Monte Carlo Simulation for Comparing The Parameters Inverse Gaussian Distributions Estimated by Bayesian Estimation Using Gamma And Weibull Prior Distribution

Abstract

This aim of this study was to compare the absolute bias (|Bias|) and the Mean Square Error (MSE) of the parameters (μ, β) in Inverse Gaussian distributions estimated by Bayesian using Gamma and Weibull prior distributions. The simulation was computed by Monte Carlo where the shape parameter (μ) was equal to 1, 5, 10, 50, 100 the scale parameter (β) was equal to 1, 5, 10, 50, 100 and the sample size (n) was equal to 50, 100, 200, 400, 1000 repeated for 10,000 times. The result showed that in the case of the sample sizes was 50 and 1000, shape parameter (μ) in Inverse Gaussian distributions using Weibull prior distribution was more efficient than using Gamma prior distribution. For the scale parameter in Inverse Gaussian distributions, using Weibull prior distribution was more efficient than using Gamma prior distribution in the case of these sample sizes; 100, 200, 400 and 1000.