Transformer Recommendation Machine for Rate Prediction
Xuyang Jin, Jiyang Shen, Tongyuan Miao, Yingjie Wu · 2022
Rate prediction has always been the focus of research in Recommender Systems (RSs), and deep learning based models have achieved remarkable results in this field. Previous works use techniques like FNN to explore new cross features and output the final prediction. Compared with this, there are still relatively few researches on attention-based models. At present, most models only process only one side of users or items through deep networks, and can not use the information of both parties to make a prediction. In this paper, we propose a TRM model that uses both users' and items' information based on attention mechanism. On this basis, it can effectively deal with the cold start problem of users or items.