Mobile Recommendation Method for Fusing Item Features and User Trust Relationship

摘要

User based collaborative filtering recommendation method does not consider the impact of user preferences on the user's similarity in the non-common score items, and the lack of traditional similarity measurement methods in sparse user score data. This paper proposed a hybrid recommendation method combining similar relationship and trust relationship of mobile users, using the EMD distance method of user preference on similar items to compute the preference similarity relation among the users, and fusing mobile user trust and similar user preferences for the target user's non-scoring items to be scored prediction. Experimental results on public data sets show that, compared to the traditional collaborative filtering recommendation algorithm based on users, this method has a lower MAE error value and higher P@N value, effectively alleviate the data sparsity and improve the performance of the recommendation system.