摘要
为提高光纤通信网络异常数据检测能力,研究基于数据挖掘的光纤通信网络异常数据检测方法。采用改进关联聚类算法划分光纤通信网络数据,通过聚类操作获取异常数据;采用基于正则化相关熵异常检测的特征提取方法,结合半二次优化技术获取光纤通信网络异常数据特征向量;以光纤通信网络异常数据特征为依据构建熵目标函数,通过大量迭代获取熵目标函数最优值,实现光纤通信网络异常数据最优检测。实验结果表明:该方法的异常数据检测结果与光时域反射仪一致,且不受数据占用率以及网络、路由所处状态影响,检测正确率及检测效率均较高。
In order to improve the abnormal data detection ability of optical fiber communication network,a data mining-based abnormal data detection method of optical fiber communication network is studied.The improved associative clustering algorithm is used to divide the optical fiber communication network data,and the abnormal data is obtained through the clustering operation.The feature extraction method based on the regularization correlation entropy abnormal detection is adopted,and the semi-quadratic optimization technology is used to obtain the abnormal data feature vector of the optical fiber communication network.The characteristics of abnormal data of communication network are based on the construction of entropy objective function,and the optimal value of entropy objective function is obtained through a large number of iterations,so as to realize the optimal detection of abnormal data of optical fiber communication network.The experimental results show that the abnormal data detection results of this method are consistent with the optical time domain reflectometer,and are not affected by the data occupancy rate and the state of the network and routing,and the detection accuracy and detection efficiency are high.
作者
王永强
李子龙
王杜鑫
张驰俊
林婷
WANG Yong-qiang;LI Zi-long;WANG Du-xin;ZHANG Chi-jun;LIN Ting(Meizhou Power Supply Bureau of Guangdong Power Grid Co.,Ltd.,Meizhou 514021 China)
出处
《自动化技术与应用》
2024年第11期111-114,共4页
Techniques of Automation and Applications
基金
国家级新一代人工智能科技项目(2020AAA0103400)。
关键词
数据挖掘
光纤通信
异常数据
相关熵
隶属度
检测
data mining
optical fiber communication
abnormal data
correlation entropy
membership degree
detection
