Sustainable development of road tourism: Model-based forecast of future trends

Objective: Road tourism plays a crucial role in sustainable regional development, involving intelligent route planning to balance tourist demand, environmental sustainability, and infrastructure capacity. Traditional methods often fail to capture the dynamic nature of visitor preferences, which are influenced by the previous behaviors, traffic congestion, and environmental factors. The research aims to address these limits by forecasting future trends in sustainable road tourism using a superior predictive model. Method: To tackle the challenges, this research proposes a Bidirectional Gated Recurrent Unit fused Dynamic Random Forest (Bi-GRUForest) model. The Bi- model integrates a Bi-GRU for tourist route recommendation and a Dynamic Random Forest (DRF) network to capture travel patterns, with a temporal attention mechanism incorporated to prioritize key travel intentions. Multi-source data, comprising environmental data, road infrastructure, and tourist movement patterns, are gathered and preprocessed utilizing techniques like outlier removal, missing value handling, and normalization to ensure consistency, reliability, and accuracy. Real-time data on road conditions, weather updates are integrated to promote eco-friendly travel choices. Result: Experimental results demonstrate that Bi-GRUForest outperforms existing models in forecasting travel trends, optimizing road network efficiency, and supporting environmentally responsible tourism development. The model achieves a recall of 90.2%, precision of 89.5%, F1-score of 87.9%, and lower error rates with MAE of 600.01, MAPE of 15.10, and RMSE of 700.01. Conclusion: The research provides valuable insights for policymakers, transportation planners, and tourism stakeholders, improving route prediction accuracy, reducing carbon emissions, and alleviating traffic congestion, contributing to the development of more sustainable road tourism and practices.