September 25, 2025EnergiesOpen Access

Short-Term Load Forecasting in the Greek Power Distribution System: A Comparative Study of Gradient Boosting and Deep Learning Models

View Full Paper

Ask AI

Bookmark

Authors

Discussion

This study compares several supervised learning models for load forecasting, highlighting gradient boosting's efficiency over deep learning.

LightGBM achieved the best performance metrics among all models tested, indicating its superiority in load forecasting.
Among various models, LightGBM showed a MAE of 69.12 MW and an R2 value of 0.9942 for the test set.
The approach used in this study involved extensive preprocessing, training on a multivariate dataset, and a 70:20:10 data split.
Findings reveal trade-offs between accuracy and complexity, establishing potential for real-world smart grid applications.

Shiblee et al. (2025) studied this question.

View Full Paper

Ask AI

Bookmark

This study compares several supervised learning models for load forecasting, highlighting gradient boosting's efficiency over deep learning.