Enhanced Solar Radiation Prediction with Machine Learning: A Comprehensive Analysis of Meteorological Data Using Random Forest

Solar radiation is a crucial source of energy from the sun, with time-varying data that impacts various sectors. The variability of solar generation poses challenges for integration and reliability of solar panels, necessitating accurate forecasting to enhance performance. Machine learning techniques, particularly Random Forest models, are commonly used for solar radiation forecasting. However, there is a need to analyze the characteristics of solar radiation to identify relevant features that have a meaningful correlation between the inputs and outputs of machine learning algorithms. To evaluate the model, we used a comprehensive dataset containing over 149,015 instances of meteorological data. The study highlighted the significance of feature engineering in solar radiation forecasting. The system can forecast radiation values for any day using different meteorological data from the previous data. We compared the model’s performance to other models such as Decision Tree, Linear Regression, and eXtreme Gradient Boosting. An elaborate evaluation of all the models was done to produce a comparison using widely used performance metrics. The results showed that the RF model performed better than the other models in terms of the lowest prediction errors across MSE, RMSE, and MAE, as well as the highest R² value and the lowest MAPE, indicating a good performance in solar radiation prediction. Overall, the RF machine-learning model is an effective method for forecasting radiation in Bangladesh. They can attain high accuracy and provide significant information for the assessment of solar energy resources.