ベイジアンネットワークを利用した個人に合わせた生活習慣改善策提案のための分析

生活習慣病は日々の生活習慣により発症する疾患である．人には改善できる習慣と妥協できない習慣があり，個人に合わせた改善策を提案する必要があると考えられる．しかし，超高齢社会で今後患者数が増加し，手間や時間などのコストがかかることや，人手による多様な提案が困難であることが問題としてあげられる．そこで，本論文では，生活習慣病を対象に，個人に合わせた生活習慣の改善策を提案するための分析を行った．ベイジアンネットワークを用いた研究は，従来「疾患に関する因果関係の分析」と「疾患のための予測・診断・治療」が主な目的であるが，本研究は従来とは異なり，個人に合わせた改善策の提案による「行動変容」を目的に分析を行っている．分析結果として，様々な生活習慣における生活習慣病の確率をシミュレーションのようにして算出可能であった．また，個人の好みや取り組みやすさを考慮して，個人に合わせた生活習慣改善策の提案可能性を示した．

Lifestyle diseases are illnesses that are caused by daily lifestyle habits. It is believed that individuals have habits that can be improved and habits that cannot be compromised, and therefore, it is necessary to propose improvement measures tailored to each individual. However, as the aging population continues to increase in the future, the issue of the increasing number of patients and the costs associated with labor and time, as well as the difficulty of making diverse proposals by human hands, is raised as a problem. Therefore, in this paper, we conducted an analysis of lifestyle diseases with the aim of suggesting remedial measures tailored to the individual. While the use of Bayesian networks has traditionally focused on “causal relationship analysis of diseases” and “prediction, diagnosis, and treatment for diseases”, this study aimed to analyze to achieve “behavior change” through proposing personalized improvement measures, which is different from traditional studies. As a result of the analysis, it was possible to calculate the probability of lifestyle diseases in various daily habits, similar to simulation. Moreover, we showed the possibility of proposing personalized lifestyle improvement measures by considering an individual's preferences and ease of implementation.