Reliability Verification, Testing, and Analysis in Engineering Design

behavior for each candidate model characterized in earlier chapters. The set of models includes the standard twoand three-parameter Weibull distributions; the inverse Weibull, exponentiated Weibull, Weibull, and inverse Weibull mixture models; Weibull and inverse Weibull competing risk models; and sectional models. The authors then use the residual sum of squares from least squares fit to the WPP of each candidate model to identify the best model. For large datasets, bootstrap and cross-validation procedures are also used as part of model selection. Three datasets are used to illustrate the procedure. The stepby-step process is explained and illustrated in a fairly clear manner, but I found this chapter lacking in some respects: