摘要
高速网络环境中数据量日益增大,安全问题日益突出,对入侵检测技术提出了更高的要求。朴素贝叶斯作为数据挖掘的重要方法之一,在入侵检测中有着重要的地位。由于其属性独立假设,使得如何在海量高维数据处理背景下快速、准确、有效地选出代表原数据的属性显得尤为重要。本文提出了一种快速属性选择方法并结合朴素贝叶斯分类模型应用于入侵检测中。实验表明,结合了该属性选择方法的朴素贝叶斯分类器有很好的分类精度及较低的时空消耗。
As a consequence of the increasing security problems and huge data in high-speed networks, it is urgent to develop new Intrusion Detection Techniques. Naive Bayesian is one of the most important approaches in data mining, and it has a high ranker in Intrusion Detection. According to its conditional independence assumption, it is very critical to select attributes, which represent raw data, quickly, and effectively when processing huge high-dimension data in networks. This paper proposes a Fast Attribute Selection idea combined with Naive Bayesian Classifier, which is used in Intrusion Detecting. Experiment＇s results show that our idea has high precision and low cost.
出处
《计算机科学》
CSCD
北大核心
2008年第4期151-153,共3页
Computer Science
基金
安徽省自然科学基金课题(编号050420207)资助
