Bayesian Estimation of Some Entropies and their Application in Aircraft Windshields
摘要
In this study, entropy estimation for the versatile inverse generalized gamma (IGG) distribution, a model well-suited for analyzing lifetime and highly skewed data, is explored. The IGG parameters are estimated using the maximum likelihood method to compute both the Shannon and the Rényi entropies. For a Bayesian approach, squared error, linear exponential, and precautionary loss functions are employed to derive entropy estimates. Due to the absence of closed-form posterior distributions, the Hamiltonian Monte Carlo No-U-Turn Sampler is utilized for efficient posterior sampling. The practical application and effectiveness of these estimation methods are demonstrated through the analysis of aircraft windshield failures for estimating entropies.