基于XGBoost与多源域迁移学习的贫资料地区负荷预测方法

唐義坤; 唐志远; 罗同桐; 谯傲; 刘俊勇

doi:10.19725/j.cnki.1007-2322.2024.0038

欢迎来到能源电力期刊网！

English Version 期刊登录

您当前的位置：

首页 >

文章列表页 >

基于XGBoost与多源域迁移学习的贫资料地区负荷预测方法

智能电网 | 更新时间：2026-06-02

- 基于XGBoost与多源域迁移学习的贫资料地区负荷预测方法
- Load Forecasting in Data-scarce Areas Based on Multi-source Transfer Learning and XGBoost
- 现代电力 2026年43卷第3期页码：464-474
- 作者机构：
  
  1. 四川大学电气工程学院,四川省,成都市,610065
  2. 成都信息工程大学物流学院,四川省,成都市,610065
- 作者简介：
  
  唐義坤(1997)，男，硕士研究生，研究方向为机器学习和负荷预测，E-mail：tangyikun@stu.scu.edu.cn
  唐志远(1992)，男，博士，副研究员，通信作者，研究方向为电力系统控制及规划运行、分布式控制及优化，E-mail：tangzhiyuan@scu.edu.cn
  罗同桐(2000)，女，硕士研究生，研究方向为机器学习在电力系统中的应用，E-mail：905951698@qq.com
  谯傲(1992)，女，博士，讲师，研究方向为绿色供应链，E-mail：qiaoao@cuit.edu.cn
  刘俊勇(1963)，男，博士，教授，研究方向为电力市场、智能配电网、电力系统大数据技术，E-mail：liujy@scu.edu.cn
- 基金信息：
  
  四川省科技计划项目(2023YFSY0033)。
- DOI：10.19725/j.cnki.1007-2322.2024.0038
  中图分类号： 3
- 纸质出版：2026
- 稿件说明：
移动端阅览
唐義坤, 唐志远, 罗同桐, 等. 基于XGBoost与多源域迁移学习的贫资料地区负荷预测方法[J]. 现代电力, 2026,43(3):464-474.

TANG Yikun, TANG Zhiyuan, LUO Tongtong, et al. Load Forecasting in Data-scarce Areas Based on Multi-source Transfer Learning and XGBoost[J]. 2026, 43(3): 464-474.
唐義坤, 唐志远, 罗同桐, 等. 基于XGBoost与多源域迁移学习的贫资料地区负荷预测方法[J]. 现代电力, 2026,43(3):464-474. DOI： 10.19725/j.cnki.1007-2322.2024.0038.

TANG Yikun, TANG Zhiyuan, LUO Tongtong, et al. Load Forecasting in Data-scarce Areas Based on Multi-source Transfer Learning and XGBoost[J]. 2026, 43(3): 464-474. DOI： 10.19725/j.cnki.1007-2322.2024.0038.

摘要

准确的负荷预测对保障电力系统安全稳定运行非常关键，贫资料地区由于缺乏相关的有效数据集，难以使用传统的机器学习预测方法。针对小样本条件下的负荷预测问题，提出一种基于极限梯度提升树（extreme gradient boosting，XGBoost）与多源域迁移学习框架。首先，为度量源域与目标域之间的可迁移性，利用XGBoost构建域数据的特征–输出映射关系，并引入一种基于夏普利可加性解释（Shapley additive explanations，SHAP）值分布的相似性度量方法，计算域之间特征–输出映射关系的相似性。然后，选择与目标域高度相似的多个源域模型作为迁移模型，并通过最优组合系数将各个源域模型的输出结果集成，以提高整体的迁移效果。最后，使用开源数据对所提出的框架进行性能评估和有效性验证，相比于多层感知机（multi-layer perception，MLP）、支持向量回归（support vector regression，SVR）、长短期记忆网络（long short-term memory network，LSTM）和XGBoost方法的直接预测，所提方法能提高至少10%的预测精度；相较于基于形态相似性的迁移学习方法，所提方法能提升选择有效源域的准确性。

Abstract

Accurate electric load forecasting

especially with small sample sets

is a critical topic in the field of power system. In this paper

we propose an attentive multi-source transfer learning framework based on extreme gradient boosting (XGBoost) for load forecasting in data-scarce scenarios. In the proposed framework

a new quantifiable framework is developed by exploiting the feature-output mapping relationship constructed by XGBoost. To effectively assess the transferability across domains

a similarity metric is introduced grounded in SHAP value distribution. Models of multiple sources that exhibit high similarity to the target domain are selected as the transfer models

and attentively integrated using the optimal combining coefficients

aiming to enhance overall transfer performance. Experiments conducted on real-world data sets illustrate that

compared with another popular transfer learning method

the proposed framework not only effectively identifies the potential source tasks that can produce positive transfer

but also achieves improved forecasting accuracy for residential electric load with limited data.

关键词

Keywords

references

KONG Xiangyu, MA Yuying, AI Qian, et al . Review on electricity consumption characteristic modeling and load forecasting for diverse users in new power system[J ] . Automation of Electric Power Systems, 2023, 47(13): 2−17(in Chinese).

