Comparison of Multivariate Statistical Analysis and Machine Learning Methods in Retailing: Research Framework Proposition
7 Pages Posted: 20 Nov 2018
Date Written: September 8, 2016
The aim of this paper is comparison of multivariate statistical analysis and machine learning methods based on the model used for the measurement of current and forecasting of the future customer profitability. Modern customer profitability analysis shows that customer-company relationship is burdened, beside costs of product, with many other different costs generated by business activities. Such costs generated by logistics, post-sale support, customer administration, sale, marketing etc. are allocated in customer’s base in non-linear way. Allocation can vary significantly from customer to customer, making the reason why each different customer’s monetary unit of revenue does not participate in profit in the same way. The research model uses RFM model to define forecasting variables and neural network, multivariate regression analysis and binary logistic regression as forecasting methods. This paper shows the ways how proposed methods can be used in process of forecasting customer profitability giving comparison of their application in that field.
Keywords: multivariate statistical analysis, RFM, machine learning, customer profitability, forecasting, knowledge
JEL Classification: C45, C53
Suggested Citation: Suggested Citation