ADeepCF: Deep Collaborative Recommendation Model Incorporating Attention Mechanisms
Weiwei Li, Xuanlin Pan, Zaiqu Fu · 2022
In real-life application scenarios there are often problems such as low user ratings of items and difficulties in collecting explicit user feedback. Since collaborative filtering algorithms that use only user and item ID information as input cannot perform its features well and the powerful feature learning ability of the attention mechanisms, this paper proposes a deep collaborative recommendation algorithm model incorporating an attention mechanism. The implicit feedback is integrated with the ID information as input, which avoids the user’s random scoring and alleviates the cold start problem. Introducing attention mechanisms to enhance the ability of deep neural networks to learn non-linear features of users and items. Finally, experiments are conducted on a publicly available dataset to verify the effectiveness of the algorithms in the paper.