Model for Estimating Ultraviolet Index Values Using Artificial Intelligence Techniques

Huaman C.W.T.; Vilca R.A.L.; Mamani E.G.M.; Calderon J.T.; Nina N.O.S. · Smart Innovation, Systems and Technologies (2026) · DOI: 10.1007/978-3-032-24262-4_23

Resumen

Accurate real-time estimation of the ultraviolet index (UVI) is crucial for public health, especially in areas with high solar radiation like Puno, Peru. Traditional methods, such as those provided by SENAMHI, rely on daily averages and broad geographic areas, failing to accurately reflect local and temporal UVI variations. To overcome this limitation, the present study developed a predictive model for maximum UVI using machine learning techniques, specifically Linear Regression and Random Forest. The models were trained on meteorological variables recorded by a DAVIS Vantage PRO 2.0 station between 2017 and 2024. Following data cleaning, exploratory analysis, and comparative modeling, it was determined that solar radiation and solar energy were the variables with the greatest influence on UVI prediction. While the Linear Regression model achieved a coefficient of determination of R2 = 0.87, the Random Forest model demonstrated superior performance (R2 = 0.95; MAE = 0.75), highlighting its ability to capture non-linear relationships. In conclusion, decision tree-based models like Random Forest are robust tools for localized, real-time UVI estimation, with relevant applications in public health, risk management, and outdoor activity planning. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2026.

Instituciones: Universidad Nacional del Altiplano; Universidad Nacional de Juliaca; Universidad Católica de Santa María