Analysing Wine Demand With Artificial Neural Networks

In this paper we analyse wine demand in Italy using microdata. Instead of estimating a traditional parametric model (like AIDS) we employed artificial neural networks (ANN) and evaluate the elasticities using two different methods, specific for the non parametric framework. We compared the performances of the two methods to estimate elasticities and put in evidence the relevance of some demographic variables together with the usual economic ones, explaining the consumer's behaviour.