Description
This paper describes a new class of space-time scan statistics designed for rapid detection of emerging disease clusters. We evaluate these methods on the task of prospective disease surveillance, and show that our methods consistently outperform the standard space-time scan statistic approach.