Massive real-time fine-grained electricity consumption data has brought short-term load forecasting (STLF) into the era of big data in electricity and this leads the high computing cost of training short-term load forecasting models (SLFM) for massive STLF scenarios. To address this issue, this paper proposes a SLFM transfer method (SLFMTM), which follows the principle of transfer learning and captures the characteristics of STLF. Firstly, temporal and distribution similarity measurements are proposed, which guarantees the upper bound of transferred model performance. Secondly, SLFMTM applies hierarchical clustering method to cluster load sequences, in which temporal similarity is used for measuring load sequence distance and distribution similarity is used for determining the best number of clusters. Thirdly, feature selection results of several STLF scenarios are taken intersection as the inputs for all scenarios in the same cluster to decrease computing time. Finally, SLFMTM proposes a bagging method for fine-tuning to decrease generalization error. The evidence from an empirical study in Guiyang, China shows that: 1) SLFMTM can maintain the SLFM performance without degrade, and 2) SLFMTM greatly reduces the SLFM computing cost by minimum of 92.66%, 92.27% and 90.9% respectively.

Short-term load forecasting (STLF) is indispensable for the efficient and cost-effective operation of power systems. The Kolmogorov-Arnold network (KAN), which is grounded in the Kolmogorov-Arnold representation theorem, exhibits superior interpretability and symbolic regression capabilities when juxtaposed with artificial neural networks (ANNs). It offers analytically expressible load formulas, making it a subject of recent research in short-term load forecasting. Nonetheless, as the depth of ANNs escalates, their predictive prowess frequently equals or even outstrips that of the KAN. To address this issue, this paper proposes a novel STLF method through the utilization and combination of Kolmogorov-Arnold Networks (KANs) and ANNs based on the idea of residual decomposition. KANs are utilized to uncover fundamental periodic and nonlinear patterns in the load, while the remaining residual component is fitted using ANNs. A hybrid model combining KANs and ANNs with residual connections is designed, explicitly satisfying load additivity. Additionally, regularization penalties are incorporated to ensure that the predictions made by KANs dominate, meeting the requirements of residual decomposition. Furthermore, a training method for the hybrid KAN-ANN model is proposed, which refines the analytical expressions obtained by KANs alone and provides more accurate forecast results. To validate the proposed method, we compared the forecasting performance of KANs, multi-layer perceptrons, and XGBoost using a public Switzerland dataset. Numerical results demonstrate that the proposed hybrid KAN-ANN method not only enhances the forecasting performance of the KAN-based method but also yields an analytical expression for STLF. Our code can be found in https://github.com/BozhenJIANG/KAN-ANN-for-STLF.

A document by Bozhen Jiang. Click on the document to view its contents.