Abstract: The paper presents two algorithms for efficient mining of partial periodic pattern with multiple time granularity.The first one prunes the number of candidate pattern by utilizing father-son relation of pattern and threshold.The second one,for each input finding all pattern which covers it,then computes pattern’s support rate and confidence rate.Experimectocl results show the proposed algorithms are efficient in mining partial periodic with multiple time granularity when input is large.

摘要： 提出了两种在多粒度时间下的部分周期挖掘算法。第1种算法利用模式的父子关系、支持度和置信度阈值裁剪候选模式集。第2种算法对每个输入寻找覆盖它的周期模式,由此来计算模式的支持度和置信度。实验表明了提出的算法在多粒度时间下挖掘部分周期是有效的。

关键词: 部分周期, 多粒度, 数据挖掘, 周期模式

CLC Number: 

• TP311.13