Comparison and evaluation of empirical and machine learning models in estimating global solar radiation in Limpopo province

Abstract

This study investigated the performance of machine learning techniques as compared to the empirical models to forecast the global solar radiation in Limpopo regions. The machine learning techniques used in this study are Support Vector Machines, Random Forest, and Artificial Neural Network, and the empirical models used are the Clemence and Hargreaves- Samani models. To assess the efficiences of the machine learning models against the empirical models, the researchers calculated and compared the models performance evaluation using statistical equations such as Coefficient of determination, Mean Square Error, Mean Absolute Error, and Root Mean Square Error. Calibaration was done to improve performance of the empirical models. The present study found that machine learning techniques perform better than the empirical models when estimating the global solar radiation in the selected Limpopo regions.