HG-PerCon: Cross-view contrastive learning for personality prediction

Meiling Li, Yangfu Zhu, Shicheng Li, Bin Wu

Abstract: Highlights•Proposes a mechanism for extracting long-lasting user historical semantic information.•Introduces a psychological knowledge graph that enhances user personality representation.•Contrastive learning is used to learn the consistency of user personality-related representations across views, thus achieving a good representation for personality prediction task.