Comparison of SAS/STAT procedures and variable neighborhood search based clustering applied on Telekom Serbia data

Abstract

In this study we analyze the possibility of improving decision qualities in the largest Serbian telecommunication provider, Telecom Serbia a.d. We compare SAS/STAT clustering with Variable Neighborhood Search based methods on three months of real data sets of prepaid customers. It appears that Variable Neighborhood Search based approach provides results with smaller error, giving more precise decision support to marketing department.