Response modeling by Gene Expression Programming method in single channel and multi channel marketing campaigns

Previously in an internship with MITACS, researcher found a piecewise linear model to approximate the dependent variable. The genetic algorithm was employed to approximate the value of sales, as the dependent variable against changes on marketing efforts, as the independent variables. The outcome became a program that can approximate the actual data with three lines. Suggested program was able to calculate the coordinates of four nodes as well as three slopes and draw the 3-pice linear approximation.

The aim of this research is to apply gene expression method to create a predictive modeling and approximation program. This program models the response curves against marketing efforts in the cases of single or multiple marketing channels (independent variables). Due to the novelty of gene expression programming applications, it is anticipated that the output of this research can be presented as one of the first applications of this fantastic scientific method.

Faculty Supervisor:

Amir Khajepour

Student:

Partner:

Discipline:

Computer science

Sector:

University:

University of Waterloo

Program:

Accelerate