Performance evaluation of machine learning algorithms for estimating reference evapotranspiration based on NASA POWER weather data: a case study in Nigeria

Abstract

) was good, with values of 0.87 and 0.72 during training and validation. The FG SVM outperformed all other models across all study locations, demonstrating its robustness in predicting ETo despite the contrasting weather conditions. Overall, this study revealed that the integration of data from NASA POWER with FG SVM accurately estimated reference evapotranspiration, which is important for effective water resource management in areas where ground climatic data is unavailable.