关于网络用户访问数据优化检测仿真

摘要

对网络用户访问数据的高效检测,可更好的提升网络的运行质量.对网络用户访问数据的检测,需要对数据集中的时间序列数据进行特征提取,计算数据正交坐标系之间最小距离,完成对访问数据的检测.传统方法对用户访问数据多元时间序列的协方差矩阵进行特征分解,但忽略了计算正交坐标系之间的距离,导致数据检测精度低.提出基于聚类划分的面向网络的用户访问数据检测方法.融合于层次凝聚方法对面向网络的用户访问数据集进行层次分解,计算数据集中任意两个对象之间的距离,采用簇间最小距离对不同类型的数据进行凝聚簇,获取高质量的簇中心,组建基于信息熵的数据微聚类过滤机制,对数据集中的时间序列数据进行特征提取,计算数据正交坐标系之间最小距离,以计算的结果为依据完成对面向网络的用户访问数据检测.实验结果表明,所提方法检测精度高,可以有效地提升网络的服务质量.%A detection method for access data of network user is proposed based on cluster division.Firstly,the hierarchical decomposition for the set of access data integrated with hierarchical clustering is carried out,and then distance between any two targets in data set is calculated,and agglomeration cluster for different types of data is carried out by using the minimum distance between clusters to acquire centre of cluster with high quality.Moreover,the filtering mechanism of data micro-cluster is built based on information entropy to carry out feature extraction for data of time series in data set,and the minimum distance between orthogonal coordinate systems is calculated.Finally,according to the result,the detection is completed.Simulation results show that the method has high detection precision.It can improve quality of service of network effectively.