Responsible tourist behaviour in Jashpur: A PLS-SEM and SHAP-based machine learning study of tribal ecotourism

Sustainable ecotourism depends on responsible visitor behaviour, particularly in emerging destinations where regulation, interpretation, and visitor-management systems remain limited. The purpose of this study is to explain and predict responsible tourist behaviour among 425 tourists in Jashpur, an emerging tribal ecotourism destination in central India. Grounded in the Value-Attitude-Behavior (VAB) framework, the study applies Partial Least Squares Structural Equation Modelling (PLS-SEM), mediation analysis, and Importance-Performance Map Analysis (IPMA) to examine relationships among perceived value, tourist satisfaction, ecotourism intention, and responsible tourist behaviour. Machine-learning models and SHAP explainability are then used to assess predictive performance and rank behavioural and demographic predictors. PLS-SEM shows that ecotourism intention is the strongest direct predictor of responsible tourist behaviour, followed by perceived value, while tourist satisfaction has a weaker effect. The intention-mediated pathway is also stronger than the satisfaction-mediated pathway. Ridge Regression performs best, explaining 64.5% of responsible tourist behaviour variance, while SHAP independently identifies ecotourism intention and perceived value as the leading predictors. The findings identify intention-building and value-enhancing interventions as priorities for sustainable tribal ecotourism management.