半监督聚类算法及其在入侵检测中的应用

针对现有入侵检测技术的不足,对基于机器学习的异常入侵检测系统进行了研究,提出了一种基于半监督聚类的异常入侵检测算法。此算法通过利用少量的标记样本,生成用于初始化算法的种子聚类,然后辅助聚类过程,对数据进行检测。实验表明,与以往入侵检测算法相比,此算法可以明显地改善入侵检测系统的性能。

Semi-Supervised Clustering Algorithm and Its Application to Intrusion Detection

Aiming at some problems in current technique of intrusion detection and studying the anomaly intrusion detection system based on machine learning,an anomaly intrusion detection system based on semi-supervised clustering is proposes.It uses a little quantity of labeled data to generate seeds which applied in initiating the algorithm,and then assists the clustering process to detect data.The experiment results manifest that comparing with the past algorithm,this algorithm can significantly improve the perform...