2013 年 12 巻 4 号 p. 455-463
The metabolic syndrome has become a significant problem in the world, and health checkup and guidance targeted on it has started from 2008 in Japan. In order to support this large undertaking from the side of information technology, we introduce ideas based on Bayesian networks in data mining technology. The Bayesian network has emerged in recent years as a powerful technique for handling uncertainty in complex domains and is expected to be an appropriate method in the health checkup domain that requires analyzing the results with medical knowledge. In this work, we constructed a Bayesian network scheme connecting the information of physical examinations and questionnaire on daily life style. By applying this network model to the field data of 12,230 subjects, the proposed method was found to provide better performance and show its potentiality for the system of specific health checkup.