摘要
提出一种进行时间序列模式挖掘的算法 ,用于对大型数据库的海量数据分析 ,从中挖掘出超过用户给定支持度和置信度的时间序列 ,从而为用户的决策支持和趋势预测提供依据 .算法分为在数据中对于频繁项集的发现和频繁序列挖掘两个部分 ,排除不可能达到支持度和置信度阈值的项集 ,缩小了挖掘中的数据扫描范围 。
An algorithm for time sequential pattern mining is presented to analyze data in large database and find time sequences with higher support rates and higher confidence rates than user′s definition,providing support for user's decision making and trend forecast. The algorithm is composed of the detection of frequent term sets and mining of frequent sequences. The efficiency of data mining is improved by eliminating the term sets with lower support rate and confidence rate than the thresholds, and reducing the data scanning domain.
出处
《山东大学学报（工学版）》
CAS
2004年第3期88-91,共4页
Journal of Shandong University（Engineering Science）
关键词
数据挖掘
时间序列
序列模式
data mining
time sequence
sequential pattern
