README.md

Good-reads-recommender

Recommendation systems (Content-based and collaborative filtering)

In the notebook there are 3 implementations for recommendation systems:

1. Content-based using the pre-trained BERT vectorizer to recommend similar books.
2. Collaborative filter using item matrix and knn to recommend similar users (friends'suggest)
3. Collaborative filter using the Surprise library, just to calcualte the precision and the recall for the data (Surprise do everything by its own we don't have to do any thing but pass out the data)

Note: if you want to run first tow implementations you need to use GCP