挖掘空间拓扑关联的有效算法  

Efficient algorithm of mining spatial topology association

作　　者：方刚[1] 魏祖宽[2] 刘雨露[1] 唐曙光[3] 

机构地区：[1]重庆三峡学院数学与计算机科学学院,重庆404000 [2]电子科技大学计算机科学与工程学院,四川成都610054 [3]西藏自治区科技信息研究所,西藏拉萨850001

出　　处：《计算机工程与设计》2010年第6期1267-1270,共4页Computer Engineering and Design

基　　金：国家自然科学基金项目(40761018)

摘　　要：在基于空间事务的横向关联规则挖掘中,为了能够在海量数据中有效地提取空间拓扑关联规则,提出一种挖掘空间拓扑关联的有效算法,其适合挖掘多层横向空间关联规则。该算法用二进制数存储空间拓扑关系,使空间事务和数字建立对应关系,用数字递增的方法产生候选频繁项。在计算支持数时,算法在用逻辑运算的同时还利用数字特性减少扫描的空间事务数,大大地提高了效率。实验结果表明,在提取多层空间拓扑关联规则时,其比现有的算法更快速更有效。In transverse association rules mining based on spatial transaction, to inefficiently extracte spatial topology association rules in large data, an efficient algorithm of mining spatial topology association is presented, which is suitable for mining multilayer transverse spatial association rules.The algorithm uses binary number to save spatial topology association and founds a relationship between spatial transaction and number, and uses the way of number ascending to generate candidate frequent itemsets.Logic operation is used to compute support of candidate, and number character is used to reduce the number of scanned spatial transaction, which efficiently improve efficiency.The algorithm is used to extract multilayer spatial topology association rules, and this experiment indicates that the efficiency is faster and more efficient than presented algorithms.

关 键 词：空间数据挖掘 拓扑关联规则 二进制 递增搜索 多层横向挖掘 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]