An unsupervised data mining approach for clustering customers of abrasive manufacturer

Abstract

Customer segmentation is the process of dividing customers into groups based on common similar characteristics such as value, location, demography etc. Companies can communicate with each group effectively and appropriately by considering these common properties. Data mining algorithms are the most utilized techniques which lead direct marketers to develop their marketing strategies tailored to particular segments and/or individuals. Clustering is one of the unsupervised data mining methods used for grouping set of objects such a way that objects in the same group have maximum similarity while between group similarities are low. K-means clustering is a commonly used non-hierarchical clustering method for performing non-parametrical learning tasks. This study aims to identify customer types according to their profitability, value and risk in order to take appropriate action for each group via clustering. In this study, data items are grouped according to coded customer profile with respect to the consumers’ total expenditures. Customers are segmented as VIP, Platinum, Gold, and Bronze into 4 groups according to their values within 2 years.