Jointly modeling intra- and inter-session dependencies with graph neural networks for session-based recommendations

Jingjing Wang, Haoran Xie, Fu Lee Wang, Lap-Kei Lee, Mingqiang Wei

Abstract: Highlights•Propose a repeat-aware local graph to solve the repeat consumption phenomenon.•Simplify neighbors retrieve from global graph built from mini-batches.•Consider the node degree as node attribute to debias the popularity influences.