Search

In this work, a simulation study is conducted to evaluate the performance of Bayesian estimators for the log–linear exponential regression model under different levels of censoring and degrees of collinearity for two covariates. The diffuse normal, independent Student-t and multivariate Student-t distributions are considered as prior distributions and to draw from the posterior distributions, the Metropolis algorithm is implemented. Also, the results are compared with the maximum likelihood estimators in terms of the mean squared error, coverages and length of the credibility and confidence intervals. [ABSTRACT FROM PUBLISHER]

Collinearity and the lack of overlap in the data are problems that affect inference based on the logistic regression model. Simulation was used to investigate how the estimators that deal with collinearity (iterative Ridge) are affected, along with separation in the data (Firth's, and Rousseeuw and Christmann's) or both problems (Shen and Gao's). These estimators were compared considering the scaled condition number of the estimated information matrix, the bias and the mean squared error. In each one of the four scenarios studied, formed by using two levels of collinearity and two sample sizes, three degrees of overlap were considered in the data. It was found that iterative Ridge and Shen and Gao's estimators have null conditioning, as well as smaller bias and mean square error. The degree of overlap and the level of collinearity strongly affect the bias and mean square error of the maximum likelihood, Firth's and Rousseeuw and Christmann's estimators. [ABSTRACT FROM AUTHOR]