Self Optimal Clustering Technique Based on Multi-Objective Genetic Algorithm

Abstract: The self optimal clustering technique is new area of research in data mining. The self optimal clustering technique increases the efficiency and scalability of partition clustering and mountain clustering technique. The concept of self optimal clustering technique used the concept of heuristic function for the selection of cluster index and centre point. In this paper proposed a novel self optimal clustering technique using multi-objective genetic algorithm. The multi-objective genetic algorithm work in two phases in first phase the genetic algorithm work for the selection of center point and merging the cluster index value based on defined fitness constraint value. In second phase of genetic algorithm check the assigned number of value of K for the process of clustering and validated the clustering according to the data sample. The proposed algorithm implemented in MATLAB software and used some reputed dataset form UCI machine learning repository.

Keywords: Data Mining, Clustering, Heuristic Function, MOGA.