A Deep Hybrid Collaborative Filtering Based on Multi-dimension Analysis

In order to solve the problem that the existing neural collaborative filtering methods are not comprehensive to mine the latent information of embedded vectors, a deep hybrid collaborative filtering based on multi-dimension analysis is proposed. The idea is to use different feature fusion methods for the embedded vectors of users and items to obtain multiple dimensional fusion features, so that the information explored by different methods can complement each other, and the model can better discover the interaction between users and items. Experimental results show that, compared with the single-method of dimension analysis, the multi-dimension analysis can effectively improve the model's ability to mine the interaction between users and items, and improve the performance of the recommender system.