No. 1112

Title:

Measurements of Effects of Serious Incidents Based on Daily POS Data

Author(s):

Fumiyo N.Kondo (Universit of Tsukuba);
Fumiaki Kubota (Universit of Tsukuba);
Takanori Maegawa (Universit of Tsukuba)

Key words:

definition of critical incidents; keywords of critical incidents; digital new data; text mining; POS dataGintervention analysisGBS ARIMA

Abstract:

The purpose of this study is to measure the effects of the critical event news on consumersf behavior that can be reflected on POS data. The records of consumersf behaviors can be considered to be the response of the sales of products related to critical incidents. Intervention analyses by ARIMA model are conducted by using POS data for the analyses on the relationships between the news and the customer purchasing behavior. The effects of BSE problem news are confirmed to be statistically significant. Further, in order to model regional differences, Bayesian Sampling-based ARIMA model based on hierarchical Bayesian techniques was constructed and model comparison between POS data analyses are conducted.