基于凝聚层次聚类的co-location模式挖掘  

Co-location Patterns Mining Based on Agglomerative Hierarchical Clustering

作　　者：高世健[1] 王丽珍[1] 冯岭[1] 陈红梅[1] 

机构地区：[1]云南大学信息学院,云南昆明650091

出　　处：《广西师范大学学报（自然科学版）》2011年第2期167-173,共7页Journal of Guangxi Normal University:Natural Science Edition

基　　金：国家自然科学基金资助项目(61063008);云南省教育厅研究基金资助项目(09Y0048);云南大学科学研究基金资助项目(2009F29Q)

摘　　要：空间的co-location模式代表一组空间对象的子集,它们的实例在空间中频繁地关联,它是空间数据挖掘的重要研究方向。本文首先介绍co-location模式挖掘的基本算法,然后提出一种新的挖掘算法,算法先对空间数据进行凝聚层次聚类,在聚类结果上挖掘co-location模式,最后对这种新的算法作实验评估。Spatial co-location patterns represent the subsets of features whose instances are frequently located together in geographic space. It is an important research in the spatial data mining. Firstly,this paper introduces the basic algorithms of co-location mining. Secondly,a new algorithm is proposed,which clusters the spatial data by the agglomerative hierarchical clustering algorithm, and mines co-location patterns based on the clustering result. Finally, the experimental evaluations of the new algorithm are presented.

关 键 词：空间数据挖掘 CO-LOCATION模式 凝聚层次聚类 参与度 

分 类 号：TP392[自动化与计算机技术—计算机应用技术]