移动核心网故障数据二重联合线性判别模型与应用  

Double joint linear discriminant model and applications on fault data recognition in mobile core network

作　　者：庞素琳[1,2] 汪寿阳[3] 

机构地区：[1]暨南大学公共管理学院/应急管理学院/金融工程研究所,广州510632 [2]广东省公共网络安全风险评价与预警应急技术研究中心,广州510632 [3]中国科学院数学与系统科学研究院,北京100190

出　　处：《系统工程理论与实践》2012年第12期2601-2610,共10页Systems Engineering-Theory & Practice

基　　金：国家自然科学基金(70871055);教育部新世纪优秀人才支持计划项目(NCET-08-0615);广东省科技计划项目(2010A032000002;2010B010600028);广东省高校高层次人才项目(2050205)

摘　　要：主要研究移动核心网BSC的TRAU数据分布规律及可疑故障识别方法.由于BSC的最大承载量是未知的,因此本文定义了一类TRAU数据突增故障现象,并在此基础上定义了可疑故障点、异常点和广义异常点.提出并建立了移动核心网故障数据二重联合线性判别模型和三类模式判别准则,用来对移动核心网TRAU数据进行三类模式("正常"、"异常"、"可疑故障")识别.实例分析表明,本文所建立的移动核心网故障数据二重联合线性判别模型,当用来对TRAU数据进行三类模式判别时,对原数据样本的回判和新数据样本的识别准确率都达到百分之百,因此能有效识别TRAU数据中的可疑故障点.This article mainly researches the distribution law of TRAU data of BSC in mobile core network and identifying method for a doubt fault. Because the maximum bearing capacity of BSC is unknown, it defines a class of the data up burst fault phenomenon of TRAU. On the basis, it further defines a doubtful fault point, the abnormal point and the generalized abnormal point. It proposes a double joint linear discriminant model for faults recognition in mobile core network and the discriminant laws with three patterns. They are used to do the three patterns （normal, abnormal, doubtful fault） recognition to the TRAU data in mobile core network. The example indicates that the proposed model can reach 100% classification accuracy within both the raw dataset and the new dataset when applied it to do the three patterns for the TRAU data. Therefore, the proposed model can availably identify the doubt fault point of TRAU data.

关 键 词：移动核心网 数据突增故障现象 可疑故障点 风险警界线 二重联合线性判别模型 

分 类 号：F830[经济管理—金融学]