A Social Network-based serendipity recommender system

The appearing of the internet brings a large amount of information, this makes searching and filtering difficult. Therefore, a kind of special data mining technique appeared, it called Recommender System. Most research of recommender system always provides the most relevant items for users or items. However, recommendations from traditional recommender system may not satisfy the new human beings because users may already know these relevant items from other information sources. We believe that there are still some unsearched but less relevant items useful for users. On the other hand, because social network has grown very quickly, we think that there are some very useful interactive information that recommender systems can use to provide recommendations. Therefore, we propose a Social Network-based Serendipity (SNS) recommender system that uses interactive information from the social network to find out which items are interesting for users but hard to discover by themselves.