一种保序序列快速挖掘算法:RSMM

RSMM:a Fast Order-preserving Sequential Pattern Mining

时间序列记录的是某一统计量按照时间推移而发生变化的数据,寻找合理的挖掘算法解决时间序列问题具有很强的现实意义。提出一种保序序列挖掘方法,通过子模式匹配结果挖掘(read the sub-pattern matching for mining,RSMM)算法,挖掘时间序列中频繁出现的趋势变化,在计算支持度时根据子模式的匹配结果得到超模式的支持度,在一遍扫描时间序列的情况下挖掘出所有的频繁保序模式。从理论上证明了RSMM是满足Apriori性质的完备性算法。在真实数据集上进行的实验表明,与其他对比算法相比,运行时间显著减少,从而验证了RSMM算法的高效性。

Time series could record the data that a certain statistic changes according to the passage of time,such as stock data.Finding a reasonable mining algorithm to solve time series problems is valuable.An order-preserving pattern mining algorithm RSMM(read the sub-pattern matching for mining)was proposed.The support of the super-pattern was obtained according to the matching results of the sub-pattern.The frequent order-preserving patterns could be mined by scanning the time series once.RSMM was a complete algorithm that satisfied the properties of Apriori was proved.Experiments were conduct on real data sets.The experimental results on real datasets indicated that compared with other comparison algorithms,the running time was significantly reduced,which verified the efficiency of RSMM algorithm.