An Implicit Information Based Movie Recommendation Strategy
Jie Chen, Junjie Peng, Yingtao Wang, Gan Chen · 2018
Movie recommendation is a common way to attract audiences and enhance audience ratings in movie fields. The effectiveness of recommendation depends on the recommendation algorithm. To the traditional recommendation algorithms, such as user based collaborative filtering, matrix factorization based collaborative filtering and so on, the recommendation effect is strongly dependent on the users' ratings on the movies.However, in the actual movie online platform, the explicit rating data is very rare, which makes the recommendation result not very satisfactory. In view of this situation, this paper proposes an implicit information based recommendation strategy which can be used to establish user's preferences and make recommendation. By analyzing the watching logs, the strategy draws out the historical behavior and preference information of users which can be utilized as the basis of recommendation. Extensive experiments show that the strategy is correct, it can effectively analyze users' preferences on different factors on different movies. Meanwhile, it can make recommendation with high accuracy which can be used as the basis of recommendation without explicit information.