Robustness assessment of Poisson rate confidence interval from a log-linear model

Elisa Norberto Ferreira SANTOS[1]

Marcelo Angelo CIRILLO1

§    ABSTRACT: Rate estimation from two Poisson populations (alfa i) with (i = 1,2,) independents was proposed by Gart and Zweifel (1967) regarding an interval method, built form premises of a log-linear model. With this purpose, the present study aims to evaluate this method robustness, through, samples generated in the presence and absence of overdispersal effect, considering different parameter values and correction of the continuity. For that reason, a Monte Carlo simulation was used, in which, results empiricaly concerning a covering probability and interval average scope. Through the overdispersal effect, it has been concluded that the present method showed more discrepant results related to the accuracy, when compared with a favorable situation for its aplication, represented by the free samples of overdispersal effect.

§    KEY-WORDS: Monte Carlo; covering probability; overdispersal; log-linear.

 



[1] Departamento de Ciências Exatas, Universidade Federal de Lavras, Caixa Postal 3037, CEP:37200-000, Lavras, MG, Brasil. Email: norelisa@hotmail.com  / macufla@gmail.com