Evaluation of rural tourism competitiveness based on multi-spatial scale data and random forest algorithm: a case study of Dujiangyan

Accurately assessing the development status and competitiveness of rural tour-ism is paramount for fostering its high-quality and sustainable growth. This paper constructs a rural tourism competitiveness evaluation system based on multi-source geospatial and socio-economic data, selecting 13 indicators across four dimensions: resource endowment, natural environment, economy and society, and policy support. Utilizing the Random Forest (RF) algorithm to analyze data from the Dujiangyan village area, the results reveal: 1) The RF algorithm overcomes subjective weight allocation issues in traditional meth-ods, enhancing the reliability and objectivity of evaluation results while im-proving the model's predictive capability and feature interpretability; 2) The optimized model achieves a root mean square error (RMSE) of 0.098, indicat-ing high predictive accuracy; 3) The highly competitive category demonstrates significant advantages in resource endowment and service facilities; the mod-erately competitive category exhibits deficiencies in service facilities and the ecological environment; the low competitiveness category displays a favorable ecological environment but weaker resource endowment. This study provides empirical references and policy recommendations to facilitate the sustainable development of rural tourism.