A K-MEANS Algorithm Applied in Network Anomaly Detection

This paper studied a known cluster centers of k-means clustering algorithm for anomaly detection of network.The algorithm avoids the low detection rate of network anomaly detection problems for traditional clustering algorithms for random selecting initial cluster centers.Experiment demonstrated that the algorithm is feasible.Resultsshowed that this algorithm comparing to traditional clustering algorithms in terms of detection rates have been greatly improved.Through the unsupervised learning method to obtain the detection of new attacks.