Integration of K-Means and SVM for Customer Segmentation Based on Customer Behaviour Using LRFM and Product Preference

In the era of Big Data, utilizing customer transaction data in the retail industry across various sectors has become crucial for understanding customer behavior and optimizing marketing strategies. This study aims to segment customers using the integration of K-Means clustering and Support Vector Machine (SVM) models based on LRFM (Length, Recency, Frequency, Monetary) and product preferences. A total of 365,736 transactions involving 24,950 customers were collected and processed, resulting in 8 customer clusters with distinct characteristics. The 8 clusters were determined based on the elbow method, a silhouette score of 0.323, and decisions made by stakeholders. Following the clustering results, SVM classification was performed, demonstrating an accuracy of 99.52%, which indicates the model's ability to effectively identify customer segments. It can be concluded that the combination of K-Means and SVM proves effective in generating relevant and actionable customer segmentation. This method allows companies to identify groups of customers based on their purchasing behavior and provides deep insights into customer preferences. Furthermore, this approach can be applied periodically, enabling segmentation to be updated in accordance with changes in consumer behavior. With the capability for automation, this process not only reduces manual workload but also enhances operational efficiency. Overall, the combination of K-Means and SVM supports the development of more focused marketing strategies and improves strategic decision-making in loyalty programs.