APPLYING MACHINE LEARNING TECHNIQUES TO ENERGY OPTIMIZATION USING LSTM, XGBOOST, AND RANDOM FOREST MODELS FOR SOLAR POWER PREDICTION

Date

2025-7-23

Author

Awan, Luqman

Metadata

Show full item record

This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.

Item Usage Stats

7771
views

0
downloads

Cite This

The growing demand for solar energy systems as a sustainable power source entails accurate forecasting and an improved model for the dependability and efficiency. This research uses the machine learning models like Random Forest, XG Boost (Extreme Gradient Boosting), and LSTM, i-e Long Short-Term Memory, for predicting the solar power generation because in the large data set these machine learning models can manage the non-linearity and complications. On the solar power system, large data set these studies try to examine whether it could be better approach of prediction by applying these machine learning models. This large dataset of solar power plant is taken from a reliable source, which is NREL stands for National Renewable Energy Laboratory, USA for training and testing purposes, including their features. Exceptional accuracy levels have been demonstrated by these models. The LSTM model achieved an R-squared value of 0.9745 and a mean absolute error of 0.6478. XG boost model obtain best accuracy among these models which is of 0.9965 and its mean absolute error is 0.1533 while random forest accuracy level is slightly down by few points, R-squared value 0.9958 and its mean absolute error is 0.1711. These exceptional results show that XG boost is best in precise power prediction and efficiency. This research explains better strategy for energy management which is pressing topic of the day while also enhance the power plant forecasting. Keywords: Machine Learning Models, Solar Power Prediction, XG Boost, Random Forest, LSTM.

Subject Keywords

Machine Learning Models, Solar Power Prediction, XG Boost, Random Forest, LSTM.

URI

https://hdl.handle.net/11511/115481

Collections

Northern Cyprus Campus, Thesis