摘要
局域网传输的数据报中携带大量与数据包相关的信息,这些信息在一定意义上反映了数据报的行为。对数据报行为进行分类可为局域网上的网络入侵检测提供重要依据。文中提出使用C4.5决策树分类算法对局域网数据报进行行为分类,并与以往常用的几种分类算法进行了比较。实验表明,C4.5算法对于该问题无论在分类效率还是在分类正确性方面均有很大的优势。
A mass of package - related information is carried by datagrana on LANs （ local area networks）. This information reflects the action and behavior of the datagram. Classify the datagram on LANs can provide some evidences of NID （network intru.sion detect）. A decision tree - based classification algorithm by using CA. 5 algorithm is introduced in this paper. Compared with other classification algorithms that usually used, this method shows distinctly advantages in both efficiency and precision in classifying datagram on LANs.
出处
《计算机技术与发展》
2006年第7期1-3,共3页
Computer Technology and Development
