Out-of-distribution Generalization in Machine Learning for Power Systems: Problem Formulation and Analysis Framework

SHANG Yuwei; WANG Siyi; GUO Jianbo; SHENG Wanxing; LIANG Ying; {{custom_author.name_en}}

doi:10.13334/j.0258-8013.pcsee.250213

English Version Journal Login

您当前的位置：

首页 >

文章列表页 >

Out-of-distribution Generalization in Machine Learning for Power Systems: Problem Formulation and Analysis Framework

更新时间：2026-10-12

- Out-of-distribution Generalization in Machine Learning for Power Systems: Problem Formulation and Analysis Framework
- Vol. 46, Issue 7, Pages: 2626-2639(2026)
- 作者机构：
  
  中国电力科学研究院有限公司，北京市海淀区,100192
- 作者简介：
- 基金信息：
- DOI：10.13334/j.0258-8013.pcsee.250213
  CLC：
- Received：05 April 2026，
  
  Published：2026
- 稿件说明：
移动端阅览
SHANG Yuwei, WANG Siyi, GUO Jianbo, et al. Out-of-distribution Generalization in Machine Learning for Power Systems: Problem Formulation and Analysis Framework[J]. 2026, 46(7): 2626-2639.
DOI：

SHANG Yuwei, WANG Siyi, GUO Jianbo, et al. Out-of-distribution Generalization in Machine Learning for Power Systems: Problem Formulation and Analysis Framework[J]. 2026, 46(7): 2626-2639. DOI： 10.13334/j.0258-8013.pcsee.250213.

摘要

为保证机器学习(machine leaning，ML)模型的泛化性能，统计学习理论通常假设数据独立同分布。然而，由于电力系统的开放性和实际数据的动态演化性，数据分布的偏移难以避免，导致训练良好的模型在应用时效果往往大打折扣。为此，该文重点研究当数据非独立同分布时的ML问题，即分布外泛化问题(out-of-distribution generalization，OODG)。首先，结合电力ML特点，建立OODG的分析框架；在此基础上，通过理论分析和案例介绍相结合，探究数据分布偏移的量化方法、OODG能力的提升策略及其评估技术。深化OODG的认识和研究，有利于提升ML学习方法的生命力和实用价值。

Abstract

To ensure the generalization performance of machine learning (ML) models

the statistical learning theory typically assumes that data are independent and identically distributed (i.i.d.). However

due to the openness of the power systems and the dynamic evolution of actual data

data distribution shifts may be unavoidable

resulting in significantly performance degradationwhen well-trained models are deployed in practical applications. This paper tackles the problems of ML in scenarios where the training and test dataset are not i.i.d.

known as out-of-distribution generalization (OODG). First

the analysis framework of the OODG problem is established for power systems applications. Then

through the combination of theoretical analysis and case studies

the quantitative methods for assessing data distribution shifts

strategies for improving and evaluating the MLs’ OODG performance

are explored. By deepening the understanding and research of OODG

the viability and application values of ML can be further enhanced.

关键词

Keywords

references

Views

223

下载量

CSCD

Alert me when the article has been cited

提交

Tools

Publicity Resources

Unpacking the optimal distinctiveness mechanisms for leading enterprises to break through key green technologies

High-voltage and High-power Hybrid Commutated Converter Based on Reverse Blocking Integrated Gate Commutated Thyristor

Short-circuit Ratio Indicator for Grid-forming DC-supported Renewable Energy Radiative Integration System

The impact of green transformational leadership on radical green innovation

The impact of negative media coverage on innovation in high-tech firms

Related Author

No data

Related Institution

No data

AI问答

⁰