Robust unbiased transformations with censoring

Damien Bousquet (1)

1 I3M - Institut de Mathématiques et de Modélisation de Montpellier

In the literature, there have been several papers about unbiased estimation in presence of censoring. In this work, we generalise the doubly robust unbiased transformation with censoring proposed by Rubin and Van der Laan and Suzukawa. We show it is the uniformly minimum variance unbiased transformation with censoring. The new more general transformation we obtain is also doubly robust and asymptotically normal.

Mots clés: asymptotic normality; bias; censoring; robustness; U statistic; variance

Domaines: Méthodologie [stat.ME]

Damien Bousquet. Robust unbiased transformations with censoring. 2013. ⟨hal-00825893⟩