Name: Llerzy Esneider Torres
Institution: Universidad del Valle, Colombia
E-mail: lletico@gmail.com
Co-authors: José Rafael Tovar
Abstract:
The main difficulties when using the Bayesian approach are obtaining information from the specialist and estimating the vector of hyperparameters of the assumed probability distribution as representative of a priori knowledge, in addition to the fact that a large part of the literature on this subject is characterized by considering a priori conjugated distributions for the parameter of interest. An indirect elicitation method is proposed to find the hyperparameters of a nonconjugated a priori distribution. The following were considered for Bernoulli trials: four a priori distributions (Beta, Kumaraswamy, Truncated Gamma and Truncated Weibull) and four scenarios for the generating process. Two necessary, but not sufficient conditions were identified to ensure the existence of a hyperparameter solution vector.
The Truncated Weibull a priori distribution performed the worst.