How do I Know What You Like ? The Ways of Collaborative Filtering

Recommendation systems are more and more needed because of the huge amount of information available on the internet. The method of making automatic predictions about the interests of a user by collecting taste information from many users is called collaborative filtering. The underlying assumption of collaborative filtering is that: If there are people with similar preferences they would rate analogous things similar. There are many different methodologies to predict an unknown user rating based on other user ratings. In this article I give an overview about the various approaches. Memory-based collaborative filtering algorithms predict the rating of an item based on the ratings of users who are similar to the selected user. It’s possible to describe the problem as finding a lowrank approximation to a partial observed rating matrix. And there are other researchers who were using neural networks or other additional data to overcome the above mentioned challenge.