摘要
针对电力通信网异常流量识别中传统方法存在的规则库不完整、识别准确率较低的问题,设计了基于自监督学习的快速识别方法。该方法采用直接规则构建异常流量识别规则库,选取无标注通信流量数据集,在自监督学习框架下,通过异常流量的概率密度筛选非线性信号,生成快速识别基线,从而实现异常流量的高效识别。实验结果表明,在多种流量数据中,该方法识别准确率均达到98%及以上,能够准确区分异常流量与正常流量,有效保障电力通信安全。
In response to the problems of incomplete rule libraries and low recognition accuracy in traditional methods for identifying abnormal traffic in power communication networks,this paper designs a fast recognition method based on self supervised learning.This method uses direct rules to construct an abnormal traffic recognition rule library,selects an unlabeled communication traffic dataset,and in a self supervised learning framework,filters nonlinear signals through the probability density of abnormal traffic to generate a fast recognition baseline,thereby achieving efficient identification of abnormal traffic.The experimental results show that the recognition accuracy of this method reaches 98%or above in various flow data,and it can accurately distinguish abnormal flow from normal flow,effectively ensuring the security of power communication.
作者
赵双梅
魏亮
ZHAO Shuangmei;WEI Liang(State Grid Jibei Electric Power Co.,Ltd.Qinhuangdao Power Supply Company,Qinhuangdao 066000,China)
出处
《电工技术》
2025年第22期173-175,共3页
Electric Engineering
关键词
自监督学习
电力通信网
异常流量
快速识别方法
识别标签
self supervised learning
electric power communication network
abnormal traffic
quick identification method
identify tags
