模糊FP-growth在入侵检测中的应用

关联规则挖掘技术目前被广泛应用于入侵检测系统中。关联规则挖掘算法之一的FP-growth算法在处理数值量的输入时需要二值化,使得准确率不高;而Fuzzy Apriori算法需要重复扫描数据库,效率较低。针对此问题,改进现有的FP-growth算法,提出模糊化FP-growth算法,从而提取模糊关联规则,用于N类异常数据的分类入侵检测。在KDDCup＇99数据集上评估,结果表明对于数值量的输入,该方法应用于入侵检测准确率高于FP-growth算法,学习效率高于Fuzzy Apriori算法。

Application of Fuzzy FP-growth to Intrusion Detection

Association rule mining is widely used in IDS.The FP-growth algorithm,as an association rule mining algorithm,needs to be binarized in processing numerical input data,resulting in low accuracy,while Fuzzy Apriori algorithm demands repeated scan of database,resulting in low efficiency.For this problem,the fuzzy FP-growth algorithm based on modification of FP-growth is proposed,thus to extract fuzzy association rules for the classification intrusion detection of N class abnormal data.This method is assessed in KDDCup＇99 data set,and the results show that,for numerical input,the detection accuracy of this method is higher than that of FP-growth algorithm,its learning efficiency is higher than that of Fuzzy Apriori.