Abstract

Customer clustering is used to understand customers’ preferences and behaviors by examining the differences and similarities between customers. Kohonen vector quantization clustering technology is used in this research and is compared with Kmeans clustering. The data set consists of customer records obtained from a mobile telecommunications service provider. The customers are clustered using various attributes that are broadly grouped under usage, revenue, handset, and service. The clustering results are examined to see the relationships between different types of attributes. The analysis leads to the discovery of several interesting facts about customers that may be of use to mobile service providers.

Share

COinS