顾及属性特征的城市设施空间聚类分析

A Spatial Clustering Methodof Urban Facilities Considering Attribute Characteristics

研究城市设施的聚集分布特征,是把握当前设施空间结构的重要内容,通过空间聚类进行相关分析能够深入挖掘城市设施的聚集模式。考虑到设施特征尺度非均质化的问题,提出了一种适用于城市设施空间聚类的新方法。首先,将体现设施特征尺度的属性评价值映射到数据场,用势值定量化表示场中对象的叠加影响;然后,通过构建势能决策图快速找到局部势值最大的对象作为聚类中心并完成聚类。以全国高等院校数据为例进行实验,结果表明,提出的方法由于顾及了设施对象的属性差异,聚类中心由设施的空间特征和属性特征同时作用产生,均为局部地区办学水平最高的院校,从而聚类结果现实意义更强。

Studying the clustering distribution characteristics of urban facilities is an important part of grasping the current urban spatial structure.Correlation analysis through spatial clustering can deeply explore the agglomeration mode of urban facilities.Considering the heterogeneity problem of facility feature scales,this paper proposes a new method for spatial clustering of urban facilities.First,this method maps the attribute evaluation reflecting the feature scale of the facilities to the data field,and uses the potential value to quantify the superimposed influence of the objects in the field;Then it quickly finds the object with the largest local potential value as the clustering center by constructing a potential decision graphand completes the clustering.Taking the data of universities across the country as an example,the results show that,considering the differences in the attributes of the facility objects,the cluster centers are generated by the simultaneous effects of the spatial and attribute characteristics of the facilities,and they are all universities with the highest level of operating schools in some areas,so that the clustering results have stronger practical significance.