Attributed graph clustering under the contrastive mechanism with cluster-preserving augmentation

Yimei Zheng, Caiyan Jia, Jian Yu

Abstract: Highlights•An attributed graph clustering method under the contrastive mechanism is proposed.•A cluster-aware contrasting view based on EBC and kNN graph is built.•Multilevel contrast and self-supervised clustering are jointly optimized.•Extensive experiments demonstrate the superiority of the model in clustering.