Network Fault Detection - A Case for Data Mining
Modern mobile communication networks are tremendously composite systems, usually capable of limited self-diagnosis. Parts of the general network fault management problem, namely, fault detection, isolation and diagnosis has been taken up in this communication. A model has been proposed to cluster the network fault data using k-mean algorithm followed by its classification through C4.5 algorithm. Side by side the clustered data are used for training using neural network. The proposed model results into betterment in terms of classification of large data set, decreased network faults, enhanced accuracy.