Um estudo comparativo entre estatísticas scan espacial baseadas em modelos de regressão com variáveis respostas assimétricas.

Abstract

Spatial Scan Statistics has been developed as a geographical cluster detection analysis tool in different probabilistic models. In this dissertation, we present some spatial scan statistics based on some regression models with an asymmetric response variable. The test statistics are based on a likelihood ratio test and evaluated using Bootstrap p-value. The statistical power, sensitivity and positive predicted value of the test are examined through a Monte Carlo simulation study. Finally, an application to pulmonary tuberculosis data in state of Maranhão is made using the model that showed good results in the simulations.