Hyperbolic multivariate feature learning in higher-order heterogeneous networks for drug-disease prediction

Jiamin Li, Jianrui Chen, Junjie Huang, Xiujuan Lei

Abstract: Highlights•Higher-order heterogeneous networks are constructed to obtain richer and reliable neighborhood information.•We explore a multivariate feature learning mechanism to enhance feature representation in hyperbolic space.•Hyperbolic graph attention mechanism is used to allow each drug to select the most related diseases to apply its filter.