HCGL: Modeling dynamics of citation network for impact prediction

Guoxiu He, Zhikai Xue, Zhuoren Jiang, Yangyang Kang, Star Zhao, Wei Lu

Abstract: Highlights•We construct hierarchical and heterogeneous graphs to record dynamics of the citation network .•We propose a novel model H2CGL to aggregate structural and temporal features.•The sensitivity of graph representations to potential citations is improved by contrastive learning.•Experimental results show that our model can reasonably make predictions.