期刊名称：Brazilian Journal of Probability and Statistics
印刷版ISSN：0103-0752
出版年度：2005
卷号：19
期号：1
页码：65-84
出版社：Brazilian Statistical Association
摘要：In this paper we introduce a Bayesian analysis for binary datain the presence of covariates and misclassifications. As a special situation indiagnostic medical testing, we obtain Bayesian inferences for the sensitivityand the specificity in the presence of covariates. We consider a situationwhere the individuals can be verified or unverified about their real diseasestatus after a test. When part or even all individuals are not verified, usuallywe have great di.culties to get classical inference results for the parametersof interest. For this situation, the introduction of latent variables gives a goodalternative to deal with missing data under the Bayesian approach, speciallyusing Markov chain monte Carlo (MCMC) methods to obtain the posteriorsummaries of interest. We illustrate the proposed methodology on three realdata sets
