Abstract: By studying distribution of the pattern number in sequential pattern mining using different support degree, this paper proposes a relation model of support and numbers of sequential pattern. Based on mining on subset of custom sequential database, it uses the relation model to provide users with the reference for determining threshold of the support degree. It uses this method in several different data sets, which gets the expected results, and demonstrates this method is correct and efficient.

Key words: data mining, sequential pattern mining, support degree

摘要： 通过对不同支持度下序列模式挖掘产生模式个数分布的研究，利用曲线拟合技术，提出一种支持度与序列模式个数的关系模型。在对客户序列数据库子集进行预挖掘的基础上，利用该模型为用户在挖掘前确定支持度阈值提供参考。在不同类型数据集上采用该方法，得到预期结果，表明该方法是正确有效的。

关键词: 数据挖掘, 序列模式挖掘, 支持度

CLC Number: 

• TP311