Parametric inference of inverse Maxwell distribution under type-I progressively hybrid censoring scheme

In this article, we establish estimation procedures for the inverse Maxwell distribution under Type-I progressively hybrid censored data. Classical inference employs the Newton–Raphson and Expectation–Maximization algorithms for estimation purposes. In Bayesian Inference, estimators are obtained using the Markov Chain Monte Carlo technique and Tierney–Kadane approximation under a squared error loss function. Confidence intervals are derived for estimators, assuming asymptotic normality, considering both complete and incomplete samples for maximum likelihood estimation. In Bayesian inference, confidence and highest posterior density intervals are obtained. To assess the performance of estimators from different methods, a simulation study is conducted for various censoring patterns. Two real datasets, Crack Pipe and Guinea Pig data, are analyzed to numerically illustrate the derived expressions.