Using collaborative filtering approaches, an RE can be built based on a user's past behavior where numerical ratings are given on purchased items. Sometimes, it can be developed on similar decisions made by other users who also have purchased the same items. From the following figure, you can get some idea of different recommender systems:
Figure 1: A comparative view of different recommendation systems
Collaborative filtering-based approaches often suffer from three problems—cold start, scalability, and sparsity:
- Cold start: Sometimes gets stuck when a large amount of data about users is required for making ...
