Knowledge enhanced graph contrastive learning for match outcome prediction

Junji Jiang, Likang Wu, Zhipeng Hu, Runze Wu, Xudong Shen, Hongke Zhao

Abstract: Highlights•A novel method called KEGC is proposed to solve the match outcome prediction problem.•The proposed framework outperforms the 5 comparison methods and 3 variants on the experiments of two real-world game match datasets.•KEGC introduces variational mechanism to model the uncertain nature of player performance in game.•A joint loss function both optimizes match outcome prediction and balance prediction to mitigate data noise.