prmr / Creco

Recommendation System for Consumer Products
Apache License 2.0
6 stars 2 forks source link

Features display #32

Closed MariamN closed 10 years ago

MariamN commented 10 years ago
asutcl commented 10 years ago

I believe this to be due to the format of the data in the database. This is why we get a lot of N/A for certain categories while other categories work better. This should be moved to milestone 3 I think and an issue should be opened regarding how to deal with theses.

One of the issues with this right now is that the PCA I am using for the attribute ranking doesn't like missing data so I have to ignore everything that is not complete.

As we explore different methods of ranking we may become more flexible in regards to this.

I've integrated your code so I am closing this.