图数据库Neo4j下的POI同位特征提取  

Co-location features extraction for POI under graph database Neo4j

作　　者：张华 ZHANG Hua(Hunan Second Institute of Surveying and Mapping,Changsha,Hunan 410013,China)

机构地区：[1]湖南省第二测绘院,湖南长沙410013

出　　处：《测绘技术装备》2024年第3期88-94,共7页Geomatics Technology and Equipment

摘　　要：本文以图结构为基础,针对城市兴趣点(POI)在空间网络中的同位特征展开挖掘,突破了传统欧式空间下的各种限制。在理论方面,本文总结了图结构下的空间分布关联特征和挖掘方法,并与欧式空间进行对比,通过剖析Apriori算法,利用网络Voronoi图构建连接图结构和关联规则挖掘算法的桥梁,并采用同位模式判定方法进行图结构与欧式空间的比较分析;在实践层面,使用图数据库Neo4j存储、管理和处理图结构,充分发挥Neo4j在复杂关联数据处理方面的优势,提升了算法效率。另外,本文还利用Neo4j进行关联规则挖掘,验证了基于图结构的方法比欧式空间更适应城市POI结构特征的结论。This paper explores the co-occurrence characteristics of urban POI in spatial networks based on graph structure,it breaks through the limitations of traditional Euclidean space.In terms of theory,the spatial distribution correlation characteristics and mining methods under graph structure were summarized and compared with Euclidean space,and the co-occurrence pattern judgment method is used to compare and analyze the graph structure with the Euclidean space by analyzing the Apriori algorithm and using the network Voronoi graph to construct a bridge between the connection graph structure and the association rule mining algorithm.At the practical level,Neo4j's advantages in complex associated data processing is brought into full play,and the efficiency of algorithm is improved by using the graph database Neo4j for storage,management,and processing of graph structures.Furthermore,Neo4j is used in this paer for association rule mining,it has been verified that the graph based method is more suitable for urban POI structure features compared with Euclidean space.

关 键 词：空间数据挖掘 同位模式 网络Voronoi图 APRIORI算法 

分 类 号：P208.1[天文地球—地图制图学与地理信息工程]