计算机科学与探索 ›› 2014, Vol. 8 ›› Issue (2): 150-160.DOI: 10.3778/j.issn.1673-9418.1306010

• 数据库技术 • 上一篇    下一篇

空间极大co-location模式挖掘研究

胡  新,王丽珍+,周丽华,温佛生   

  1. 云南大学 信息学院 计算机科学与工程系,昆明 650091
  • 出版日期:2014-02-01 发布日期:2014-01-26

Mining Spatial Maximal Co-Location Patterns

HU Xin, WANG Lizhen+, ZHOU Lihua, WEN Fosheng   

  1. Department of Computer Science and Engineering, School of Information Science and Engineering, Yunnan University, Kunming 650091, China
  • Online:2014-02-01 Published:2014-01-26

摘要: 空间co-location模式代表了一组空间特征的子集,它们的实例在空间中频繁地关联。挖掘空间co-location模式的研究已经有很多,但是针对极大co-location模式挖掘的研究非常少。提出了一种新颖的空间极大co-location模式挖掘算法。首先扫描数据集得到二阶频繁模式,然后将二阶频繁模式转换为图,再通过极大团算法求解得到空间特征极大团,最后使用二阶频繁模式的表实例验证极大团得到空间极大co-location频繁模式。实验表明,该算法能够很好地挖掘空间极大co-location频繁模式。

关键词: 空间数据挖掘, 空间极大co-location模式挖掘, 极大团

Abstract: A spatial co-location pattern is a group of spatial features whose instances are frequently located in the same region. The mining spatial co-location pattern problem had been investigated in the past, but a little for mining spatial maximal co-location patterns. This paper proposes a novel algorithm for mining spatial maximal co-location patterns. Firstly, the size2 co-location frequent patterns are generated based on the data sets, and then the size2 co-location frequent patterns are converted into a graph. Secondly, the maximal cliques in the graph are found through a maximal clique algorithm. Finally, spatial maximal co-location frequent patterns are obtained by verifying the maximal cliques based on table instances of size2 frequent patterns. The extensive experiments demonstrate that this algorithm is effective and efficient in mining spatial maximal co-location frequent patterns.

Key words: spatial data mining, spatial maximal co-location pattern mining, maximal clique