Application Research On The Network Fault Management Technique Based On Rules Mining

The stability of computer network always holds more people's attention. A main method ofimproving the network stability is that diagnosing the fault of the working network in time andinstructing the network administrator's decision. It can improve the reliability, stability of thenetwork management and the validity and velocity of fault diagnosing that applying the data miningtechnology in diagnosing the fault of the network.This thesis in-depth researched the FP-growth algorism in association rules, put forward aimproved algorism and testified its validity with experiments;it researched the updating problem onthe association rules in the condition that the min-support and min-confidence were fixed and thedata in network alarm database was increasing, put forward an efficient incremental updatingalgorism for association rules;it studied the time series in network alarm events as well and broughtforward an algorism of temporal association rules based on sliding-window;it devised andimplemented a network fault diagnosis system based on rules mining, mined the network alarmdatabase with this system. The result suggested that this system is potent.Wu Yingzhi (computer application technology)...