@article{oai:nagoya.repo.nii.ac.jp:00013297, author = {宮崎, 慧 and MIYAZAKI, Kei}, issue = {1}, journal = {経済科学}, month = {Jun}, note = {In recent times, the development of information technology has resulted in the easy availability of disaggregated data. In particular, active researches have been conducted on the analysis methods of scanner panel data in which purchase history of each household is recorded. In this paper, I review representative studies regarding the analysis of consumer purchase behavior and competitive market structure analysis that use scanner panel data in relation to the analysis of purchase data of substitute or complementary products. When dealing with the purchase behaviors of several categories or brands of products, it is important to research the interrelated influences among them. Granger causality is one of the useful methods for understanding the time series variation of purchase behaviors, but in marketing science, few researches directly deal with Granger causality. Thus, I developed a model that could determine the Granger causality among purchase behaviors of each product for each consumer segment. This model enables to determine the latent switching behavior of the consumer across multiple products or brands in the absence of promotional activities. I applied the proposed method to toothbrush and toothpaste data from the IRI market data (Bronnenberg, Kruger and Mela, 2008), and the results were noteworthy.}, pages = {43--56}, title = {非集計データを用いた複数商品の競合・補完関係の分析:日用品の補完財への応用}, volume = {59}, year = {2011} }