都市の脱炭素化や地域交通の持続可能性が求められる中、電動自動車やオンデマンド交通・カーシェアといった新たなモビリティサービスの導入が進んでいる。しかし、それらに対する利用者の受容性は都市文脈や制度条件により大きく異なる。受容性の計測手法として、統計分析は利用者意識の把握にとどまり、機械学習手法は高精度ながら要因解釈や政策応用に限界があった。本研究は、行動理論に基づく因果推論と説明可能AIを融合し、個人レベルの意思決定構造を可視化しつつ政策シナリオの社会的効果を評価する新たな方法論を構築する。

As cities pursue decarbonization and seek more sustainable regional transportation systems, new mobility services such as electric vehicles, on-demand transportation, and car sharing are increasingly being introduced. However, users’ acceptance of these services varies significantly depending on urban contexts and institutional conditions.
Existing approaches to measuring social acceptance have limitations: statistical analyses are often limited to describing user attitudes, while machine learning methods, despite their high predictive accuracy, often lack interpretability and policy relevance.
This research aims to develop a new methodological framework that integrates causal inference based on behavioral theory with explainable artificial intelligence (XAI). By visualizing individual-level decision-making structures, the proposed approach evaluates the social impacts of alternative policy scenarios and provides insights for designing effective mobility policies.