TITLE
MACHINE-GENERATED NEURAL NETWORKS FOR SHORT-TERM LOAD FORECASTING
AUTHOR(S)
Gergana Vacheva1, Plamen Stanchev2, Nikolay Hinov1*
ABSTRACT
This paper presents a comprehensive study on machine-generated neural networks for short-term load forecasting (STLF), focusing on their ability to predict power demand accurately over short periods. Effective STLF is vital for utility companies to maintain balance between electricity supply and demand, optimizing operational efficiency and reducing costs. This study examines neural networks generated through neural architecture search (NAS), an automated machine learning approach that optimizes neural network structures specifically for load forecasting tasks. By leveraging NAS, this approach enhances forecasting accuracy and adaptability by dynamically adjusting to patterns in energy consumption data. Results indicate that machine-generated networks outperform traditional and manually designed models in STLF, highlighting the potential of automated network design in complex time-series forecasting applications.
DOI
http://www.doi.org/10.70456/KQVP8332
DOWNLOAD
https://unitechsp.tugab.bg/images/2024/1-EE/s1_p143_v1.pdf
How to cite this article:
Gergana Vacheva, Plamen Stanchev, Nikolay Hinov, MACHINE-GENERATED NEURAL NETWORKS FOR SHORT-TERM LOAD FORECASTING, UNITECH – SELECTED PAPERS - 2024