SDH告警中的知识发现

TMN在电信网中起着越来越重要的作用，通信网的发展，要求网络管理向智能化方向发展．运用人工智能中知识发现和数据挖掘的方法，可以实现知识的自动获取．故障管理是网络管理的重要方面，本论文运用知识发现的方法对TMN子系统-SDH网管系统中传输告警的关联关系进行了挖掘．论丈运用知识发现的方法，以中国电信四川SDH网管系统中的告警数据为例，实现了SDH传输告警关联关系的挖掘算-Apriori算法；分析了重复告警对单前题规则可信度的影响，对算法做了改进，提高了规则的可信度，有利于发现重大告警；讨论了逆规则对规则有效性的影响，对算法进行了进一步的改进，减少了无效规则；对冗余规则进行了压缩，得到了更简明的挖掘结果；分析了部分挖掘结果和算法复杂度．

Knowledge discovery in SDH alarms

TMN is playing a very important role in telecommunication network. With the development of telecommunication network, the intelligent technology about network management is requested. This thesis mainly describes the knowledge discovery in SDH alarms. Firstly, this thesis introduces the concepts of TMN and SDH, presents the principle, procedure and method of knowledge discovery in database. Secondly, this thesis implements association mining by using Apriori algorithm in SDH alanns. Because duplicate alarms reduce the confidence of rule with single precondition, the algorithm is improved to increase the confidence of rule by handling the duplicates. This thesis also discusses the influence of reversed rules, compresses the redundant rules, analyzes a part of results of data mining, and analyzes the complication of the algorithm. Finally, this thesis describes other possible applications by using knowledge discovery in TMN.