AN ANALYSIS AND ADAPTIVE PREDICTION OF CONSUMER ATTRITION RATE USING FUZZY COGNITIVE MAP (CARM)

Abstract

Consumer retention is a major challenge faced in today’s day to day business. Identifying approaches to predict consumer attrition or retention at an early rate is a major research work which is demandable among industry members and survey shows that business intelligence is always challenging research. This work CARM adopts consistent set of consumer data over varying time period over metrics such as accuracy of prediction and Consumer Life Time (CLV) to analyze on reasons behind attrition rate. CARM uses Fuzzy Cognitive Map as a modelling tool to determine on prediction of attrition over time period. Proposed approach is compared with traditional approaches such as Genetic algorithm, Fuzzy K-means and ANN whose performance shows that CARM shows an improved prediction accuracy of attrition rate (%) and at an early time (msecs). FCM is well adaptable to prediction compared to traditional approaches due to its early susceptibility to optimality condition.