Spatiotemporal Surveillance Using Biaxially Weighted Scan Statistic

Abstract

This paper focuses on signaling the clustering of anomaly with mean shifts. The shift magnitude, location, coverage size, and change point time of clustering are often unknown, which makes the monitoring process challenging. Spatiotemporal scan statistics are effective to detect the existence of clustering. However, none of the existing method is optimal when several parameters about clustering are unspecified. The spatial structure is informative but partially used in the calculation of the scan statistics. This paper proposes a spatiotemporal scan statistic that exponentially weights data according to their time of occurrence and distance to a reference point in space. The results show that the proposed method could be more sensitive than the existing methods depending on the smoothing parameter, size of scan window and other factors. An example of detecting the counties with high incidence of male thyroid cancer is provided to illustrate the effectiveness of the method.

Key Words:

• Clustering
• EWMA
• scan window
• spatiotemporal statistics.

Additional information

Notes on contributors

Chen-ju Lin

Chen-ju Lin is an Associate Professor in the Department of Industrial Engineering and Management at Yuan Ze University, Taiwan. She received her B.S. degree in Industrial Engineering and Management from the National Chiao Tung University in 2003, and M.S. and Ph.D. degrees in Industrial and Systems Engineering from the Georgia Institute of Technology in 2004 and 2007, respectively. Her research interests include multiple comparison techniques, quality control, and spatiotemporal statistics.

Yen-Ting Chen

Yen-Ting Chen received his master degree in Industrial Engineering and Management from Yuan Ze University, Taiwan.