摘要
用挖掘跨事务关联规则的方法分析多元时间序列,可以找到序列中不同采样点观察值之间相互影响的关系。本文为实现这一目的,提出一种新的分析方法:ES-Appiori。此方法通过减少数据库扫描次数,优化内存分配,能够高效地分析多元时间序列之间的关联规则。试验表明,用此方法分析中国证券市场的股票时间序列非常有效。
We can use methods that mine inter-transaction association rules to analysis multiple time series for finding their relationships. In this paper, a new algorithm named ES-Apriori will be presented to mine inter-transaction association rules in multiple time series. This algorithm needs to search the whole database only one time, with the rational arrangement of memory usage it can successfully mine the association rules in time series. Experiments have shown that this method can efficiently analyze the time series of Chinese Stock Market.
出处
《计算机科学》
CSCD
北大核心
2004年第3期108-111,共4页
Computer Science
