Comparison of K-Means, Fuzzy C-Means and Fuzzy Logic Based Clustering Algorithms for Customer Segmentation

Customer segmentation is a critical tool for businesses to optimize their marketing strategies and increase customer loyalty. This study aims to segment customers based on their buying behavior using K-Means, Fuzzy C-Means (FCM) and Fuzzy Logic based clustering algorithms. While the K-Means algorithm divides the data into a certain number of clusters and determines a center for each cluster, the FCM algorithm allows data points to belong to more than one cluster with varying degrees of membership. Fuzzy Logic based clustering provides a more nuanced segmentation by using a system of fuzzy membership functions and rules defined for each feature. In this study, the performance of all three algorithms is evaluated with Silhouette score, Calinski-Harabasz score and Davies-Bouldin score using Online Retail dataset. The results show that K-Means generally delivers higher Silhouette scores and Calinski-Harabasz index values, FCM provides flexibility to manage similarities across clusters, and the Fuzzy Logic-based approach better captures complex relationships between features. The findings provide valuable insights for customer segmentation and optimization of marketing strategies.