ANALYSIS AND VISUALIZATION OF MARKET SEGEMENTATION IN BANKING SECTOR USING KMEANS MACHINE LEARNING ALGORITHM

Abstract

Segmentation is a way of assigning each dataset to a segment called cluster. It is widely applied in different area of human endeavor such as banking sector, health sector, retail, media etc. Many organizations are faced with problems of ineffective customer care services and intelligent management decisions because of inability to effectively analyze customer data that will give insight to the nature of customers to help in effective customer services and intelligent management decision.  Kmeans algorithm is the widely used algorithm for market segmentation, normally the k value of Kmeans algorithm are randomly picked. Picking the optimal k value is usually a challenge in application of Kmeans algorithm and this usually affects the performance of Kmeans algorithm. This work applies elbow method to obtain the optimal k value that was applied to analyze dataset from banking sector (in this case United Bank of Africa) for better insight, business management and marketing strategy. The customer cluster created was evaluated using visual plots and cluster centers. The optimal k value of six (6) was obtained using the elbow function. The dataset was thus segmented based on the optimal k value of 6 obtained. The clustering results obtained showed high intra cluster similarity (data within a cluster are similar) and low inter cluster similarity (data from different clusters are dissimilar). The result also showed that customers in cluster 3 and 4 has similar marketing needs and can be served together