基于XGBoost与DBSCAN算法的停电故障聚类分析
Analysis of Power Outage Fault Clustering Based on XGBoost and DBSCAN Algorithms
许国杰 1孙亚林 1张龙 1潘虹 1宋醒亚1
作者信息
摘要
阐述配电网停电故障数据特征评估和筛选,采用DBSCAN算法对筛选后的特征进行密度聚类.实验表明,该方法使故障分类准确率提升至93.5％,较传统方法提高16.8％,且能准确识别异常故障类型.
Abstract
This paper describes the feature evaluation and screening of power outage fault data in distribution networks,and uses the DBSCAN algorithm to perform density clustering on the screened features.The experiment shows that this method improves the accuracy of fault classification to 93.5％,which is 16.8％higher than traditional methods,and can accurately identify abnormal fault types.
关键词
故障聚类/XGBoost算法/DBSCAN算法/特征工程Key words
fault clustering/XGBoost algorithm/DBSCAN algorithm/feature engineering引用本文复制引用
出版年
2024
NETL
NSTL
万方数据