A Novel Multi-Objective Genetic Algorithm for Clustering

• Published in: Intelligent Data Engineering and Automated Learning - IDEA 2011 Vol. 6936; pp. 317 - 326
• Main Authors: Kirkland, Oliver, Rayward-Smith, Victor J., de la Iglesia, Beatriz
• Format: Book Chapter
• Language: English
• Published: Germany Springer Berlin / Heidelberg 2011; Springer Berlin Heidelberg
• Series: Lecture Notes in Computer Science
• Subjects: Artificial intelligence; Information retrieval
• ISBN: 9783642238772; 3642238777
• ISSN: 0302-9743; 1611-3349
• DOI: 10.1007/978-3-642-23878-9_38

In this paper, we introduce a new Multi-Objective Clustering algorithm (MOCA). The use of Multi-Objective optimisation in clustering is desirable because it permits the incorporation of different criteria for cluster quality. Since the criteria to establish what constitutes a good clustering is far from clear, it is beneficial to develop algorithms that allow for multiple criteria to be accommodated. The algorithm proposes a new implementation of multi-objective clustering by using a centroid based technique. We explain the implementation details and perform experimental work to establish its worth. We construct a robust experimental set up with a large number of synthetic databases, each with a pre-defined optimal clustering solution. We measure the success of the new MOCA by investigating how often it is capable of finding the optimal solution. We compare MOCA with k-means and find some promising results. MOCA can generate a pool of clustering solutions that is more likely to contain the optimal clustering solution than the pool of solutions generated by k-means.