Title
Non/Semiparametric Regression Using Local Kernel Method for Correlated Survival Data
Speaker
Zhangsheng Yu, The Ohio State University
Abstract
The Local Kernel (or Polynomial) method has been studied extensively in the past twenty years. I will first present a local kernel method for correlated survival data assuming a marginal proportional hazard model. Independent and weighted working kernel estimating equations (EE) derived from local partial likelihood are studied. We show that the nonparametric estimator of the covariate functionals derivative is consistent for any arbitrary working correlation matrix and the asymptotic variance is minimized by assuming working independence. We also study a semiparametric time-dependent coefficient model for correlated survival data using a working independent profile likelihood method. The estimator is shown to be asymptotically normal. Simulation results and application study will also be presented.
Meet the speaker in Room 212 Cockins Hall at 4:30 p.m. Refreshments will be served.
