Market Segmentation Through Finite Mixture Regression Models with Generalized Normal Distribution and Hierarchical Clustering

Abstract

Market segmentation through mixture regression models  received lot of impetus  to its ready applicability in market analytics, business analytics, financial analytics, supply chain analytics , Human Resource analytics etc. In regression analysis it is common to assume that error term follows a normal distribution. Normal distribution has several drawbacks such as being mesokurtic and the model may not fit well for all types of data. Hence, in this paper we develop market segmentation method though mixture of regression models with Generalized Normal Distributed (GND) errors. The GND includes leptokurtic, platykurtic and normal distribution as special cases. The parameters of the proposed model are estimated using Expectation Maximization (EM) algorithm. The initialization of the model parameters is done by using hierarchical clustering algorithm. The segmentation algorithm is obtained through component maximum likelihood under Bayesian framework. The applicability of the proposed algorithm is demonstrated with market segmentation data. The performance of the algorithm is evaluated by computing   segmentation performance metrics. It is observed that this method performs much better than the earlier segmentation methods having normal distributed   and generalized normal distributed errors with k-means algorithm for the data sets having leptokurtic and platykurtic response variables.