08.recommender systems
The two most common types of recommender systems are
- Content-Based
- Collaborative Filtering (CF)
Collaborative filtering produces recommendations based on the knowledge of user's attitude to items, that is it uses the 'wisdom of the crowd' to recommend items.
Content based recommender systems focus on the attributes of the items and give you recommendations based on the similarity between them.
In general, Collaborative filtering (CF) is more commonly used than content-based systems because it usually gives better results and is relatively easy to understand (from an overall implementation prespective.
ALS is basically a matrix factorization approach to implement a recommendation algorithm. We decompose our large user item matrix into a lower missional user factors in item factors.