Maximum Likelihood Estimation in Semiparametric Transformation Models with Length-Biased Data

• Activity Number: 315
• Type: Contributed
• Date/Time: Tuesday, August 11, 2015 : 8:30 AM to 10:20 AM
• Sponsor: Section on Statistics in Epidemiology
• Author(s): Yu-Jen Cheng*
• Companies: National Tsing Hua University
• Keywords: survival analysis

Abstract

In this article, we consider nonparametric likelihood estimation in semiparametric transformation models for survival data subject to length-biased sampling. Under length-biased sampling, the survival times are left-truncated and the truncation time follows a uniform distribution. In contrast to the conventional inverse weighting or bias-adjusted risk set approaches, the proposed full likelihood approach can incorporate the information about truncation time distribution and allow the distribution of censoring time to depend on covariates. Moreover, from our extensive simulation studies, the proposed approach performs well and is more efficient than the existing methods. Using modern empirical process theory, the proposed nonparametric likelihood estimator is shown to be consistent and asymptotically normally distributed. We apply the proposed approach to a real data.