Co‐Clustering Analysis of Multi‐Layer Directed Networks: A Spectral Approach

ABSTRACT This article focuses on community detection in multi‐layer‐directed degree‐corrected block models. A spectral co‐clustering method building upon a bias‐adjusted aggregation matrix is proposed to identify the community memberships. We theoretically demonstrate the efficacy of our approach by controlling the proportion of misclustered nodes. Both simulations and real‐data analysis provide empirical evidence supporting the effectiveness of our proposed methodology.