Community-Aware Role-Based Graph Embedding

Graph embeddings are essential for analyzing complex networks across domains like social networks, biology, and ecology. However, existing methods often fail to integrate structural roles and mesoscopic community patterns, treating these aspects separately. To address this gap, we propose Community-Aware Role-Based Graph Embedding (CRGE), a novel framework that combines role-based feature extraction with hierarchical random walks tailored to extract community structures. CRGE extracts robust structural features through recursive aggregation and integrates them with community-aware random walks to generate comprehensive node embeddings. Experiments on diverse benchmark datasets demonstrate CRGE’s superior performance, achieving up to 3.5% higher accuracy in link prediction task compared to state-of-the-art baselines. By bridging the strengths of role-based and community-aware approaches, CRGE offers a scalable and adaptable solution for graph analysis, with applications in social network analysis, bioinformatics, and recommendation systems.