序列模式挖掘在物流中的应用

The Application of Sequence Mining in Logistics

当前第三方物流管理系统中以物流活动为对象的数据库庞大,难以发现有价值数据。为此,本文提出一种针对物流数据分析的经典方法:IGSP(improved sequential patterns)算法。该方法通过对原始序列数据库筛选,选出路径序列长度大于或等于候选序列长度的路径序列,进而有针对性地产生过度候选序列,经约减产生候选序列。利用这种产生候选序列的方式,能够有效地减少候选序列数量,进而产生物流中有意义的规则。案例和理论分析表明,该方法不仅缩小了扫描数据库的规模,而且减少了生成频繁序列的候选序列集合。

Currently the database for logistic actions in third party logistic system is very huge,so it is very difficult to find valuable data.Therefore,an efficient algorithm-IGSP(improved sequential patterns)for analyzing logistic data is presented.In this method the original database is screened to find the path sequences that is greater than or equal to the candidate sequences in the length,and then generate the candidate sequences through generating the transitional candidate sequences.This method could effectively generate frequent patterns by reducing the volume of sequences,and then find the valuable rules.Theoretical analysis shows that the method not only reduce the size of scanning database but also reduce the candidate sequences set.