基于AC算法的比特流频繁序列挖掘

为解决比特流频繁序列挖掘效率不高以及易受用户数据影响而导致准确率低的问题,首先从理论上论证了短频繁序列挖掘存在的局限性,根据不同长度的频繁序列挖掘时存在的特点,将其分为长频繁序列与短频繁序列,提出比特流协议头部字段定位算法；基于AC多模式匹配算法分别针对长、短频繁序列挖掘的不同特点,提出了相应的挖掘方法,提高了挖掘结果的准确性。最后通过实验验证了所提算法的有效性。

Frequent Pattern Mining in Bit Stream Based on AC Algorithm

The existing method of frequent pattern mining in bit stream is inefficient and the precision of the results is low under the influence of redundant data.In order to solve the problem,it was proved that the mining of short frequent pattern has great limitations.According to the different features when the patterns are mined with different lengths,the frequent sequences are divided into two types:long frequent pattern and short frequent pattern.An algorithm of finding the header fields of the protocol in bit stream was proposed,and the efficient algorithm of mining the long frequent pattern and the short frequent pattern were proposed based on AC multi-pattern matching algorithm.Simulation results on the Ethernet show that the proposed algorithm is effective.