FANG Na, LI Junxiao, CHEN Hao, et al . Multi-frequency combination short-term power load forecasting with convolutional neural networks - bidirectional gated recurrent unit-multiple linear regression based on variational mode decomposition[J ] . Mode rn Electric Power, 2022, 39(04): 441−448(in Chinese).

HAN F , PU T , LI M, et al . Short-term forecasting of individual residential load based on deep learning and k-means clustering[J ] . CSEE Journal of Power and Energy Systems, 2021, 7(2): 261-269.

FANG Lei, CHU Chengbo, HE Yinghong, et al . Power load forecasting in station areas based on asgmd -mlr-clstm[J ] . Modern Electric Power, 2025, 42(04): 840−846(in Chinese).

GAN Zhongxue, YU Xiangxiang, XU Yuli, et al . Air-conditioning cooling and heating load prediction based on periodic arma-svr model[J ] . Control Engineering of China, 2020, 27(2): 380−385(in Chinese).

FAN Shixiong, LIU Xingwei, YU Yijun, et al . Multi-source data and hybrid neural network based ultra-short-term bus load forecasting[J ] . Power System Technology, 2021, 45(1): 243− 250(in Chinese).

LI Mingjie, LIANG Zhifeng, XU Tao, et al . Prediction and application of maximum daily load in summer and winter based on spatial distribution of sensitive temperatures[J ] . Power System Technology, 2023, 47(03): 1088−1098(in Chinese).

HU Wei, ZHANG Xinyan, LI Zhenen, et al . Short-term load forecasting based on an optimized VMD-mRMR-LSTM model[J ] . Power System Protection and Control, 2022, 50(01): 88−97(in Chinese).

SAJJAD M , KHAN Z , ULLAHET A , et al . A novel cnn-gru-based hybrid approach for short-term residential load forecasting[J ] . IEEE Access, 2020, 8: 143759-143768.

ZHUANG F, QI Z, DUAN K, et al . A comprehensive survey on transfer learning[J ] . Proceedings of the IEEE, 2021, 109(1): 43-76.

OUYANG Fulian, WANG Jun, ZHOU Hangxia. Short-term power load forecasting method based on improved hierarchical transfer learning and multi-scale CNN-BiLSTM-Attention[J]. Power System Protection and Control, 2023, 51(2): 132−140(in Chinese).

CAI L, GU J , JIN Z. Two-layer transfer-learning-based architecture for short-term load forecasting[J]. IEEE Transactions on Industrial Informatics, 2020, 16(3): 1722-1732.

WU D, LIN W. Efficient residential electric load forecasting via transfer learning and graph neural networks[J]. IEEE Transactions on Smart Grid, 2023, 14(3): 2423−2431.

D’INCECCO M, SQUARTINI S, ZHONG M. Transfer learning for non-intrusive load monitoring[J]. IEEE Transactions on Smart Grid, 2020, 11(2): 1419−1429.

ZHOU Z, XU Y, REN C. A transfer learning method for forecasting masked-load with behind-the-meter distributed energy resources[J]. IEEE Transactions on Smart Grid, 2022, 13(6): 4961−4964.

ZENG P, WU D, JIN M. Compress-filtering and transfer-expanding of data set for short-term load forecasting[C]//2017 International Joint Conference on Neural Networks, Anchorage(IJCNN). AK, USA: IEEE, 2017: 1095-1101.

WANG Z, DAI Z, POCZOS B, et al . Characterizing and avoiding negative transfer[C ] //2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). Long Beach, CA, USA: IEEE, 2019: 11285-11294.

WEI Zetao, LIU Youbo, SHEN Xiaodong, et al . Ultra-short term power generation modeling and prediction for small hydropower in data-scarce areas based on sample data transfer learning[J ] . Proceedings of the CSEE, 2023, 43(07): 2652−2666(in Chinese).

TIAN Y, SEHOVAC L, GROLINGER K. Similarity-based chained transfer learning for energy forecasting with big data[J]. IEEE Access, 2019, 7: 139895−139908.

WU D, WANG B, PRECUP D, et al . Multiple kernel learning-based transfer regression for electric load forecasting[J ] . IEEE Transactions on Smart Grid, 2020, 11(2): 1183−1192.

WEI Z, SHEN X, QIN G, LIN Y, LIU J. An integrated transfer learning method for power generation prediction of run-off small hydropower in data-scarce areas[J]. IEEE Transactions on Smart Grid, 2024, 15(1): 1030−1041.

ZHU Jiangxing, ZOU Xiaosong, XIONG Wei, et al . Short-term power load forecasting based on prophet and xgboost mixed model[J ] . Modern Electric Power, 2021, 38(03): 325−331(in Chinese).

WANG M, ZHENG K, YANG Y, et al . An explainable machine learning framework for intrusion detection systems[J ] . IEEE Access, 2020, 8: 73127−73141.

WANG Yixing. Power data mining technology based on deep learning and transfer learning[D]. Hangzhou: Zhejiang University, 2019(in Chinese).

浏览量

175

下载量

CSCD

文章被引用时，请邮件提醒。

提交

工具集

关联资源

基于时间卷积网络与迁移学习的短期负荷预测

基于K-means聚类与LSTM模型的多能源耦合电力负荷预测

基于长短期记忆网络模型的联邦学习居民负荷预测

一种改进Transformer的电力负荷预测方法

基于数据驱动响应约束的配电网可靠性评估方法