一种改进的压缩频繁模式挖掘算法

On Improved Algorithm for Mining Compressed Frequent Patterns

传统的频繁模式挖掘算法产出大量的频繁模式,这些大量的频繁模式对于后期的分析、处理以及理解都带来了相当大的挑战.针对上述问题,该文提出了一种改进的压缩频繁模式挖掘算法,该算法基于CFP-树通过对传统频繁模式挖掘算法的改进能够从大量的频繁模式中选择出规模更小的频繁模式代表集合,从而起到减小庞大的频繁模式挖掘结果规模的目的.实验中还将该算法和现有的RPlocal算法进行了对比,结果表明改进的压缩频繁模式挖掘算法能够在合理的时间及容错范围内获得更小的频繁模式代表集,同时大大降低了时间复杂度,提高了挖掘效率.

Frequent pattern mining often produces a large number of frequent patterns, which brings a great challenge to understanding and further analysis of the generated patterns. In this paper, in response to this shortcoming, an improved compressed frequent pattern mining method based on CFP-Tree has been proposed. It can be found that this method serves as a smaller number of representative patterns to best approximate all other patterns than the traditional methods and achieves the purpose of reducing the huge frequent pattern mining results scale. The experimental evaluation shows that the new method can get smaller frequent pattern represents set within a reasonable time and fault-tolerant range than the tradition- al RPlocal algorithm and be faster. Thus, it improves performance greatly.