摘要
研究城市公共设施的热点分布,对把握当前城市发展现状具有重要意义。针对传统算法的相关研究存在参数设定受人为因素影响较大的问题,本文采用优化的密度峰值聚类算法进行热点选取和聚类分析,在原算法的基础上,利用指数隶属函数计算局部密度并自动生成阈值,削弱了参数选取对结果的影响和反复调整参数带来的不便,结合决策图和簇中心权值进行热点的选取,提高了聚类结果的准确性。以郑州市辖区教育培训机构分布热点为例进行实验分析,结果表明,基于该算法的热点分布分析能够快速掌握设施的分布模式和聚集程度,为调整发展思路、优化设施布局提供决策支持。
Studying the hotspot distribution of urban public facilities is of great significance for grasping the current status of urban development.The related research on traditional algorithms has the problem that the parameter settings are greatly affected by human factors.This paper selects optimized density peak clustering algorithm for hotspot selection and cluster analysis.Based on the original algorithm,it calculates local density using exponential membership function and automatically generates threshold,which weakened the impact of parameter selection on the results and the inconvenience of repeatedly adjusting the parameters.By combining decision graphs and cluster center weights to select hotspots,it improves the accuracy of the clustering results.Taking the distribution of hotspots of education institutions in Zhengzhou as an example for experimental analysis,the result shows that analysis based on this algorithm can quickly grasp the distribution pattern and degree of aggregation of facilities,and provides decision support for adjusting development ideas and optimizing facilities layout.
作者
康磊
刘海砚
陈晓慧
张付兵
李元復
KANG Lei;LIU Haiyan;CHEN Xiaohui;ZHANG Fubing;LI Yuanfu(Information Engineering University,Zhengzhou 450001,China)
出处
《测绘与空间地理信息》
2021年第3期26-29,35,共5页
Geomatics & Spatial Information Technology
基金
河南省自然科学基金项目(182300410005)
国家自然科学基金项目(41801313)资助。
关键词
公共设施
聚类分析
局部密度
热点分布
public utilities
cluster analysis
local density
hotspot distribution
