RECOVERY OF THE CONSUMER MULTIATTRIBUTIVE UTILITY MAXIMIZATION PROBLEM

Abstract

This article contains a prediction model for the demand for the consumer commodities. I consider the classical
model of the consumer utility function maximization in a given budget constraint where there are two products:
the first one for which the demand is being estimated, and the rest of the consumption bundle which is the second
product. The utility function is introduced as multiattributive utility function with an unidentified number of
attributes. An approach to estimate the exact number of attributes and the parameters of the model in a given class
of utility functions for each attribute was proposed. The estimation is derived through the optimization of
corrected Akaike information criterion, where the parameters of the utility function are continuous and the number
of attributes is integer and positive. This model was tested on the prediction of the homogenous product demand
with the Giffen effect.