An application of General Maximum Entropy to Utility

Methodologies related with information theory have been increasingly used in studies in economics and management. In this paper we use Generalized Maximum Entropy as an alternative to the Ordinary Least Squares in the estimation of utility functions. We estimated linear, logarithmic and power utility functions as well as confidence intervals and in order to compare both methodologies. Results point to the greater accuracy of Generalized Maximum Entropy. Linear function seems to be the one with best goodness of fit.