Automating consumer segmentation in e-commerce using clustering and the NSGA-II algorithm

The rapid expansion of e-commerce has created a pressing need for intelligent, automated methods of customer segmentation capable of capturing complex behavioral patterns and supporting large-scale personalization. Traditional segmentation techniques—such as demographic grouping, rule-based approaches, and basic clustering—struggle to handle multi-criteria trade-offs involving compactness, separability, and marketing value. To address these limitations, this study proposes a hybrid decision-support framework integrating the NSGA-II multi-objective evolutionary algorithm with clustering techniques to automate consumer segmentation. Three optimization objectives are incorporated: intra-cluster compactness, inter-cluster separation, and predicted marketing value estimated using gradient-boosted machine learning models. Experimental results show the formation of three distinct Pareto-optimal segmentation structures, reflecting trade-offs between structural quality and business value. The findings confirm the effectiveness of the proposed hybrid approach for supporting automated, scalable personalization strategies in modern e-commerce ecosystems.