基于改进SVM算法的电力工程异常数据检测方法设计  

Design of power engineering abnormal data detection method based on improved SVM algorithm

作　　者：王楠[1] 周鑫 周云浩 苏世凯 王增亮 WANG Nan;ZHOU Xin;ZHOU Yunhao;SU Shikai;WANG Zengliang(Electric Power Construction Engineering Consulting Branch,State Grid Beijing Electric Power Company,Beijing 100021,China)

机构地区：[1]国网北京市电力公司电力建设工程咨询分公司,北京100021

出　　处：《电子设计工程》2024年第4期162-166,共5页Electronic Design Engineering

基　　金：北京电力公司输变电工程应用项目(SGBJJS00XSJS2100639)。

摘　　要：针对传统电力工程数据异常检测过程中存在准确度差且主观性较强的问题,文中提出了一种基于改进支持向量机的电力工程数据异常检测模型。其在传统支持向量机的基础上加入了二叉树多分类算法,从而使模型具备多特征分类能力。同时通过引入AdaBoost分类器,来改善支持向量机弱特征分类能力较差的不足。为进一步提高准确度,还使用鲸鱼算法对模型惩罚项、核函数及迭代次数进行优化。在实验测试中,所提算法的检测准确度相较其他三种对比算法分别提升了5.35%、2.17%和5.35%,说明该算法具备更为理想的性能,并可有效提升电力工程数据检测的准确度,故能为电力基建工程验收与管理提供数据支撑。Aiming at the shortcomings of poor accuracy and strong subjectivity in the process of traditional power engineering data anomaly detection,this paper proposes a power engineering data anomaly detection model based on improved support vector machine.The binary tree multi classification algorithm is added to the traditional support vector machine,which makes the model have the ability of multi feature classification.At the same time,in order to improve the weak feature classification ability of support vector machine,AdaBoost classifier is also introduced.In order to further improve the accuracy,the whale algorithm is used to optimize the penalty term,kernel function and iteration times of the model.In the experimental test,the detection accuracy of the proposed algorithm is improved by 5.35%,2.17%and 5.35% respectively compared with the other three comparison algorithms,indicating that the algorithm has more ideal performance,effectively improves the detection accuracy of power engineering data,and can provide data support for the acceptance and management of power infrastructure projects.

关 键 词：支持向量机 ADABOOST算法 鲸鱼优化算法 二叉树结构 异常数据分析 

分 类 号：TN-9[电子电信]