Multi-objective C-means Data Clustering Algorithm using Self-Adaptive Differential Evolution

: This paper proposes a Multi-objective C-means Data Clustering algorithm using Self-Adaptive Differential Evolution (DE) for improving the performance of data clustering by introducing three data clustering validity indices.. The proposed algorithm composed of three objectives: including the symmetry-index to maximize similarity within clusters, the compactness index to maximize dissimilarity among clusters, and validity Silhouette index to improve the validity of data clustering. Selfadaptive DE is similar to the traditional DE algorithm except two changes in the mutation and the crossover operations [19], where DE is a global optimization technique [13]. The proposed algorithm is implemented and evaluated using twenty benchmark data sets and compared with different 5 data clustering algorithms that MOSAC-Means, GenClustMOO, MOCK, VGAPS, and GenClustPESA2. The experimental results showed that the proposed algorithm is performing well compared with the previous algorithms.