Abstract
The authors propose a system to analyze the effect of sales promotions by neural networks and fuzzy inference. We firstly make the neural network learn the dynamics between promotional activities and sales quantities based on retail store level scanner data, and use it for the modeling of consumer's purchase behavior. Secondly, we analyze some effects of sales promotions to sales quantities and sales profits by using the neural network after learning as the "knowledge" of consumer's purchase behavior. Moreover, a fuzzy inference is used to evaluate the promotional effective balance between sales quantities and sales profits for each sales promotion. Finally, we apply the system to the instant coffee market as a typical example, and discuss the promotional balance and some effects for various promotions.
