一种基于费歇(Fisher)判别法的异常检测模型设计

Design of an anomaly detection model based on Fisher-discriminance

网络上的入侵事件层出不穷,这对信息资源的安全构成了严重威胁。应对这些恶意行为的重要措施之一就是入侵检测。该技术的一个重要分支———异常入侵检测技术在目前的网络安全领域研究中十分活跃。文中提出一种基于统计理论———费歇(Fisher)判别法的异常入侵检测模型,该模型与其他的异常检测模型相比,在对异常事件响应能力的实时性与精确性方面有了较为显著的提高,应用该模型也简化了入侵检测系统设计的复杂性。

With the growth of the Internet, events of intrusion become more and more usual, which does serious harm to information resources. To prevent these malicious behaviors, one of the important measures is the intrusion detection technique. In recent years, anomaly detection - an important branch of intrusion detection is a highlighted topic in the studies of the network security. Based on the Fisher-discriminance, a statistical theory, a novel anomaly detection model, which, compared with other anomaly detection models, would improve apparently the performance in real-time and accuracy was presented. Also, the complexity in the design of intrusion detection system would be simplified when this model was employed.