Recommendation System for Retail Business Using Customer Segmentation: Case Study of Tuenjai Company in Surat Thani, Thailand

This research aimed to develop a recommendation system (RS) for the retail store using customer segmentation based on customer purchase behavior by RFM (Recency, Frequency, and Monetary) analysis and k-means clustering. Since inventory management and promotions are important for increasing sales and attracting customer attention, this system was specifically designed for those purposes. In this RS, customer segmentation was performed automatically by calculating the Silhouette Coefficient (SC), Calinski-Harabasz (CH) index, and Davies-Bouldin (DB) index to obtain the optimal number of customer segments. The recommended products were generated from association rule mining to find the relationships between products that customers previously purchased. This study used the Frequent Pattern Growth (FP-Growth) algorithm to find the association rules. The system was developed to be easy to use, user-friendly, and to display results in a format that users could easily understand. The result can be used for the benefit of retail business.