Statistical evaluation of the bootstrap estimate of spatial intensity of disease cases

João Domingos Scalon, Flávio Mattar Silva

Resumo


The aim of the present work is to investigate the performance of the method bootstrap for estimating the sampling distribution of intensity of disease cases located in different geographical areas when sampling is made through cases located in a sample of square area. Simulated data from a design factorial experiment were applied to study the effects

of type of pattern, number of quadrats, number of quadrats sampled, number of bootstrap samples and number of events in the pattern on the bootstrap confidence intervals for spatial intensity. It was observed that the bootstrap estimate of the standard error and consequently of the confidence interval for spatial intensity depends on all the main factors except for the number of bootstrap samples. The performance of the method bootstrap was dubious to estimate sampling distribution for spatial intensity when the sampling is based on a sample of quadrats. These results can help the public health planners with results about estimates of the number of disease cases in a geographical area, starting from a sample of locations of cases within specific regions.


Palavras-chave


bootstrap; configurações de padrõoes espaciais; intensidade espacial; localização de doenças

Texto completo:

PDF


Revista Ciências Exatas e Naturais - RECEN. Universidade Estadual do Centro-Oeste - UNICENTRO/PR, BRASIL.

Creative Commons License  Licenciada sob uma Licença Creative Common

ISSN 2175-5620 ON LINE; 1518-0352 IMPRESSO