基于支持量的并发序列模式挖掘方法  

Concurrent sequential pattern mining method based on support vector

作　　者：王翠青[1] 杨晓彤[1] 陈未如[1] 

机构地区：[1]沈阳化工大学计算机科学与技术学院,辽宁沈阳110142

出　　处：《计算机工程与设计》2016年第1期156-162,共7页Computer Engineering and Design

基　　金：辽宁省科技攻关基金项目(2012219001);辽宁省教育厅科技基金项目(L2013157)

摘　　要：在研究并发序列模式性质的基础上,设计实现基于支持量的并发序列模式挖掘算法。通过为每个序列模式及并发序列模式集求支持量,在遍历序列模式同时利用支持量的求与操作尽可能多的将序列模式加入并发集中。在生成数据和真实数据上的实验验证了该算法效率较高,得到了有意义并发序列模式集。Based on analyzing properties of the concurrent sequential patterns,an algorithm was developed using support vector.The support vector of every sequential pattern was obtained,and each sequential pattern was made as seed-sequence of concurrent sequential pattern and the sequential pattern set was scanned.Sequential patterns were put into the concurrent sequential pattern as more as possible.The algorithm was running on real datasets as well as synthetic datasets,and the meaningful concurrent sequential patterns were got.

关 键 词：结构关系模式 并发序列模式 支持量 序列模式 数据挖掘 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程] TP311[自动化与计算机技术—控制科学与工程]