SolarPred: Hybrid Machine Learning Based Time Series Prediction Model for Solar Energy Applications

The excessive use of fossil fuels has a detrimen-tal effect on the environment, leading to a global increase in the need for energy harvesting technologies from diverse renewable energy sources. Solar energy is a widely recognized and extensively utilized renewable energy source for generating power in various applications. The performance of Photovoltaic (PV) panels is significantly impacted by environmental variables. This led to fluctuating power output, raising concerns about energy security and iddentified as a significant research interest. Therefore, it has become crucial to have a precise framework for predicting solar power output. This article presents an advanced hybrid Machine Learning (ML) model that utilizes time series data to accurately forecast the output power of a solar system. Publicly available dataset has been used for training and testing of the proposed model. The proposed work surpasses the benchmark ML models by exhibiting coefficient of determination (R2) as high as 0.828 and Mean Absolute Error (MAE) as low as 0.231.