stagelife: Random Number Generation and Inferences for Two-Stage Life-Testing Data

Provides random number generation, estimation, and hypothesis testing for two-stage life-testing experiments in which units begin on a baseline stress and, at a fixed or random change point, some or all surviving units move to an elevated stress ("stage life testing"), following the cumulative-exposure model of Laumen and Cramer (2019), "Stage life testing", <doi:10.1002/nav.21874>. Twelve baseline lifetime distributions are supplied (Weibull, Power Lindley, Exponentiated Lindley, Exponentiated Exponential, Marshall-Olkin Exponential, Lomax, Exponentiated Teissier, Log-Logistic, Chen, Log-normal, Gamma, and Power Muth/Power Teissier) together with three parameter-sharing cases (common, shape1, shape11). The package implements data simulation under both the fixed-proportion (typeP) and fixed-count (typeM) rules and the simple-step-stress special case, maximum likelihood estimation using a hybrid Nelder-Mead and BFGS optimisation routine, asymptotic (Wald) and non-parametric/parametric bootstrap confidence intervals (including log-scale variants that guarantee a positive interval), a stage-1-observation planning probability, Kolmogorov-Smirnov goodness-of-fit testing with a parametric-bootstrap correction for estimated parameters, likelihood-ratio and information criteria for nested/non-nested model comparison, and diagnostic plots comparing fitted and empirical distributions, including the cumulative distribution function (CDF), survival function, probability-probability (P-P), and quantile-quantile (Q-Q) plots.

install.packages("stagelife")

Try the stagelife package in your browser