Publication

Abstract : Accurate forecasting of renewable energy is essential for efficient grid management and sustainability. This study compares machine learning models—Lasso Regression, Ridge Regression, Decision Tree, Random Forest, Artificial Neural Networks (ANN), and Long Short-Term Memory (LSTM)—using time-series data from the Open Power Systems database in Germany. Results show that ANN and LSTM achieved the best performance, with R2 values of 0.9967 and 0.9966, and the lowest error metrics, outperforming traditional models. The project also delivers a cloud-based web interface where users input parameters such as humidity, air pressure, and cloud cover to instantly predict renewable energy output. The platform enhances usability and accessibility, contributing to reliable and sustainable energy forecasting.