WSO-based Spatial Outlier Detection Algorithms

Based on the idea of Weighted Spatial Outlier (WSO), this study identifies the influences of spatial attributes on the calculation of spatial outlying degree, combines these influences with non-spatial attributes, and proposes two revised spatial outlier detection algorithms, Improved Z-value (IZ-value) algorithm and Weighted Difference Algorithm (WDA). The proposed algorithms are detailed in the paper and the time complexities of them are analyzed. The influencing factors of spatial outlier detection are considered from a global perspective, so that the determinations of outliers are more reasonable, and the detection result are more accurate. To verify the accuracy and effectiveness of the two proposed WSO-based spatial outlier detection algorithms, real-world dataset FMR and WNV are used for experiments, and the results are compared with classical algorithm. It is demonstrated that the proposed spatial outlier detection algorithms are feasible