A Spatiotemporal Graph Neural Network for session-based recommendation

Huanwen Wang, Yawen Zeng, Jianguo Chen, Zhouting Zhao, Hao Chen

Abstract: Highlights•Simulate users’ behavior patterns in the session without destroying the click order•Highlight the critical preferences of users during the simulate process.•Use gated fusion functions to combine global behavioral interest and current interest.•Auto-regressive moving average convolution filter evaluates the user preferences.•Use graph neural networks to obtain richer potential information.