基于项目位置索引的序贯模式增量挖掘方法  

Item Position Index Based Pattern Growth Method for Sequential Pattern Mining

作　　者：梁铁柱[1] 李德毅[2] 宋云娴[3] 

机构地区：[1]解放军理工大学通信工程学院,南京210016 [2]总参61所 [3]空军工程大学工程学院

出　　处：《西安交通大学学报》2002年第10期1032-1036,共5页Journal of Xi'an Jiaotong University

基　　金：国家"九七三"重点基础研究发展规划资助项目 (G1 9980 30 50 8- 4 )

摘　　要：通过前缀序列的引入 ,将搜索空间划分为若干个子空间 ,利用模式增量技术对序贯模式进行有效搜索 ,并提出了项目位置索引的概念 ,即将原始序列数据库信息转换到项目位置索引 (IPI)中 ,从而在搜索序贯模式时避免了复杂的多维候选序列的测试 ,仅需对各前缀序列对应的扩展的项目位置索引库(IPIDBs)做简单的序列数目累加操作 ,将复杂的高维序贯模式搜索问题巧妙地转换为一维频繁项目的搜索 ,降低了算法复杂度 。Using prefix sequence, the search space is divided into many subspace small enough to be processed in the memory, and sequential patterns are searched with the pattern growth method. The concept of item position index (IPI) is also introduced here. After transferring the original sequence database into IPIs, instead of testing the candidate sequence, the searching of sequential patterns need only count the sequence number of each frequent item appeared in the corresponding IPIDBs of prefix sequences. This sequential patterns mining method successfully transfers the complicated problem of high-dimension sequential patterns mining into the searching of frequent items so that it is much easier, can reduce the complexity of algorithm and can raise the efficiency of computation.

关 键 词：序贯模式 项目位置索引 模式增量挖掘 数据库 人工智能 知识挖掘 搜索空间 

分 类 号：TP311.13[自动化与计算机技术—计算机软件与理论] TP18[自动化与计算机技术—计算机科学与技术]