Data-driven Clustering analysis in E-commerce User Segmentation

Abstract

E-commerce user segmentation is crucial for enterprises aiming to develop precise marketing strategies and efficiently manage their customer base. Traditional clustering algorithms often face challenges due to noisy data and isolated points, which can degrade segmentation quality. This investigation focuses on improving e-commerce user segmentation through innovative cluster analysis techniques that enhance traditional clustering methods by effectively excluding noisy and isolated data points and selecting high-quality initial cluster centers. A novel Efficient Coral Reefs Optimized Kernel Density-Based Spatial Clustering of Applications with Noise (ECRO-Kernel DBSCAN) algorithm is useful for user segmentation based on the customer behavior on e-commerce platforms. The dataset includes customer transaction and interaction data such as purchase history, transaction frequency, monetary value, browsing behavior, and other engagement metrics. First, Kernel DBSCAN is used to detect core dense regions and remove noise; then, ECRO is applied to the core points to identify clusters. The analysis reveals five distinct customer groups: Platinum, Gold, Silver, Copper, and Iron segments asked their purchasing behavior and engagement levels. A detailed estimation of these segments informs the design of customized marketing strategies tailored to each group's unique characteristics. The effectiveness of this targeted marketing approach is further assessed through enhancements in sales performance and platform user satisfaction. Experimental outcomes demonstrate 94% accuracy, 93% precision, 92% recall, and 94% F1-score, highlighting superior segmentation robustness and practical effectiveness in targeted marketing and customer engagement compared to traditional clustering methods. This research provides valuable insights into e-commerce user segmentation, offering a practical framework for businesses to optimize customer targeting, increase engagement, and drive sustainable growth through data-driven cluster analysis.