A combination of forensic-based investigation algorithm and density peak-based fuzzy clustering for custom segmentation

Custom segmentation is a process of classifying potential customers based on their mutual features such as shopping habits, consumption trends, and demand to provide an effective marketing campaign for each customer group. Data clustering is one of the most common methods for custom segmentation. This study proposed a novel clustering method that employs density peak-based fuzzy c-means (DP-FCM) and forensic-based investigation (FBI) algorithms. The proposed method (denoted as DP-FBI-FCM) aims to provide an effective clustering technique that can exploit the global optimal solution for custom segmentation problems. Besides, the proposed DP-FBI-FCM is used to segment wholesale customer data of a supermarket. As a result, four distinct customer groups are classified. Businesses can implement different strategies in each cluster to retain and attract their customers.