摘要
针对在时间序列数据流中挖掘频繁串行情节的问题,提出了一种具有可持续挖掘的方法—TFSE(Tim e-tab le-joined Frequent Serial Ep isodes)。该方法引入了情节时间表和再次挖掘的概念,一个情节模式对应一个情节时间表,通过在情节时间表之间做数据库联接操作,生成相应新的情节时间表。情节时间表记录个数即是该情节的支持数。
Aiming at mining frequent episodes in time serial data stream, an approach of Time -table -joined Frequent Serial Episodes (TFSE) is presented, which includes the conception of the episode - time - table and the re - mining. An Episode - time - table corresponds to an episode pattern and a new Episode - time - table is generated by the joining databases. The amount of records in the Episode - time - table is equal to the number of episodes occurring in the time series.
出处
《武汉理工大学学报(信息与管理工程版)》
CAS
2006年第5期12-16,共5页
Journal of Wuhan University of Technology:Information & Management Engineering
关键词
数据挖掘
时间序列
情节时间表
频繁串行情节
data mining
time series
episode - time - table
frequent serial episodes