IPv6远程监控网络下无状态通信数据的多尺度离群点挖掘算法

doi:10.11959/j.issn.1000-0801.2023149

摘要/Abstract

摘要：

为了准确挖掘离群点，降低离群点对通信数据造成的影响，对IPv6远程监控网络无状态通信数据多尺度离群点挖掘算法进行了研究。通过IPv6远程监控网络获得无状态通信数据，依据提取的无状态通信数据的季节性、趋势性和自相似性特征，运用傅里叶变换将无状态通信数据分为两类。再用K均值法对两类分别进行聚类，确定无状态通信数据的邻域，将其作为基础。采用卷积神经网络对无状态通信数据进行离群点挖掘，初始化卷积神经网络；根据卷积神经网络输出值，判别该网络是否符合停止条件，反复重复卷积神经网络的运算步骤，挖掘全部离群点，实现无状态通信数据多尺度离群点挖掘。实验结果表明，无状态通信数据类别的个数越少，挖掘效率越高；所提方法能准确挖掘IPv6远程监控网络无状态通信数据多尺度离群点的个数，准确分析离群原因。

关键词: IPv6, 远程监控网络, 无状态, 通信数据, 多尺度, 离群点挖掘

Abstract:

In order to accurately mine outliers and reduce the impact of outliers on communication data, a multi-scale outlier mining algorithm for stateless communication data in IPv6 remote monitoring network was investigated.The stateless communication data were obtained through an IPv6 remote monitoring network, and based on the seasonality, trend, and self-similarity characteristics of the extracted stateless communication data, the Fourier transform was used to divide the stateless communication data into two classes.Then, the K-mean method was used to cluster the two classes to determine the neighborhood of the stateless communication data, which was used as the basis for outlier mining using a convolutional neural network on the stateless communication data.The convolutional neural network was initialized, and according to the output value of the convolutional neural network, it was determined whether the network met the stopping condition.The operation steps of the convolutional neural network were repeated, all the outlier points were mined, and the multi-scale outlier mining of stateless communication data was achieved.The experimental results showed that the fewer the number of stateless communication data categories, the higher the mining efficiency; the proposed method can accurately mine the number of multiscale outliers of stateless communication data in IPv6 remote monitoring network and accurately analyze the reasons for the outliers.

Key words: IPv6, remote monitoring network, stateless, communication data, multi-scale, outlier mining

中图分类号:

TP181

刘琨, 张晓涵, 曹汝坤, 李帅. IPv6远程监控网络下无状态通信数据的多尺度离群点挖掘算法[J]. 电信科学, 2023, 39(8): 118-126.

Kun LIU, Xiaohan ZHANG, Rukun CAO, Shuai LI. Multi-scale outlier mining algorithm for stateless communication data under IPv6 remote monitoring network[J]. Telecommunications Science, 2023, 39(8): 118-126.

