UVBoost: An erythemal weighted ultraviolet radiation estimator based on a machine learning gradient boosting algorithm. (April 2023)

• Record Type: Journal Article
• Authors: Corrêa, Marcelo de Paula
• Is Part Of: Journal of quantitative spectroscopy & radiative transfer. Volume 298(2023)
• Publication Date: 2023-04
• Subjects: Ultraviolet index -- Artificial intelligence -- Skin cancer -- vitamin D -- photoprotection; Spectrum analysis -- Periodicals; Radiation -- Periodicals; Analyse spectrale -- Périodiques; Rayonnement -- Périodiques; Radiation; Spectrum analysis; Periodicals; 543.0858
• DOI: 10.1016/j.jqsrt.2023.108490
• Languages: English
• ISSNs: 0022-4073

Abstract

Highlights: UVBoost uses the Catboost machine learning tool to manage categorical data. UVBoost calculations were accurate with differences up to <5% relative to the RTM. UVBoost is 8.7 × 10 3 times faster than the conventional RTM. UVBoost is open-source and uses the Python coding language. UVBoost is a useful tool that can be widely applied by researchers in any field of knowledge. Abstract: This article presents UVBoost, an ultraviolet radiation (UVR) estimator based on a Supervised Machine Learning (SML) regression model powered by high precision calculations provided by a conventional Radiative Transfer Model (RTM). The proposed regression model takes UVR as a dependent variable, and the Solar Zenith Angle (SZA), Total Ozone Content (TOC), and Aerosol Optical Depth (AOD), as the independent predictive variables. UVBoost was developed to increase computational speed for conducting calculations with large databases, without sacrificing accuracy. Furthermore, this method employs a user-friendly code, which can be used by laymen or researchers in other areas. UVBoost can be employed to disseminate UVR data in different spatiotemporal scales, or for climatological projection studies on a global scale. UVBoost was developed by comparing seven regression SML tools via cross validation. Results were validated using non-parametric statistical tests. Of all the tested tools, the Categorical Boosting (CatBoost) method showed the best accuracy at the lowest computational cost. Two additional studies were carried out, one at the global scale, and another at the local scale, to compare the traditional RTM vs. the UVBoost results. The first study simulated a global UVR field (1°x1°) with 64, 800 gridpoints. The differences between RTM and UVBoost were less than ±5% for approximately 95% of all points, except for high SZA datapoints. The computational speed of UVBoost surpassed that of the RTM by more than three orders of magnitude. The second study simulated the daily UVR at eight different locations on Earth. The results showed that the UVBoost was very efficient in simulating accumulated UVR doses during the day, with negligible differences (< ±3%), which means it can be used in studies on UVR and human health. In the future, UVBoost will include other geophysical parameters and be extended to other bands in the electromagnetic spectrum.