空间关联规则挖掘研究进展  

Research on Progress of Spatial Association Rule Mining

作　　者：张雪伍[1] 苏奋振[1] 石忆邵[2] 张丹丹[1] 

机构地区：[1]中国科学院地理科学与资源研究所资源与环境信息系统国家重点实验室,北京100101 [2]同济大学测量与国土信息工程系,上海200092

出　　处：《地理科学进展》2007年第6期119-128,共10页Progress in Geography

基　　金：国家重点基础研究发展计划资助项目(2006CB701305);国家自然基金项目(40571129)

摘　　要：随着空间数据获取技术的进步,空间数据量日益增大,已超出人们的分析能力。传统的空间数据分析方法只能进行简单的数据分析,无法满足人们获取知识的需要。空间关联规则是空间数据挖掘一个基本的任务,是从具有海量、多维、多尺度、不确定性边界等特性的空间数据中进行知识发现的重要方法。本文从基本概念、分类、挖掘过程、挖掘方法、目前研究成果等方面对其进行综述,重点阐述了空间关联规则挖掘效率的改进策略、基于不确定空间信息的挖掘方法、挖掘过程及结果的可视化、弱空间关联规则的挖掘方法等。通过对现有空间关联规则研究成果和存在问题的深入剖析,指出了其未来主要的发展方向。With the progress of spatial data technologies, the volumes of the spatial data enhance gradually, far exceeding people＇s ability to analyze it. Traditional spatial data analysis methods can only carry out simple data analysis, having no way to satisfy people＇s need of gaining knowledge. Spatial association rule mining approach, which is used to acquire underlying spatial knowledge from spatial database managing complex, multiple-dimension, large, and flexibility border space spatial data, is a fundamental mission of the spatial data mining. The authors make an annotated review of basic concepts, classification, mining process, current research achievements and so on, especially paying attention to approaches improving mining efficient, mining method based on the uncertain space information and attribute information, visualization of mining results and processes, and negative spatial association rule mining. After deeply analyzing research achievements and existing problems, the authors bring forward the future main development directions of spatial association rule mining.

关 键 词：空间数据 数据挖掘 空间关联规则 地理信息系统(GIS) 

分 类 号：TP311.13[自动化与计算机技术—计算机软件与理论]