Chi-square goodness-of-fit test for randomly censored data (Q1096280)

The problem of testing goodness of fit of a parametric family \(\{F(\cdot,\theta), \theta\in\Theta\}\) of survival distributions from arbitrary right-censored data is considered. Weak convergence of the process \(\sqrt{N}[\hat F_ N(\cdot)-F(\cdot,{\hat \theta}_ N)]\) is obtained where \(\hat F_ N\) is a Kaplan-Meier estimator and \({\hat \theta}_ N\) is a maximum likelihood estimator of the parameter \(\theta\). Asymptotic distributions of the modified Pearson statistic and the generalized Pearson statistic are obtained both for nonrandom partitions and partitions with random boundaries.