摘要
空间聚类方法是分析城市规划、市场营销、社区发现等问题的一种重要手段,然而当前大多数的空间聚类方法只考虑了空间对象之间的位置关系,而忽视了它们之间的社会关系.针对这一问题,本文提出一种融合空间对象位置关系和社会关系的空间聚类方法,该方法首先构建了空间对象的地理-社会关系模型,然后提出了空间对象之间的社会关系紧密度评估方法,在提升空间聚类算法执行效率方面提出了采用邻接表索引方法进行聚类的索引结构和执行方法.实验结果表明,本文方法得到的空间聚类结果更加合理,聚类内部地点的社会联系更加紧密,并且算法具有较高的执行效率.
Spatial clustering is an important method to analyze the problems of urban planning,marketing,community discovery,and etc. However,most of the current spatial clustering methods only consider the relationship between spatial objects while ignore the social relationship between them. To solve this problem,this paper proposes a spatial clustering method,which combines the spatial object location and social relationship,to cluster the spatial objects. In this method,the spatial object＇s geographic-social relation model is constructed firstly. And then,a method for evaluating the social relationship between spatial objects is proposed. To improve the efficiency of the spatial clustering algorithm,a novel index structure which is called adjacency list index and the corresponding implementation algorithm is also presented. The experimental results demonstrate that the spatial clustering results obtained by our method are more reasonable,the social connections within the cluster are more closely,and the algorithm also has a good efficiency.
出处
《小型微型计算机系统》
CSCD
北大核心
2017年第11期2523-2528,共6页
Journal of Chinese Computer Systems
基金
国家青年科学基金项目(61401185)资助
辽宁省自然科学基金项目(20170540418)资助
辽宁省教育厅一般项目(LJYL018)资助
关键词
地理社会网络
基于密度聚类
空间索引
geo-social network
density-based clustering
spatial indexing
