Resumen de A Hybrid Bio-Inspired Tabu Search Clustering Approach

Dragan Simić , Zorana Banković, José Ramón Villar Flecha , José Luis Calvo Rolle , Svetislav D. Simić, Svetlana Simić

• The purpose of a data clustering process is to group a set of objects into multiple classes so that the objects in each class – cluster are similar according to certain rules or criteria, where the definition of similarity can be problem dependent. This paper is focused on a new bio-inspired clustering approach based on a model for combining tabu search algorithm (TS) and firefly algorithm (FF). The proposed hybrid bio-inspired system is tested on two well-known Iris and Wine data sets. Finally, the experimental results are compared with the parallel tabu search clustering algorithm. The proposed bio-inspired TS-FF clustering system shows a significantly better accuracy value for Iris data set.