Multi-View Relational Evidential C-Medoid Clustering with Adaptive Weighted

Relational data, where objects are defined by similarities or dissimilarities, is omnipresent and essential in real clustering applications. In this context, relational clustering, which aims to identify groups of similar objects based on their mutual relationships, has become a necessity. However, most existing relational clustering methods cannot effectively handle multi-view data sets while representing uncertainty and imprecision when faced with objects in overlapping clusters. To address this gap, we introduce a new relational clustering method, called Multi-View Evidential C-Medoid clustering with adaptive weightings (MECMdd). Our approach is based on the theory of belief functions to characterize the partial knowledge in cluster assignment. It integrates view weight assignments, estimated locally for each cluster and globally in a collaborative learning framework. We have evaluated our proposition via several experiments using different real-world datasets, compared to other related and state-of-the-art methods, in terms of their advantages and overall effectiveness.