Edge computing and transfer learning-based short-term load forecasting for residential and commercial buildings

Energy and Buildings, Год журнала: 2025, Номер 329, С. 115273 - 115273

Опубликована: Янв. 7, 2025

Язык: Английский

---

Similarity-driven truncated aggregation framework for privacy-preserving short term load forecasting

Internet of Things, Год журнала: 2025, Номер unknown, С. 101530 - 101530

Опубликована: Фев. 1, 2025

Язык: Английский

---

Residential Load Forecasting Based on Long Short-Term Memory, Considering Temporal Local Attention

Sustainability, Год журнала: 2024, Номер 16(24), С. 11252 - 11252

Опубликована: Дек. 22, 2024

Accurate residential load forecasting is crucial for the stable operation of energy internet, which plays a significant role in advancing sustainable development. As construction internet progresses, proportion electricity consumption end-use increasing, peak on grid growing year year, and seasonal regional power supply tensions, mainly household consumption, grow into common problems across countries. Residential can assist utility companies determining effective pricing structures demand response operations, thereby improving renewable utilization efficiency reducing share thermal generation. However, due to randomness uncertainty user data, remains challenging. According prior research, accuracy using machine learning deep methods still has room improvement. This paper proposes an improved load-forecasting model based time-localized attention (TLA) mechanism integrated with LSTM, named TLA-LSTM. The composed full-text regression network, date-attention time-point network. network consists traditional while networks are local constructed CNN LSTM. Experimental results real-world datasets show that compared standard LSTM models, proposed method improves R2 by 14.2%, reduces MSE 15.2%, decreases RMSE 8.5%. These enhancements demonstrate robustness TLA-LSTM tasks, effectively addressing limitations models focusing specific dates time-points data.

Язык: Английский

---

Federated Graph Learning Based on Graph Distance Computing

Open Journal of Applied Sciences, Год журнала: 2024, Номер 14(11), С. 2985 - 2995

Опубликована: Янв. 1, 2024

Язык: Английский

---