基于互关联后继树的多时间序列关联模式挖掘  

Mining Relationship Patterns in Multiple Time Series Based on IRST

作　　者：曾海泉[1] 刘永丹[1] 宋扬[1] 胡运发[1] 

机构地区：[1]复旦大学计算机与信息技术系,上海200433

出　　处：《计算机研究与发展》2003年第7期934-940,共7页Journal of Computer Research and Development

基　　金：国家自然科学基金 ( 60 173 0 2 7)

摘　　要：时间序列是现实生活中常见的数据形式之一 ,在时间序列中发现频繁模式是分析时间序列变化规律的一项重要任务 提出基于互关联后继树的多时间序列关联模式挖掘算法 该算法首先用Allen逻辑位置关系来描述序列状态关系 ,根据这些关系在时间窗口内顺序或并行出现情况 ,获得一个由这些关系组成的特殊序列 在此基础上提出了一个基于互关联后继树的新型挖掘模型 ,实现了序列间关联模式的挖掘 与其他方法相比 ,该算法简单、直观 ,而且整个挖掘过程不需要生成候选模式 。Time series are an important type of data, and discovering frequent patterns in time series is a basic task to predicate the changing trend In this paper, an algorithm for the discovery of frequent patterns in multiple time series is proposed In this algorithm, firstly the states relationship between in time series is represented in Allen temporal logic, and then a sliding window is used to examine the order or occur rence relationship of states and obtain a particular sequence On the basis of the sequence, a frequent pattern method called inter related successive trees(IRST), is developed to find the frequent relationship patterns from multiple time series Compared with the previous methods, the method is more simple, flexible, efficient and more applied value

关 键 词：时间序列 数据挖掘 关联模式 互关联后继树 

分 类 号：TP182[自动化与计算机技术—控制理论与控制工程]