图/表 9

图1

图2

图3

图4

图5

表1

图6

表2

图7

参考文献 15

[1]	丑义凡, 易波, 王兴伟 ,等. IPv6网络中基于MF-DL的DDoS攻击快速防御机制[J]. 计算机学报, 2021,44(10): 2047-2060.
	CHOU Y F , YI B , WANG X W ,et al. A rapid defense mechanism based on MF-DL for DDoS attack in IPv6 networks[J]. Chinese Journal of Computers, 2021,44(10): 2047-2060.
[2]	罗晓媛, 赵丽艳, 刘君 ,等. 神经网络技术下多尺度时序数据离群点挖掘[J]. 计算机仿真, 2021,38(1): 231-235.
	LUO X Y , ZHAO L Y , LIU J ,et al. Mining outliers in multi-scale time series data based on neural network technology[J]. Computer Simulation, 2021,38(1): 231-235.
[3]	张倩倩, 于炯, 李梓杨 ,等. 基于近邻传播的离群点检测算法[J]. 计算机应用研究, 2021,38(6): 1662-1667.
	ZHANG Q Q , YU J , LI Z Y ,et al. Outlier detection algorithm based on affinity propagation[J]. Application Research of Computers, 2021,38(6): 1662-1667.
[4]	康耀龙, 冯丽露, 张景安 ,等. 基于谱聚类的高维类别属性数据流离群点挖掘算法[J]. 吉林大学学报(工学版), 2022,52(6): 1422-1427.
	KANG Y L , FENG L L , ZHANG J A ,et al. Outlier mining algorithm for high dimensional categorical data streams based on spectral clustering[J]. Journal of Jilin University (Engineering and Technology Edition), 2022,52(6): 1422-1427.
[5]	王晓辉, 宋学坤, 王晓川 . 基于邻域密度的异构数据局部离群点挖掘算法[J]. 计算机仿真, 2021,38(7): 281-285.
	WANG X H , SONG X K , WANG X C . Local outlier mining algorithm for heterogeneous data based on neighborhood density[J]. Computer Simulation, 2021,38(7): 281-285.
[6]	梁勇, 刘承启 . IPv6 远程监控网络无状态双向通信方法[J]. 计算机仿真, 2021,38(2): 119-123.
	LIANG Y , LIU C Q . IPv6 remote monitoring network stateless two-way communication method[J]. Computer Simulation, 2021,38(2): 119-123.
[7]	蔡顺, 耿豪鹏, 郑炜珊 ,等. 基于傅里叶变换的谷间距特征信息提取及其影响因素研究[J]. 地球信息科学学报, 2020,22(3): 399-409.
	CAI S , GENG H P , ZHENG W S ,et al. Valley spacing character information and its influencing factors based on the Fourier transform[J]. Journal of Geo-Information Science, 2020,22(3): 399-409.
[8]	孟笑天, 徐艳蕾, 王新东 ,等. 基于改进 K 均值特征点聚类算法的作物行检测[J]. 农机化研究, 2020,42(8): 26-30.
	MENG X T , XU Y L , WANG X D ,et al. Crop line detection based on improved K-means feature point clustering algorithm[J]. Journal of Agricultural Mechanization Research, 2020,42(8): 26-30.
[9]	马永军, 汪睿, 李亚军 ,等. 利用聚类分析和离群点检测的数据填补方法[J]. 计算机工程与设计, 2019,40(3): 744-747,761.
	MA Y J , WANG R , LI Y J ,et al. Data filling using cluster analysis and outlier detection[J]. Computer Engineering and Design, 2019,40(3): 744-747,761.
[10]	崔斌, 张永红, 闫利 ,等. 一种基于卷积神经网络的 SAR 变化检测方法[J]. 测绘科学, 2019,44(6): 170-175,186.
	CUI B , ZHANG Y H , YAN L ,et al. A method of SAR change detection based on convolutional neural networks[J]. Science of Surveying and Mapping, 2019,44(6): 170-175,186.
[11]	马京晖, 潘巍, 王茹 . 基于K-means聚类的三维点云分类[J]. 计算机工程与应用, 2020,56(17): 181-186.
	MA J H , PAN W , WANG R . 3D point cloud classification based on K-means clustering[J]. Computer Engineering and Applications, 2020,56(17): 181-186.
[12]	李晓莉, 韩鹏, 李晓光 . 基于典型样本的卷积神经网络技术[J]. 计算机工程与设计, 2020,41(4): 1113-1117.
	LI X L , HAN P , LI X G . Convolutional neural network based on typical samples[J]. Computer Engineering and Design, 2020,41(4): 1113-1117.
[13]	许璟琳, 彭阳, 余芳强 . 基于 K-means 聚类和离群点检测算法的医院建筑节能诊断方法[J]. 计算机应用, 2021,41(S1): 288-292.
	XU J L , PENG Y , YU F Q . Energy-saving diagnosis method for hospital buildings based on K-means clustering and outlier detection algorithm[J]. Journal of Computer Applications, 2021,41(S1): 288-292.
[14]	神显豪, 李驰, 桂琼 ,等. 基于卷积神经网络的网络节点异常数据检测方法[J]. 机床与液压, 2020,48(22): 18-23.
	SHEN X H , LI C , GUI Q ,et al. A method for detecting abnormal data of network nodes based on convolutional neural network[J]. Machine Tool ＆ Hydraulics, 2020,48(22): 18-23.
[15]	赵晓永, 王宁宁, 王磊 . 基于主动学习的离群点集成挖掘方法研究[J]. 计算机工程与应用, 2020,56(12): 112-117.
	ZHAO X Y , WANG N N , WANG L . Research of outlier ensemble mining based on active learning[J]. Computer Engineering and Applications, 2020,56(12): 112-117.

数据集/万	类别个数/个	本文方法挖掘次数比	本文方法时间比	文献[4]方法挖掘次数比	文献[4]方法时间比
10	30	1.298	2.10	1.775	3.02
	20	1.134	1.09	1.593	2.16
	10	1.008	1.01	1.487	2.02
30	30	1.197	3.25	1.768	3.46
	20	1.090	1.43	1.647	2.98
	10	1.015	1.05	1.319	2.61
50	30	1.101	4.78	1.541	5.88
	20	0.981	2.42	1.316	3.29
	10	0.981	1.08	1.149	3.02
70	30	1.098	5.58	1.532	6.46
	20	0.987	3.06	1.311	5.13
	10	0.889	1.14	1.135	3.64

离群点	离群原因
1	环境变化导致原本的数据不在传感器收集范围内
2	外部噪声太大导致数据异常
3	暴雨天气使传感器收集通信数据造成误差
4	数据获取来源不同
5	数据存储混乱
6	数据记录错误
7	远程监控系统出现损坏
8	服务器接收错误
9	计算错误