Towards an Introduction to Collaborative Filtering

This paper makes a comprehensive introduction to collaborative filtering aiming to facilitate readers to understand this field. Firstly the research background is introduced, including mainstream collaborative filtering applications, research conferences and journals. Then this paper briefly describes the current generation of collaborative filtering that are usually classified into the following two main categories: heuristic-based and model-based methods. In addition, the widely adopted benchmarked datasets and their characteristics are exhibited. This paper also describes various limitations of current collaborative filtering methods and discusses possible extensions that can improve recommendation capabilities and make collaborative filtering systems applicable to an even broader range of application.