Abstract
AbstractThe starting point of any hierarchical clustering method is the definition of the similarity measure between objects. In case of a social network the similarity must be inferred from the adjacency relationships between vertices. The main contribution of the paper is to investigate the impact of similarity measure on the quality of clusters (communities) detected in an organizational social network by agglomerative hierarchical clustering method. Three different similarity measures have been considered in the computational experiment. The quality of communities generated by the proposed method has been compared with the quality of clusters generated by the Louvain algorithm using modularity. Because of the fact that the computational experiment has been carried out on the network referring to real public organization, the results of the experiment have been also compared with the structure of the investigated organization.KeywordsOrganizational network analysisCommunity detectionSimilarityHierarchical clustering
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have