Scholarlink

<p>Urban mobility needs to shift towards sustainability to address the environmental, social, and economic problems, challenges, and implications arising from the increasing use of private vehicles in rapidly emerging countries like India. The proposed research will use a hybrid modeling framework grounded on Principal Component Analysis (PCA) and explainable ensemble machine learning to test the willingness to adopt sustainable transport modes in two large Indian Metropolitan regions. A perception-based survey of 780 regular private vehicle commuters was conducted, and 21 behavioral indicators were recorded using a structured questionnaire survey. The PCA identified four latent constructs: Social Norms and Environmental Responsibility; Trust in Public Transport and Perceived Effectiveness; Health motivation and Policy support; Environmental Awareness and Preference for Sustainable Transport. These components, along with the original variables, have been used to train different Machine learning classifiers with an 80/20 train-test split using fivefold cross-validation. Ensemble models showed high predictive accuracy (RF: 93%; XGBoost: 89%), and the service accessibility, habitual resistance and EV-readiness were identified as dominant predictors of adjustment using SHapley Additive exPlanations. Model-based sensitivity analysis of interaction effects as a policy lever is expressed as reduced factorial simulations, which are operationalized as binary baseline conditions and hypothetical intervention conditions. The findings indicate that combined structural and behavioral interventions have a greater predictive value for adoption than single measures. This study would be useful for evidence-based policymaking, enabling the shift to sustainable mobility.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Modeling Behavioral Determinants of Commuters’ Transition to Sustainable Mobility Using an Explainable Machine Learning Approach in Indian Cities

  • 期刊论文

  • 发表日期：

    2026年4月25日

查看全文
  • Rupam Sam,
  • Sudip Kumar Roy,
  • Subojit Debnath

摘要

Urban mobility needs to shift towards sustainability to address the environmental, social, and economic problems, challenges, and implications arising from the increasing use of private vehicles in rapidly emerging countries like India. The proposed research will use a hybrid modeling framework grounded on Principal Component Analysis (PCA) and explainable ensemble machine learning to test the willingness to adopt sustainable transport modes in two large Indian Metropolitan regions. A perception-based survey of 780 regular private vehicle commuters was conducted, and 21 behavioral indicators were recorded using a structured questionnaire survey. The PCA identified four latent constructs: Social Norms and Environmental Responsibility; Trust in Public Transport and Perceived Effectiveness; Health motivation and Policy support; Environmental Awareness and Preference for Sustainable Transport. These components, along with the original variables, have been used to train different Machine learning classifiers with an 80/20 train-test split using fivefold cross-validation. Ensemble models showed high predictive accuracy (RF: 93%; XGBoost: 89%), and the service accessibility, habitual resistance and EV-readiness were identified as dominant predictors of adjustment using SHapley Additive exPlanations. Model-based sensitivity analysis of interaction effects as a policy lever is expressed as reduced factorial simulations, which are operationalized as binary baseline conditions and hypothetical intervention conditions. The findings indicate that combined structural and behavioral interventions have a greater predictive value for adoption than single measures. This study would be useful for evidence-based policymaking, enabling the shift to sustainable mobility.