Article: Short-term electricity price forecasting based on Attention-TCN-GRU Journal: International Journal of Energy Technology and Policy (IJETP) 2025 Vol.20 No.3/4 pp.317 - 333 Abstract: The electricity price is dynamic along with the demand-supply balance in electricity market. With the large-scale integration of the renewable energy sources and the flexible demand, the short-term electricity price fluctuates with much uncertainty risk. In view of the problems such as incomplete consideration of relevant influencing factors and insufficient information mining, this paper proposes a short-term electricity price prediction model based on Attention-TCN-GRU. The model considers the behaviour of different groups of electricity consumers. The temporal convolution network (TCN) and gated recurrent unit (GRU) are combined to facilitate the dual learning of the correlation between various factors and the price of electricity and the internal connection of multiple factors. The attention mechanism is used to further enhance the weight of key factors and build a deep prediction model of short-term electricity price in a complex market environment. The experimental results verify the validity and practicability of this model. Inderscience Publishers - linking academia, business and industry through research

Title: Short-term electricity price forecasting based on Attention-TCN-GRU

Authors: Shengjie Yang; Youwei Gong; Wenjun Zhao; Zhijun Zhou; Liang Ge

Addresses: School of Computer, Hunan University of Technology and Business, Changsha, 410205, China ' School of Computer, Hunan University of Technology and Business, Changsha, 410205, China ' School of Computer, Hunan University of Technology and Business, Changsha, 410205, China ' School of Computer, Hunan University of Technology and Business, Changsha, 410205, China ' Hunan Electric Power Trading Center Co., Ltd., Changsha, 410029, China

Abstract: The electricity price is dynamic along with the demand-supply balance in electricity market. With the large-scale integration of the renewable energy sources and the flexible demand, the short-term electricity price fluctuates with much uncertainty risk. In view of the problems such as incomplete consideration of relevant influencing factors and insufficient information mining, this paper proposes a short-term electricity price prediction model based on Attention-TCN-GRU. The model considers the behaviour of different groups of electricity consumers. The temporal convolution network (TCN) and gated recurrent unit (GRU) are combined to facilitate the dual learning of the correlation between various factors and the price of electricity and the internal connection of multiple factors. The attention mechanism is used to further enhance the weight of key factors and build a deep prediction model of short-term electricity price in a complex market environment. The experimental results verify the validity and practicability of this model.

Keywords: electricity price forecast; temporal convolutional network; TCN; gated recurrent unit; GRU; power user behaviour; attention mechanism.

DOI: 10.1504/IJETP.2025.151061

International Journal of Energy Technology and Policy, 2025 Vol.20 No.3/4, pp.317 - 333

Received: 26 Oct 2024
Accepted: 25 Jul 2025
Published online: 12 Jan 2026 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article

Title: Short-term electricity price forecasting based on Attention-TCN-GRU

Keep up-to-date