Abstract: This paper discusses a problem of temporal approximate periodicity. It presents an approximate periodic pattern on the basis of temporal type, introduces the concepts of approximate precision and approximate periodic pattern mantle, and proves some relative properties. The paper discusses an algorithm based on self-organizing map to find approximate periodic pattern. Experiments show that proposed algorithms are efficient.

Key words: Data mining, Temporal type, Approximate periodic pattern, SOM

摘要：

研究了时态近似周期的挖掘问题，提出了近似周期模式，引进了近似精度、近似周期模式覆盖等概念及性质，提出了一个基于SOM (自组织特征映射)聚类来寻找近似周期模式的算法，实验表明算法是有效的。

关键词: 数据挖掘, 时态型, 近似周期模式, SOM

CLC Number: 

• TP311