基于Co-location模式挖掘城市空间分布特征
发布时间：2019-05-19 21:39
【摘要】：Co-location模式挖掘是在空间对象集中找到位置上频繁出现在一起的模式。例如,假设两种范畴的商业行为在空间上通常会有邻近关系,则它们很有可能会被定义为co-location模式；一些不同的生物会生长在具有邻近关系的区域内,它们很有可能具有co-location模式。co-location模式的发现在现实生活中具有非常重要的意义,比如广告商会在人群聚集的地方放置特定广告等等。 本文主要研究如何在GIS平台上实现空间数据挖掘的co-location算法。首先简要介绍了关于数据挖掘、空间数据挖掘和空间关联规则的一些基本概念,以及空间数据与传统数据的区别。其次,详细介绍了关于GIS空间数据挖掘技术,包括GIS的工作原理和GIS挖掘的一些方法,为后面的开发奠定基础。然后介绍了本文核心算法的有关概念以及算法的基本思想和执行的过程,并对算法的数据结构进行了设计。接下来,使用ArcGIS Engine组件和.NET平台的结合,开发了co-location算法的挖掘模块,通过嵌入式GIS的基本功能,为算法的执行提取空间数据库中的数据,实现了较好的操作性和可视化的效果。 最后,使用了昆明市区的地图为例,从该地图的图层中选择了三个空间特征,即教育机构、金融机构和医疗卫生机构,将它们通过所在行业的类型各自划分为几类进行了分析。使用本文开发的系统,通过设置不同的参数,挖掘不同条件下的co-location规则,发现这些规则反映的实际情况,说明该算法具有较强的实际应用意义。
[Abstract]:Co-location pattern mining is a pattern that frequently appears together in the spatial object set. For example, assuming that the two categories of business behavior usually have adjacent relationships in space, they are likely to be defined as co-location patterns. Some different organisms will grow in adjacent areas, and they are likely to have co-location patterns. The discovery of co-location patterns is of great significance in real life. For example, advertising chambers place specific advertisements in places where people gather, and so on. This paper mainly studies how to implement the co-location algorithm of spatial data mining on GIS platform. Firstly, some basic concepts of data mining, spatial data mining and spatial association rules are briefly introduced, as well as the differences between spatial data and traditional data. Secondly, the GIS spatial data mining technology, including the working principle of GIS and some methods of GIS mining, is introduced in detail, which lays the foundation for the later development. Then the related concepts of the core algorithm, the basic idea and the execution process of the algorithm are introduced, and the data structure of the algorithm is designed. Next, using the combination of ArcGIS Engine component and. Net platform, the mining module of co-location algorithm is developed. through the basic function of embedded GIS, the data in spatial database is extracted for the execution of the algorithm. The good operation and visualization effect are realized. Finally, using the map of Kunming urban area as an example, three spatial characteristics are selected from the layer of the map, namely, educational institutions, financial institutions and medical and health institutions, and they are divided into several categories through the types of their industries. Using the system developed in this paper, by setting different parameters and mining co-location rules under different conditions, the actual situation reflected by these rules is found, which shows that the algorithm has strong practical application significance.
【学位授予单位】：云南大学
【学位级别】：硕士
【学位授予年份】：2011
【分类号】：TP311.13
[Abstract]:Co-location pattern mining is a pattern that frequently appears together in the spatial object set. For example, assuming that the two categories of business behavior usually have adjacent relationships in space, they are likely to be defined as co-location patterns. Some different organisms will grow in adjacent areas, and they are likely to have co-location patterns. The discovery of co-location patterns is of great significance in real life. For example, advertising chambers place specific advertisements in places where people gather, and so on. This paper mainly studies how to implement the co-location algorithm of spatial data mining on GIS platform. Firstly, some basic concepts of data mining, spatial data mining and spatial association rules are briefly introduced, as well as the differences between spatial data and traditional data. Secondly, the GIS spatial data mining technology, including the working principle of GIS and some methods of GIS mining, is introduced in detail, which lays the foundation for the later development. Then the related concepts of the core algorithm, the basic idea and the execution process of the algorithm are introduced, and the data structure of the algorithm is designed. Next, using the combination of ArcGIS Engine component and. Net platform, the mining module of co-location algorithm is developed. through the basic function of embedded GIS, the data in spatial database is extracted for the execution of the algorithm. The good operation and visualization effect are realized. Finally, using the map of Kunming urban area as an example, three spatial characteristics are selected from the layer of the map, namely, educational institutions, financial institutions and medical and health institutions, and they are divided into several categories through the types of their industries. Using the system developed in this paper, by setting different parameters and mining co-location rules under different conditions, the actual situation reflected by these rules is found, which shows that the algorithm has strong practical application significance.
【学位授予单位】：云南大学
【学位级别】：硕士
【学位授予年份】：2011
【分类号】：TP311.13
【参考文献】
相关期刊论文 前6条
1 巨s，
本文编号：2481087
本文链接：https://www.wllwen.com/wenyilunwen/guanggaoshejilunwen/2481087.html
