A Bayesian via Laplace Approximation of Power Lomax Model with Censored Data
Ogunde A. A., Phillips S. A. Ajayi B. and Oboh I. C.
Keywords:Laplace Approximation, Markov Chain Monte Carlo Methods, Power Lomax model, Survival models.
Abstract
Power Lomax distribution is an extension of Lomax distribution that is more flexible, versatile and provides a better fit for skewed and censored data. In this work we introduce a solution to the closed form expression of the survival function of the model, which shows the model’s adaptability and flexibility for modelling real lifetime data. Alternatively, Bayesian estimation by MCMC simulation via Laplace approximation was employed. Comparison using Akaike Information Criterion (AIC) and Bayesian Information Criterion (BIC) shows that the Power Lomax model is a good choice for fitting the survival models and Information Criterion (AIC) simulations by Markov Chain Monte Carlo Methods. Findings show that this procedure and method are better options for modelling Bayesian regression and survival/reliability analysis. However, the results of the censored data have been clarified by the simulation results.