MPSQAR:无损语义的量化关联规则挖掘算法(英文)  

MPSQAR:Mining Quantitative Association Rules without Loss of Semantics

作　　者：曾春秋[1] 唐常杰[1] 李川[1] 段磊[1] 

机构地区：[1]四川大学计算机学院,成都610065

出　　处：《计算机科学与探索》2009年第4期392-404,共13页Journal of Frontiers of Computer Science and Technology

基　　金：The National Natural Science Foundation of China under Grant No.600773169;the National Great Project of Scientific and Technical Supporting Programs Funded by Ministry of Science & Technology of China During the 11th Five-year Plan under Grant No.2006BAI05A01~~

摘　　要：在挖掘量化关联规则的过程中,由于对量化值的划分,将产生语义损失。为避免这种情况,提出基于无损语义的算法MPSQAR来处理量化关联规则的挖掘。主要工作包括:(1)提出规泛化量化值的新方法;(2)提出反映属性值分布的属性权重设计方法;(3)扩展加权关联规则模型以处理量化关联规则,避免量化值的划分;(4)提出挖掘传统布尔关联规则和量化关联规则的集成方法;实验表明算法MPSQAR的有效性和时间消耗随时间趋势呈线性增长。During the process of mining quantitative association rules, the semantics may be lost due to the discretization of quantitative values. To avoid the loss of semantic information, a novel algorithm, MPSQAR （mining preserving semantic quantitative association rule）, is proposed to handle the quantitative association rules mining. The main contributions include ： （ 1 ） Propose a new method to normalize the quantitative values ; （ 2 ） Propose a method to assign a weight for each attribute to reflect the values distribution; （3） Extend the weight-based association model to tackle the quantitative values in association rules without partition; （4） Design a integrated and uniform method to mine the traditional Boolean association rules and quantitative association rules; Experiments show the effectiveness and linear scalability of the new method on time consuming.

关 键 词：量化关联规则 MPSQAR算法 语义信息损失 

分 类 号：TP301[自动化与计算机技术—计算机系统结构]