| 基于聚类分析的非监督式异常检测研究 |
| |
| 引用本文： | 刘燕,梁云娟.基于聚类分析的非监督式异常检测研究[J].河南科技学院学报,2006,34(2):80-83. |
| |
| 作者姓名： | 刘燕 梁云娟 |
| |
| 作者单位： | [1]西安电子科技大学计算机学院,陕西西安710071 [2]新乡医学院计算机教研室,河南新乡453003 [3]河南科技学院,河南新乡453003 |
| |
| 摘 要： | 传统的异常检测方法要求训练数据集完全由已标记为正常的实例所构成，但在实际应用中，很难得到这样的训练数据集。本文提出了一种基于聚类分析的非监督式异常检测方法，该方法的优点在于不需要任何标记数据，并且能够实现网络连接数据的实时检测。实验采用KDD99数据集进行测试，结果表明，该方法具有比较高的检测性能。
|
| 关 键 词： | 入侵检测 异常检测 非监督 聚类 |
| 文章编号： | 1673-6060（2006）02-0080-04 |
| 收稿时间： | 2006-01-09 |
Unsupervised Anomaly Detection Research Based on Clustering Analysis |
| |
| LIU Yan ,et al..Unsupervised Anomaly Detection Research Based on Clustering Analysis[J].Journal of Henan Institute of Science and Technology(Natural Science Edition),2006,34(2):80-83. |
| |
| Authors: | LIU Yan |
| |
| Institution: | 1. School of Computer Engineering, Xidian Univ. , Xian 710071, China ;2. Section of Computer Science, Xinxiang Medical College, Xinxiang 543003, China |
| |
| Abstract: | Traditional anomaly detection requires a set of purely normal data from which to train their model. But in practice,we do not have either labeled data or purely normal data. This paper presents a clustering based method for unsupervised anomaly detection. The benefit of method is that it neednt unlabeled data and it can perform the real - time detection of the network connection data. Using the data sets of KDD99, this approach is proved to be with high performances by resuit. |
| |
| Keywords: | intrusion detection anomaly detection unsupervised clustering |
| 本文献已被 维普 万方数据 等数据库收录！ |
