本文提出了一种基于随机指纹模型的Wu and Manber(WM)算法(Randomizing Fingerprint WM,RFP-WM),它通过为每一个模式串计算唯一指纹可以有效降低误报率.与WM算法相比,RFP-WM算法极大地降低了哈希冲突率,提高了命中率,在海量模式集上这一效果更为显著.实验结果表明,相对于传统WM算法,该算法的匹配效率更高,而且模式集的规模越大,性能越优越.
Abstract
This paper presents a randomizing fingerprint-based Wu and Manber(WM) algorithm(RFP-WM)
which can effectively reduce false positives rate by calculating a unique fingerprint for each pattern.Compared with WM algorithm
RFP-WM algorithm greatly reduces the hash collision rate and increases the hit rate
especially in the massive patterns set.Experiment results show that the performance of the RFP-WM algorithm is more superior than traditional Wu and Manber(WM) algorithm on the larger pattern set.