Differential evolution for multi-objective clustering

This work describes a Differential Evolution (DE) for multi-objective clustering. Compared to traditional clustering algorithms, we evaluate two objectives--the compactness in clusters and the looseness between different clusters. DE is a population based search algorithm, which requires few control variables and it is robust, easy to implement. To optimize the two measures simultaneously, a weighing factor method is used for the clustering problem. Experiments on several data sets from UCI machine learning repository show that this new clustering method can achieve good results.