EcoSta 2022: Start Registration
View Submission - EcoSta2022
A0960
Title: Cluster Detection Using Scan Statistics for High Dimensional Nonhomogeneous Poisson Process Authors:  Tung-Lung Wu - Mississippi State University (United States) [presenting]
Abstract: Cluster/anomaly detection based on scan statistics for one or two dimensional homogeneous Poisson processes have been widely stud- ied. However, there is little work on the scan statistics method for higher dimensional Poisson processes, especially the nonhomogeneous Poisson processes. In this talk, we develop an algorithm to implement the scan statistics method for high-dimensional nonhomogeneous Pois- son processes by converting the high-dimensional problem into one- dimensional problem through projections along the coordinate axes.