search

ankane / disco

Recommendations for Ruby and Rails using collaborative filtering
MIT License
565 stars 11 forks source link

Building a Hybrid Approach #11

Closed jesster2k10 closed 3 years ago

jesster2k10 commented 3 years ago

Hey, thanks for this library! I'm trying to work out how I could go about building a hybrid recommendation system that combines collaborative filtering, content based filtering and also knowledge based recommendations rather than one of them exclusively. Do you have any suggestions how to go about this, or where to start? I've read through the README and you suggested using the https://github.com/jpmckinney/tf-idf-similarity gem for content-based recommendations to remedy the cold-start issue; but how would this look like in a real-world example as a hybrid system?

ankane commented 3 years ago

Hey @jesster2k10, the simplest hybrid approach would be to use Disco for user-based and/or item-based recommendations and tf-idf-similarity or similar for content-based recommendations, and then show users recommendations from each method. You can use has_recommended to store recommendations generated from any method.