Proposal of a Model Verification Method by Virtual Questionnaire Survey to Understand Consumer's Purchasing Decision Process.

Abstract

In order to understand consumers' purchasing decision process, two research approaches are well known. One is analytic approach that statistically analyze questionnaire surveys to derive it, and the other is constitutive approach that verify the assumption for it by dynamic simulation such as MAS (multi-agent simulation). Conventional constitutive researches typically verify the assumption model by comparing simulated trend of sales volume with actual trend as observable metrics. In this article, the authors propose a model verification method, which can compare simulator agents' decision process with actual questionnaire survey. Then, the proposed method has been implemented on generic MAS simulator, by adding two features that are logging agents' decisions and answering to questionnaire. Finally, pilot trials of this method are discussed.