CLUSTERING ANALYSIS FOR CREDIT CARD USER SEGMENTATION USING K-MEANS ALGORITHM AND PRINCIPAL COMPONENT ANALYSIS

Abstract

Customer segmentation is a process used by companies to group customers based on common characteristics. The goal is to understand customer needs and preferences better so that companies can provide products and services that match customer needs. One way to segment customers is to use clustering algorithms, such as k-means. This algorithm groups data into adjacent clusters with randomly selected centroids. In the case of credit card customer segmentation, the k-means algorithm can be used to group customers based on characteristics such as number of transactions, amount of payments, and credit history. Thus, companies can better understand the needs and preferences of credit card customers and determine more effective marketing strategies. The advantages of the k-means algorithm and the clustering method are that the developed models can help companies determine more effective marketing strategies, easy-to-use algorithms with fast computation time and accurate results, and the PCA algorithm is also used to reduce dimensions and makes data visualization easier. Based on the test results and analysis of credit card customer data, the performance of the k-means algorithm is considered relatively good for segmentation with the number of clusters = 3 and the Davies Bouldin value = -0.778.