Predicting Market Responses with a Neural Network: the Case of Fast Moving Consumer Goods

Market response modeling is well covered in the marketing literature. However, much less research has been undertaken in the use of neural networks for market response modeling. In this paper experiments are described to t neural networks to a data set taken from a fast moving consumer goods market. The neural network approach is compared with several other possible models. We focus on the out of sample performance of the models. A method for adjusting the neural network architecture is described that leads to better performance on out of sample data. Sample Performance.