Abstract: Concerning the quantitative analysis of attribute correlation in spatial outlier detection, a kind of spatial outlier detection based on quantitative value of attributive correlation is introduced. Values of attributive correlation are computed according to quantitative analysis, it uses the matrix of attributive correlation and R-tree dynamic index structure to search spatial statistics. Some experiments are done on the cadastre data. The range of application and performance of this approach are analyzed, and it can get good results under the multi-attributive correlation.

Key words: Spatial outlier, Spatial data mining, R-tree, Quantitative analysis

摘要： 针对空间离群数据(spatial outlier detection)中缺乏对多属性相关定量分析的特点，进行了基于定量分析多属性相关的空间离群数据研究。在空间统计学的基础上，采用定量相关分析对属性之间的关系进行度量，引用了多属性相关矩阵，采用R-tree动态索引结构来搜索空间离群点，有效地解决了这一问题。在多非空间属性中，准确地发现空间离群点，通过对杭州地籍数据的分析，验证了方法的正确性和有效性。

关键词: 空间离群点, 空间数据挖掘, R-tree, 定量分析

CLC Number: 

• TP301.6