Abstract
The study of Social Networks displays a hierarchy of communities which highlights the structure and the hierarchical interactions between social network users'. In fact, the hierarchical community detection algorithms can be divided into two groups, namely the agglomerative and the divisive algorithms. However, we propose, in this paper, hybrid hierarchical method for community detection based on both hierarchical groups. The introduced hybrid hierarchical algorithm assumes the existence of an initial partition. Because the input of the algorithm has a significant influence its output, we aim at avoiding the random generation of this initial partition. To achieve this purpose, we treated community detection problem as a combinatorial optimization issue, more especially as a Multi Objective Knapsack Problem. We also proposed Tabu Search metaheuristic to solve this issue. Moreover, we performed some comparative experiments to enhance the quality of the clustering results and to show the effectiveness of our algorithm.