A Bio-inspired Optimization Technique for Cluster Ensembles Optimization

Abstract

Several clustering algorithms have been applied to a great variety of problems in different application domains. Each algorithm, however, has its own advantages and limitations, which can result in different solutions for the same problem. In this sense, combining different clustering algorithms (cluster ensembles) is one of the most used approaches, in an attempt to overcome the limitations of each clustering technique. The main aim is to combine multiple partitions generated by different clustering algorithms into a single clustering solution (consensus partition). To date, several approaches have been proposed in literature in order to provide optimization, or continuously improve the solutions found by the cluster ensembles. Therefore, as a contribution to this important subject, this paper presents a new bio-inspired optimization technique to optimize the cluster ensembles. In this proposed technique, the cluster ensembles are heterogeneously created and the initial partitions are combined through a method which uses the Coral Reefs Optimization algorithm, resulting in a consensus partition. In order to evaluate the feasibility of the proposed technique, an empirical analysis will be conducted using 15 different problems and applying two different indexes in order to examine its efficiency and feasibility.