分布式网络动态数据异常区域时序挖掘仿真  

Distributed Network Dynamic Data Anomaly Region Timing Mining Simulation

作　　者：张凯斐 ZHANG Kai-fei(Department of Computer Science and Technology,Lvliang University,Shanxi Lvliang 033000,China)

机构地区：[1]吕梁学院计算机科学与技术系

出　　处：《计算机仿真》2019年第9期361-364,401,共5页Computer Simulation

摘　　要：为改善当前分布式网络异常区域动态时序数据挖掘过程中受冗余、干扰数据影响,造成挖掘准确率不高、误检率和漏检率居高不下的问题,提出了基于小波发分析的分布式网络异常区域动态时序数据挖掘方法,该方法通过采用小波分析的方法对分布式网络中产生的动态时序数据进行多尺度分解和平滑滤波处理,消除了冗余和干扰数据影响;在此基础上,引入网格作为索引计算将分布式网络中的动态时序数据活动空间进行网格划分,同时结合二元正态密度核函数和二进制序列法挖掘分布式网络异常区域以及异常区域动态数据的活动周期规律,实现了分布式网络异常区域动态时序数据挖掘。在MATLAB软件环境下模拟分布式网络场景,选取检测率、误检率、漏检率作为评价指标,测试了注入不同异常类型后所提方法的挖掘性能,并对比了注入不同比例异常动态时序数据时所提方法与其它方法的挖掘准确性,充分证明了所提方法的有效性与优越性。This article proposed a method to mine the dynamic time series data of abnormal region in distributed network based on wavelet analysis.Firstly,this method used wavelet analysis method to perform multi-scale decomposition and smoothing filtering on dynamic time series data in distributed network,so as to eliminate the influence of redundancy and interference data.On this basis,the grid was introduced as the index computing to mesh the active space of dynamic time series data in distributed network.Meanwhile,the binary normal density kernel function and the binary sequence method were combined to mine the active cycle law of the anomaly area in distributed network and the dynamic data of abnormal area.Thus,dynamic time series data of abnormal area in distributed network was mined.In addition,the distributed network scene was simulated in MATLAB software environment.The detection rate,false detection rate and missed detection rate were selected as the evaluation indicators.The mining performance of the proposed method after introducing different abnormal types was tested.Finally,the mining accuracy of the proposed method is compared with the mining accuracy of other methods,when the different proportions of abnormal dynamic time serried data are introduced,so that the effectiveness and superiority of the proposed method can be proved.

关 键 词：分布式网络 异常区域 动态 时序数据 挖掘 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]