Deep Collaborative Recommendation Algorithm Based on Attention Mechanism

Representation learning-based collaborative filtering (CF) methods address the linear relationship of user-items with dot products and cannot study latent nonlinear applied to implicit feedback. Matching function CF directly learn complicated mapping functions that map user-item pairs matching scores, which has limitations in identifying low-rank relationships. To this end, we propose a deep recommendation algorithm based on attention mechanism (DACR). First, before representations are input into DNNs, utilize adaptively assign different weights representations, captures hidden information After that, corresponding representation learning modules. Finally, concatenate prediction vectors learned from dimensions predict scores. The results show can improve expression ability model while taking account not only feedback, but also relationships obtain more accurate predictions. Through detailed experiments two datasets, find ranking capability DACR is enhanced compared other baseline models, evaluation metrics HR NDCG increased by 0.88–1.19% 0.65–1.15%, respectively.