Research of Spatial Outlier Detection Based on Quantitative Value of Attributive Correlation

Finding spatial outlier from its neighbor domain is a challenging problem in spatial databases. While previous work focused on the discovery of outliers when the attributive correlation values aren't quantitatively considered. We present a novel method that finds spatial outliers in spatial continuous data to overcome the disadvantages. In particular, our algorithm mines the outlier under quantitatively attributive correlation by using correlation matrix and R-tree index. We conduct experiments with the cadastre data and the results indicate that the new algorithm is very effective.