Segmentação de Mercado e Modelos Mistura de Regressão para Variáveis Normais

The purpose of this work is to provide an overview of what is perhaps the most common analysis context in market research – that of regression models for normally distributed data. In fact, examples of applications of these models continue to accumulate in the marketing literature, given their relative advantages. Moreover, these models are ease implemented due to its incorporation in many commercial packages of marketing research. We aim at presenting the background for the development of mixture regression models (switching regressions, clusterwise regression and finite mixture models) and review the formulation of the basic model and its main extensions in the context of panel data analysis and conjoint studies.