基于模糊关联规则挖掘的网络入侵检测算法

Network intrusion detection algorithm based on fuzzy association rules mining

为了有效解决当前网络入侵检测算法存在的缺陷,提高网络的安全性,提出基于模糊关联规则挖掘的网络入侵检测算法。首先收集网络数据,提取网络入侵行为的特征;然后采用模糊关联规则算法对入侵行为特征进行挖掘,选择入侵行为最有效的特征,减少特征之间的关联度;最后支持向量机根据"一对多"的思想建立网络入侵检测的分类器,以KDD CUP数据为例对网络入侵检测性能进行分析。结果表明,该算法的网络入侵检测正确率超过了95%,检测结果要明显好于其他检测算法,易实现,可以用于大规模网络的在线入侵检测分析。

In order to solve the shortcomings existing in the current network intrusion detection algorithm effectively,and improve the network security,a network intrusion detection algorithm based on fuzzy association rules mining is proposed.The network data is collected to extract the features of the network intrusion behavior.The fuzzy association rules algorithm is used to mine the intrusion behavior features,select the most effective feature of intrusion behavior,and reduce the correlation among the features.The support vector machine is used to establish the classifier of the network intrusion detection according to thethought of"one-to-many".The KDD CUP data is taken as an instance to analyze the performance of network intrusion detection.The results show that the network intrusion detection accuracy of this algorithm is higher than95%,its detection result is obviously better than that of other detection algorithms,the algorithm is simple to implement,and can be used to the online intrusion detection analysis of the large?scale network.