Quantcast

Robust Wald-Type Tests under Random Censoring

Research paper by Abhik Ghosh, Ayanendranath Basu, Leandro Pardo

Indexed on: 31 Aug '17Published on: 31 Aug '17Published in: arXiv - Statistics - Methodology



Abstract

Randomly censored survival data are frequently encountered in applied sciences including biomedical and reliability applications. We propose Wald-type tests for testing parametric statistical hypothesis, both simple as well as composite, for randomly censored data using the M-estimators under a fully parametric set-up. We propose a consistent estimator of asymptotic variance of the M-estimators based on sample data without any assumption on the form of the censoring scheme. General asymptotic and robustness properties of the proposed Wald-type tests are developed. Their advantages and usefulness are demonstrated in detail for Wald-type tests based on a particular M estimator, namely the minimum density power divergence estimator.