<![CDATA[Accurate prediction of river water quality is essential for effective environmental monitoring and sustainable water resource management. The increasing availability of open environmental sensor data enables continuous observation of key physicochemical parameters, offering new opportunities for data-driven prediction and evaluation. This study evaluates the performance of multiple predictive models for river water quality using open-access sensor datasets derived from diverse river systems. A comparative framework was applied to assess predictive accuracy, robustness, and generalization across indicators, spatial contexts, and temporal scales. Model performance was examined using standardized statistical metrics and benchmarked against official monitoring reference values reported in prior studies and regulatory reports. The findings indicate that predictive reliability varies across models and water quality parameters, with robustness strongly influenced by data quality, sensor characteristics, and environmental heterogeneity. The results further highlight the importance of transparent performance evaluation for operational monitoring, policy relevance, and decision-support applications. Overall, this study underscores that structured evaluation using open sensor data enhances the credibility, transferability, and practical applicability of river water quality prediction systems. ]]>
Copyrights © 2026
