A hybrid recommendation algorithm based on time factor

摘要

With the development of social network, helping users find their interesting information become a primary objective for recommend systems. As a most popular recommend algorithm, collaborative filtering recommendation still remains some shortcomings such as data sparseness, cold start and neglecting of variable user interests. With this in mind, a hybrid recommendation algorithm based on time factor is proposed in this paper. A hybrid recommendation model based on time factor aiming to improve the accuracy of user similarity calculations is proposed. This recommendation model includes the user rating, content feature and time factor. Then, the particle swarm optimisation (PSO) algorithm is exploited to optimise the searching space. The experimental results show that the proposed algorithm can effectively improve accuracy while solving data sparseness and cold start. It can be used in the social network and e-commerce.