基于空间同位模式的贝叶斯网构建  

Learning a Bayesian network from spatial co-location pattern

作　　者：周园[1] 云正富[1] 赵国庆[1] 任友理[1] 郑毅[1] 

机构地区：[1]云南国土资源职业学院

出　　处：《云南大学学报（自然科学版）》2010年第S2期5-12,共8页Journal of Yunnan University(Natural Sciences Edition)

摘　　要：空间同位挖掘方法是空间数据挖掘中挖掘空间点要素之间潜在关联规则的方法,通过对空间同位挖掘算法和贝叶斯网的分析,提出了一种从空间同位挖掘算法构建贝叶斯网的方法.该方法主要通过把空间同位挖掘算法中挖掘出的空间同位作为贝叶斯网中的结点变量,并把挖掘过程中定义的支持度作为条件概率,构建出一个基于空间同位模式的叶斯网,给出了算法和实验.利用贝叶斯网可以构建出关于空间同位模式的全概率空间.spatial co-location rule is a way to discovery potential pattern in spatial database,this paper analysed spatial co-location discovery algorithms and Bayesian networks,proposed a method to learning a Bayesian networks form spatial co-location pattern.this method mainly consider a co-location pattern as the node in Bayesian network meanwhile seem condition probability in spatial co-location pattern as the condition probability in Bayesian networks.They build a Bayesian networks bases on spatial co-location pattern and give the algorithms and experiment of learning Bayesian networks from spatial co-location pattern.

关 键 词：空间同位 贝叶斯网 空间推理 

分 类 号：TP311.13[自动化与计算机技术—计算机软件与理论]