一种新颖的关联规则数据挖掘算法  

A novel association rule data mining algorithm

作　　者：李珺[1] 宋文龙[1] 朱良宽[1] 

机构地区：[1]东北林业大学机电工程学院,黑龙江哈尔滨150040

出　　处：《辽宁工程技术大学学报（自然科学版）》2014年第6期846-849,共4页Journal of Liaoning Technical University (Natural Science)

基　　金：黑龙江省自然科学基金资助项目(QC2013C035);黑龙江省教育厅科学技术研究基金资助项目(12533014);哈尔滨市科技创新人才研究专项基金资助项目(2010RFQXN094)

摘　　要：为提高挖掘大项目集的速度,有效建立给定数据集中各项集之间的关联规则,提出了一种0-1矩阵关联规则数据挖掘算法.算法将事务数据库转化成M×N的矩阵,构造一个矩阵关联图表示频繁1-项目集中每两个项目之间的关联关系,通过遍历构造的关联矩阵有效地缩减事务数据库的大小,产生所有的频繁项集.利用模拟实验结果证明了所提算法可行性及有效性.In order to improve the speed of mining large item sets and effectively establish association rules between items of the preset database, a 0-1 matrix association rule data mining algorithm was proposed which transforms data bank to M×N matrix and constructs a matrix association graph indicating the relationship between two items in frequent 1-item set. It compresses the transaction database through the association matrix of callbacks, thus produces all the frequency sets. The feasibility and validity of this kind of algorithm was proved by simulation and experimental results.

关 键 词：关联规则 数据挖掘 0-1矩阵 支持率 置信度 

分 类 号：TP301[自动化与计算机技术—计算机系统结构]