Seminar - Penalized likelihood estimation of baseline hazard and regression coefficients

13 April 2012

Full title: Penalized likelihood estimation of baseline hazard and regression coefficients in proportional hazard models: Application to SAFE TBI data

Presenter: Associate Professor Jun Ma, Statistics Department, Macquarie University

In this talk we will discuss an alternative method to fit the proportional hazard model from independent survival times subject to right censoring. Different from the traditional Cox's partial likelihood method, our approach is based on maximizing directly the penalized log-likelihood function. We first adopt an approximation (such as discretization) to the baseline hazard function, and then estimate this approximated baseline hazard and the regression coefficients simultaneously. A simulation study reveals that this method can be more efficient than the partial likelihood, particularly for small to moderate samples. In addition, the new estimator is substantially less biased under informative censoring. We will also apply this method to the Traumatic Brain Injury Study (SAFE TBI study).

About the speaker: Jun Ma received the B.Sc. degree in Mathematics from Anhui University, China, in 1983, and the PhD degree from Macquarie University in statistics in 1996. He is currently an Associate Professor in Statistics Department at Macquarie University, Australia. His research interests include medical imaging, image restoration, density estimation and biostatistics.

Hosted by The George Institute for International Health

Time: 11am

Location: 341 George Street, Level 6, training room 1 (Wynyard)

Contact: Serigne Lo

Phone: 02 9657 0329

Email: 4158240f3f1f0740575c061b20471b43111122452e38231c552d