Evolutionary algorithms for robust data clustering: a comprehensive literature review

Abstract

Clustering is a fundamental technique in data mining and unsupervised learning that enables the identification of hidden structures in unlabeled datasets. However, traditional clustering algorithms often encounter limitations when applied to noisy or complex datasets. Evolutionary algorithms have emerged as effective alternatives for addressing optimization challenges in clustering tasks. This study presents a comprehensive literature review examining clustering algorithms and evolutionary optimization methods applied to clustering problems. The analysis focuses on classical clustering techniques, fuzzy clustering models, and differential evolution algorithms. The reviewed studies indicate that evolutionary approaches improve clustering robustness by reducing sensitivity to initialization parameters and local optima. Hybrid models combining evolutionary algorithms with conventional clustering methods demonstrate improved performance in complex data environments. The results suggest that evolutionary optimization represents a promising direction for improving clustering accuracy in data mining applications.