XGBoost算法下供电线路停电敏感度识别
Sensitivity Identification of Power Supply Line Outages Under XGBoost Algorithm
卢海明 1辜小琢 1陈晓瑜 1李文珊 1王滢桦 1方立勤 1王思烨1
作者信息
- 1. 广东电网有限责任公司汕头供电局,广东 汕头 515041
- 折叠
摘要
由于忽略了数据时序特征,供电线路停电识别结果ROC曲线的AUC值不高.因此,提出XGBoost算法下供电线路停电敏感度识别方法.采用缺失补充以及归一化的方法处理供电线路数据时序特征,并分析相应的停电相似度量,采用XGBoost集成机器学习算法学习分析该相似度量的敏感度特征,结合贡献度分析特征属性值以识别出供电线路的停电敏感度.实验结果表明:应用供电线路停电敏感度识别方法后,得出的识别结果表现出的ROC曲线AUC值较高,识别准确度较高,满足了供电线路运维工作中对停电敏感度的信息需求.
Abstract
Due to ignoring the temporal characteristics of the data,the AUC value of the ROC curve for identifying power outa-ges in the power supply line remains suboptimal.Therefore,a sensitivity identification of power supply line outages based-on XGBoost algorithm is proposed.The missing supplementation and normalization methods are used to process the temporal characteristics of power supply line data and analyze the corresponding power outage similarity measure.What's more,the XG-Boost algorithm integrated machine learning algorithm is used to learn and analyze the sensitivity characteristics of this similari-ty measure and combine the contribution analysis feature attribute values to identify the power outage sensitivity of the power supply line.The experimental results show that the recognition results obtained after the application of the proposed method ex-hibit a better ROC curve AUC value and recognition accuracy,it be able to meet the information requirements for power outage sensitivity in power supply line operation and maintenance work.
关键词
供电线路/停电敏感度/XGBoost算法/敏感度识别/ROC曲线/离散化处理/数据时序特征/集成机器学习算法Key words
power supply lines/power outage sensitivity/XGBoost algorithm/sensitivity identification/ROC curve/discreti-zation processing/data temporal characteristics/integrated machine learning algorithms引用本文复制引用
出版年
2024