一种基于密度的无监督聚类算法

A Density-based Unsupervised Cluster Algorithm

分析了k-means算法的缺陷、入侵检测特点和网络中数据的特点,提出了一种基于密度的无监督2次聚类算法—KD算法。该算法聚类使用改进的k-means算法并引入基于密度聚类算法的优点,以提高对单种入侵数据集及混合入侵数据集的检测效果。实验结果表明,该算法具有较高的检测率和较低的误检率。

Focusing on the defects of k-means algorithm and the features of intrusion detection,an improved cluster algorithm is promoted,called KD algorithm.This algorithm makes use of the improved k-means algorithm and takes the advantages of density-based cluster algorithm,so it can improve the invasion detection result for single and mixed intrusion detection data sets.The experimental results show that this algorithm has effectively detection rate and lower false alarm rate.