News for hackers. Stuff that matters.

A comment on Item2Vec: Neural Item Embedding for Collaborative Filtering. Parent comment

Re:Item2Vec: Neural Item Embedding for Collaborative Filtering by etangent on Monday July 11, 2016 @07:41AM (#12070348)
"[before computing SVD], we normalized each entry according to the square root of the product of its row and column sums."

Why didn't they use something that usually works better, like PMI?

[ Parent | on HN ]
Re:Item2Vec: Neural Item Embedding for Collaborative Filtering by rspeer on Monday July 11, 2016 @11:34AM (#12071729)
This is a normalization that I have used and seen other people use. I don't think it's a foregone conclusion that PMI is better for every task.
[ Parent | on HN ]