Genetic modelling in database marketing: The GenIQ Model

Abstract

Using a variety of techniques, data analysts in database marketing aim to build models that maximise expected response and profit from solicitations. Standard techniques include the statistical methods of classical discriminant analysis, as well as logistic and ordinary regression. A recent addition to the data analysis arsenal is the machine learning (ML) method of neural networks. The GenIQ model is a hybrid ML-statistics method that is presented in full detail in this paper. First, a background on the concept of optimisation will be helpful, since optimisation techniques provide the estimation of all models. Genetic modelling is the �engine� for the GenIQ model, and is discussed next as an ML optimisation approach. Since the objectives of database marketing are to maximise expected response and profit from solicitations, the author will demonstrate how the GenIQ model serves to meet those objectives. Actual case studies will further explicate the potential of the GenIQ model. [ABST