Answers are grounded in this paper and cite their sources.
Accurate electricity load forecasting is crucial for the scientific formulation of power dispatch plans and the achievement of supply-demand balance in power systems. To address the current challenges, namely, the difficulty in extracting key features from electricity load data and the insufficient consideration of multi-channel data characteristics, which result in low forecasting accuracy, a parallel load forecasting model optimized by an attention mechanism is proposed in this paper. The model architecture consists of two parallel branches: in Branch One, an Effective Channel Attention Network (ECANET) is employed to optimize Temporal Convolutional Networks (TCN), while an Informer network is utilized in Branch Two. In addition, elastic net regression is applied to reduce the dimensionality of high-dimensional meteorological data, by which the accuracy of load forecasting is further enhanced. Experimental results based on real-world datasets demonstrate that the proposed method achieves an RMSE of 170.65 kW, an MAE of 17.35%, and an R 2 of 0.9732, which outperforms several benchmark models, including LSTM, BiLSTM, Transformer, and Informer. Compared with the LSTM model, the RMSE of the proposed method is reduced by 57.9%, while the lowest MAE and the highest R 2 among all compared approaches are achieved. Furthermore, generalization experiments under different batch sizes show that stable performance is maintained by the proposed model, and the effectiveness of the attention mechanism and dual-branch architecture is confirmed by ablation studies.
No takes yet. Share an insight, caveat, or question.
Xu et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: