MovieRecommender about machinelearning-samples HOT 3 CLOSED

I added a movie recommendation sample based upon Matrix Factorization.

https://github.com/dotnet/machinelearning-samples/tree/master/samples/csharp/getting-started/MatrixFactorization_MovieRecommendation

There are also two other kind of samples we will work on adding for recommendation.

1. One Class Matrix Factorization: In some use cases a web portal or store might only have data on what items a user has purchased but not have ratings on these products. For such a scenario One Class Matrix Factorization will work better.

2. Field Aware Factorization Machines: Matrix Factorization is limited to two features (u, v) representing the user ids and movie ids per user in the movie recommendation example, what happens when you want to use more features like movie genre, movie cast, year of release etc. Field Aware Factorization Machines based recommendation approach would work better in that scenario.

Hope that helps.

from machinelearning-samples.

I have also added the Field Aware Factorization Machines based sample to the repo. You can try it out here.
https://github.com/dotnet/machinelearning-samples/tree/master/samples/csharp/end-to-end-apps/Recommendation-MovieRecommender

from machinelearning-samples.

@asthana86 I have one issue with Field Aware Factorization Machines, as the described, it uses more features like movie genre, movie cast, but in your demo sample, it still uses userId and movieId, would you mind sharing us the complete demo with other features?

from machinelearning-samples.