Papers › Residential Load Forecasting: An Online-Offline Deep Kernel Learning Method

Residential Load Forecasting: An Online-Offline Deep Kernel Learning Method

28 Jul 2023IEEE Transactions on Power Systems 2023 7archive 2025-07-28

Yuanzheng Li, Fushen Zhang, Yun Liu, Huilian Liao, Hai-Tao Zhang, Chiyung Chung

Residential load forecasting (RLF) is critical for power system operations. Different from traditional system-level load forecasting, studying RLF faces the challenges of high uncertainty. Besides, learning temporal dynamics within the residential load sequences is important. However, existing methods fail to effectively tackle the fore-mentioned challenges simultaneously. In this article, a deep kernel is proposed by integrating the deep soft Spiking Neural Networks (sSNN), which is then applied to perform Gaussian Process (GP) regression. The constructed regressor investigates the temporal dynamics within the residential load sequence and retains the probabilistic advantages for uncertainty estimates. Furthermore, to better address the high uncertainty of RLF, a learning scheme combing both offline and online learning is specifically developed for the regressor. Such a learning scheme contributes to fully exploring historical information while learning the uncertainty from real-time data. The effectiveness of the proposed method is demonstrated on three public and actual residential load datasets.

PaperPDFCode

Code

Repository list and official/mentioned flags are the archive's, frozen 2025-07-28. Reachability, where shown, is from one Syntology probe window (2026-09-16 to 2026-09-18); repositories not probed show nothing. GitHub stars are not tracked.

Code Syntology ran Syntology

Not run by Syntology. Nothing on this page verifies that the listed code works.

Tasks

Load ForecastingProbabilistic Time Series Forecasting

Results from the paper archive 2025-07-28

No leaderboard rows for this paper in the archive.

Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections