Spatial Cluster Detection for Compound Poisson Count Data
- 2015-08-03 (Mon.), 10:30 AM
- 中研院-統計所 2F 交誼廳
- 茶 會:上午10:10統計所二樓交誼廳
- Prof. Hsing-Ming Chang(張欣民 教授)
- 國立成功大學統計學系
Abstract
Traditional spatial cluster detection methods are designed to analyze Bernoulli data (a case or not a case) to search for hotspots of cases. For example, one may be interested in finding the district or county (or a collection of districts or counties) in Taiwan that is the primary cluster of individuals who have Asthma. In this talk, we will introduce a spatial event-cluster detection method to analyze compound Poisson count data. For example, one may be interested in finding or county (or a collection of districts or counties) in Taiwan that is the primary cluster of hospital emergency department visits related to individuals due to Asthma. We will introduce the spatial scan method and illustrate the method by analyzing administrative health data. Keywords: spatial cluster detection, event-cluster, compound Poisson data.