Abstract
Although model-based Collaborative Filtering approaches have been widely used in recsys in the past few years. But In practical application, where users’ rating data arrives sequentially and frequently, the model-based approaches have to re-trained completely for new records. For users’ social information has been succeed used in recommendation system in previous work. In this paper, we proposed several online collaborative filtering algorithms using users’ social information to improve the performance of online recommender systems. The algorithms can better use the prior rating and the social network information, which compute fast and scalable in large data. The contribution of this paper are mainly two-fold: (1) We propose an online collaborative filtering algorithm which can better use the social information and prior knowledge; (2) We solve the problem of cold start and users with few ratings.
Published Version
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call