BAYES ESTIMATE OF THE EQUILIBRIUM PARAMETER IN A MARKOV CHAIN: A MEDICAL APPLICATION

Carlos Alberto de Bragança PEREIRA[1]

André ROGATKO[2]

§    ABSTRACT: The results of a temporal study of 30 diabetic pregnant women are analyzed using a Markov Chain model and a Bayesian estimator for the parameter of interest, the invariant probability measure, is obtained. It is shown that cartdiotocography results of a fetus are positively dependent on the mother's glycemic levels. Then, these levels can be used as a calibration tool for the expensive cardiotocography exam.

§    KEYWORDS: Transition probability; invariant probability; high likelihood region.

 

 

 



[1]Instituto de Matemática e Estatística -USP - 05389-970 - São Paulo - SP - Brasil.

[2] Fox Chase Center - 7701 Burholme Av. - Philadelphia, PA 19111.