基于粗糙集约简算法的配置文本聚类方法研究  

Research on Clustering Method of Configuration Text Based on Rough Sets Reduction Algorithm

作　　者：唐启涛[1] 张燕[1] 彭利红[1] 

机构地区：[1]长沙医学院计算机科学与技术系,湖南长沙410219

出　　处：《计算机技术与发展》2015年第11期105-109,共5页Computer Technology and Development

基　　金：湖南省教育科学技术研究项目(14C0114)

摘　　要：当前的网络设备的配置文本日趋复杂,在网络设备出现故障时采集到的数据量也随之倍增。但是,并不是所有的配置文本信息都是有用的。文中提出了通过对配置文本应用基于DAG思想的配置元集无关性算法,实现消除配置文本中的冗余信息,只保留配置文本中有用的信息。对于每一次网络设备配置故障诊断,因所采用通信信道、采集设备的不同致使获得的信息无法保证它的完备性和正确性,因此,想获得理想的故障诊断结果通过传统的方法行不通。在对配置文本信息进行了预处理后,为了便于对网络设备配置文本故障进行智能、快速的诊断,提出了一种基于粗糙集约简算法的配置命令文本聚类方法,实现按功能的不同对预处理后的设备配置命令文本进行分类。最后,利用Simulink仿真软件比较配置文本归类与不归类在故障诊断时的差别。Configuration text of the current network equipment is becoming more complicated, the amount of information acquisition is more and more in network equipment malfunction. However, not all information of configuration text is useful. The configuration element of ideologies of independent algorithm is put forward based on DAG used in configuration text, which eliminates redundant information, retains only the useful information in the configuration text. For the every configuration fault diagnosis of network equipment, because the acquisition equipment, communication channel and other factors make the information obtained cannot guarantee the complete and correct- ness, therefore, the fault diagnosis results with traditional methods is not ideal. After the configuration text information is preprocessed,in order to facilitate the intelligent and rapid diagnosis for the network equipment configuration text fanlt,present a method of configuration text clustering based on rough sets reduction algorithm, realizing classification of network equipment configuration text command by different function after pretreatment. Finally, compare the configuration text difference of text categorization and not classified in fault diagnosis by using Simulink simulation software.

关 键 词：网络设备 配置元集 故障诊断 粗糙集 有向无环图 

分 类 号：TP301[自动化与计算机技术—计算机系统结构]