Detecting Community Structures in Patients with Peripheral Nervous System Disorders

Identifying sub-networks showing similar properties, referred to as community detection, is a challenging task in network analysis. This challenge becomes even more formidable in bipartite networks. The focus of this study is the patients with problems in their Peripheral Nerve System. To this aim, we engaged the assistance of spinal specialty clinics in the collection of necessary Data. We employ the bipartite network to represent the relationship between the patients and their symptoms and disorders. The resulting bipartite network showcases unequally sized sets of nodes, making community detection more challenging. The principal purpose of this study is to develop a new, practically relevant method for finding communities inside such networks. As such, we propose the Bi-MRComSim algorithm which applies different methods to transform the bipartite network to a unipartite one that can find meaningful communities between patients that coincide 85% of the time with diagnoses issued by physicians.