K-Modes Clustering Algorithm Based on Weighted Overlap Distance and Its Application in Intrusion Detection

In order to better apply the K-modes algorithm intrusion detection, this paper overcomes problems of existing based on rough set theory. Firstly, for problem clustering in initial class center selection, an selection Ini_Weight weighted density and overlap distance is proposed. Secondly, algorithm, a new WODKM Thirdly, applied detection obtain unsupervised model. The model detects by dividing clusters result into normal abnormal analyzing average object x be detected each cluster overlapping point. We verified performance KDD Cup 99 dataset. experimental results current study show that proposed achieves efficient solves present-day system some extent.