Sequential Recommendation through Graph Neural Networks and Transformer Encoder with Degree Encoding

Predicting users’ next behavior through learning preferences according to the historical behaviors is known as sequential recommendation. In this task, sequence representation by modeling pairwise relationship between items in capture their long-range dependencies crucial. paper, we propose a novel deep neural network named graph convolutional transformer recommender (GCNTRec). GCNTRec capable of effective item user’s sequence, which involves extracting correlation target node and multi-layer neighbor nodes on graphs constructed under heterogeneous information networks an end-to-end fashion (GCN) with degree encoding, while capturing encoder model. Using multi-dimensional vector representation, related user can be easily predicted. We empirically evaluated multiple public datasets. The experimental results show that our approach effectively predict subsequent relevant outperforms previous techniques.