设施POI的局部空间同位模式挖掘及范围界定

Regional Co-location Pattern Mining and Scoping from Urban Facility POI

不同类型城市基础设施之间具有较强的空间相关性,旨在发现该类知识的空间同位模式挖掘(Spatial Co-location Pattern Mining)可为商业布局、城市规划和区域管理等应用提供重要的决策支持。然而,传统的同位模式分析方法往往基于全局统计指标,容易忽略空间关系在局部地理范围内的强显著性。该文针对该缺陷,给出了局部普遍性度量指标,并基于地理学第一定律的区位影响构建邻近关系影响域,进而界定同位特征显著的局部范围。通过深圳市POI点同位模式挖掘实验,分析了POI基础设施在城市空间分布上的关联模式与依赖关系,评价结果表明,该方法较其他挖掘方法具有更高的可靠性。

Different types of urban facilities usually appear together, which have strong spanal associanon. The spanal co-loca- tion pattern mining, which aims to discover this type of knowledge, can be utilized to support many applications, such as com- mercial layout, urban planning and regional management. However, traditional co-location pattern mining usually adopts the global statistical index, which tends to neglect the strong significance of spatial relationship in regional geographic areas. Consid- ering this problem, a regional prevalence index is proposed in this paper, and the impact area of neighborhood is constructed based on the regional impact of the First Law of Geography, and thereby significant areas of co-location patterns can be delimita- ted. The actual data experiment for mining the POIs＇ co-location patterns is used to analyze spatial association characteristics of the POI infrastructure in Shenzhen City. The assess result indicates that this approach is more reliable than other methods.