The Study Of Rough Set Theory In Network Fault Dlagnosis

As computer networks increase in size, complexity and pervasiveness, effective fault diagnosis of such networks becomes more important and more difficult. In order to satisfy the modern network fault diagnosis and elimination, a special tool for the network management operator is required urgently. Network fault diagnosis is base on the state of network devices when the faults occurred, then by analyzing and judging these information to find the cause of the faults. And the most efficient and direct measure is collecting MIB(management information base) data in network devices . MIB data can reflect most state of network, and according to these data we can diagnose the network faults.In our task, we will combine rough set theory and network fault diagnosis. We can apply rough set algorithm in MIB data to find the relationship between MIB and network faults. First a decision table is constructed by collecting original MIB data from networks, then from the point of view of attribute significance and attribute frequency, we propose a heuristic rough set reduct algorithm using significance and frequency. Based on attribute core and user preference attribute set, the algorithm not only makes use of advantage of these two algorithm, but also the universality of core, and user background knowledge, domain experience. At the same time, we introduce the minimal support degree and minimal belief degree into rough set theory for decision rules discovery and propose a decision rule discovery algorithm based on minimal support degree and belief degree. This algorithm can discover the interested decision rules using minimal support and belief preset. At last this algorithm based on rough set, is implemented to remove inconsistent or erroneous MIB data, and accordingly concise information of network faults and relative MIB data is reserved. In this way the diagnostic rules for network faults can be obtained.