基于相邻模式段组合的生物序列模式挖掘算法

Joined pattern segment-based sequential patternmining algorithm for biological datasets

传统的序列模式挖掘算法应用在生物序列上有其局限性,根据生物序列的特点,提出了基于相邻频繁模式段的模式挖掘算法-JPS。首先产生相邻频繁模式段,然后对这些频繁模式段进行组合,产生新的频繁模式。通过实验分析,该方法在相似性很强的序列数据库中比传统的PrefixSpan算法效率高。通过对真实的蛋白质序列家族库的处理,证明该算法能有效处理生物序列数据。

Traditional algorithms for sequential pattern mining have limits when dealing with biological datasets.Biology sequence has its own characters.Based on these characters,the author develops Joined frequent Pattern Segment approach,JPS,for mining biological sequences.First,the joined frequent pattern segments are produced.Then,longer frequent patterns can be obtained by combining the above segments.The experiment shows JPS has better performance than PrefixSpan.Through dealing with the real protein family database,it is proved that the algorithm can deal with biology sequence data efficiently.