【摘要】 针对客户交易数据的特点,提出了一种基于前缀映射累加树的客户时序关联规则发现方法。将时间窗口内频繁项的信息映射到前缀映射累加树中,以降低频繁时序模式的搜索空间,提高时序关联规则的挖掘效率。另外,通过为特定的频繁项建立前缀映射累加树,可以挖掘特定的时序关联规则,并能以较精确的方式,发现具有一定模糊性的客户时序关联规则。实验结果表明,所提出的方法能够提高客户时序关联规则的挖掘效率。更多还原

【Abstract】 According to the characteristics of customer’s transaction database,a novel customer temporal sequence association rules mining method was proposed which was based on prefix projected accumulation tree.The searching space of frequent temporal sequence patterns could be reduced and the efficiency of association rule mining could be improved by projecting frequent items in time windows to prefix projected accumulation trees.Through building of prefix projected accumulation tree for specific frequent items,the proposed method could regard specific rule as a mining goal and find fuzzy rules in precise way.Experimental results indicated that the proposed method has improved the efficiency of customer temporal sequence association rule mining.更多还原