AI-Driven Customer Segmentation for Enhanced Decision-Making in Retail Markets

People have recently become more cautious about online shopping. This has led to the development of online retail platforms and a deeper knowledge of customer purchase behavior. Retail organizations face the challenge of managing a large volume of customer purchases, requiring efficient methods to create more precise and effective consumer segmentation. Customer segmentation is an effective marketing weapon that supports consumer-centric services and, as a result, increases profitability. In this research, we propose a novel consumer segmentation model based on K-Means Clustering (KMC), Particle Swarm Optimization (PSO), and Central Limit Theorem (CLT) to enhance decision-making processes for companies. To handle large volumes of customer data, the PSO model is combined with the KMC algorithm to detect the optimal centroids in the clusters, which are then further optimized using CLT, resulting in more effective segmentation. For the experiments, we used a UK-based online retail dataset. We compared several state-of-the-art clustering techniques, including K-Means and K-Means++. The model was validated using four metrics: the Silhouette Index, Dunn's Index, Davies-Bouldin Index, and execution time. The proposed K-Means model outperforms existing clustering algorithms, with a Silhouette Index of 0.8124, a Dunn's Index of 0.8045, and a Davies-Bouldin Index of 0.0254.