Community-Based Collaborative Filtering to Alleviate the Cold-Start and Sparsity Problems

摘要

Recommender systems help users in the discouraging task of selecting through large quantities of information in order to select relevant information or items. It relies on most similar users or items, when the information is large huge number of neighbors gain importance where the goal is to obtain a set of users with whom a target user is likely to match. Forming communities allows us to reveal like-minded users and also reduce the challenges of collaborative filtering like data-sparsity and cold-start problems. This paper proposes a community based collaborative filtering approach based on high correlation and shortest neighbor in the community. We carried out experiments on m1-1m dataset available on MovieLens datasets and experimental results indicate that the quality of the recommendation method is improved compared with traditional algorithms.