Nouvelle approche de clustering par kernel-pattern via la densité en triades

Community detection has become a major active area of research in recent years. A plethora of relevant methods have been implemented for directed graphs. Most of them focus on the density of links, and consider the relationship between nodes as symmetric by ignoring links directionality during their clustering step, this leading to non-semantic results. This paper propose an efficient method based on the extraction of kernels through the distribution of triads in the graph, using Kernel Degree Clustering (KDC) a novel metric to judge the quality of a community partitioning, demonstrated to yield superior results over other commonly used metrics like modularity in conformity with centrality. To validate our approach, we conduct experiments on some networks which show that it has better performance over some of the other state-of-the-art methods and uncovers expected communities. MOTS-CLÉS : Réseaux orientés, Détection des communautés kernel, Clusters basés sur la structure, Triade.