An interactive web-based solar energy prediction system using machine learning techniques
Priyanka Chawla, Jerry Gao, Teng Gao, Chengchen Luo, Huimin Li, Yiqin We, Yiqin We
Abstract
Solar energy being one of the most inexpensive renewable energy sources, has shown to be a viable alternative to traditional fossil-fuel and wood-based electricity generation. For the purpose of creating a more trustworthy and successful energy planning strategy, accurate projections of sun irradiation, solar energy generation, and revenues are crucial. Hence, in this work we have proposed web-based optimal prediction system that estimates solar radiation based on location and meteorological data using Machine Learning techniques. Furthermore, an interactive dashboard solar digital map has been developed that enables real-time investigation of solar energy consumption, production, solar radiation, and investment potential for a specific county in California. The model's performance has been measured using Root Mean Square Error (RMSE), Mean Square Error (MSE), Mean Average Error (MAE), and Mean Absolute Percentage Error (MAPE) scores. Experimental results demonstrate that stacking model outperformed all the models with the lowest RMSE, MSE, and MAE.