<![CDATA[This research evaluates the performance of a hybrid clustering model that integrates K-Means and DBSCAN, enhanced through Feature Weight Optimization (FWO) using a Genetic Algorithm (GA), to achieve more precise consumer data segmentation. Two benchmark datasets, Customer Personality Analysis (CPA) and Online Retail (OR), were utilized to examine how different clustering techniques respond to variations in data structure. The feature weighting process was optimized using GA to improve the representational contribution of each variable toward the final cluster configuration. The Silhouette Score was adopted as the primary evaluation metric to measure intra-cluster cohesion and inter-cluster separation. Experimental findings reveal that for the CPA dataset, the Hybrid + FWO method achieved the best performance with a Silhouette Score of 0.9600, while the K-Means + FWO method recorded the highest score of 0.9804 on the OR dataset. Across all scenarios, the inclusion of FWO consistently enhanced clustering stability and interpretability. These results highlight that algorithm selection must consider dataset characteristics, and that feature weight optimization is pivotal in strengthening segmentation quality and ensuring more meaningful insights in consumer behavior analytics.]]>
Copyrights © 2026
