Bootstrap approximation with censored data under the proportional hazard model
Titel:
Bootstrap approximation with censored data under the proportional hazard model
Auteur:
Dikta, G. Ghorai, J. K.
Verschenen in:
Communications in statistics
Paginering:
Jaargang 19 (1990) nr. 2 pagina's 573-581
Jaar:
1990
Inhoud:
Let [image omitted] and [image omitted] be two independent i.i.d. sequences from a distribution function F, resp. G, on R+ = (0,∞) with F(0)=G(0)=0, where F and G are connected by the assumption of the proportional hazard model, i.e. [image omitted] . Based on the observable variables [image omitted] where [image omitted] and [image omitted] and 1A denotes the indicator function of the set A,Abdushukurov (1984) and Cheng and Lin (1984)proposed the m.l.e. of SF. An invariance principle for this estimate was shown by Cheng and Lin (1984). In this article we prove that, under an appropriate resampling scheme, the bootstrap approximation can be used to construct a confidence band for SF.