrocklin/6fc759ab829a44c4f1969a6d6fc9dd28/raw/4fc465c6c202eb8b715762664abc954a1c2c5f4f/als-example.ipynb

dsevero commented 7 years ago

Yeap. Twas I!

Interesting. I work with recommender systems, so I was thinking of the classic collaborative filtering problem for implicit feedback.

I recently started working with textual data and I also wanted to port my ~very obscure~ version of a distributed TF-IDF to this project.

I think this is a pretty strategic decision, since it will draw attention to the project. Maybe we could survey the scientific community to see what makes sense?

jjerphan commented 5 years ago

This issue is related to #139. Should this one be closed?