Online Program Home
My Program

Abstract Details

Activity Number: 542
Type: Contributed
Date/Time: Wednesday, August 3, 2016 : 10:30 AM to 12:20 PM
Sponsor: Biometrics Section
Abstract #320264
Title: Pseudo-Value Method for Ultra-High-Dimensional Semiparametric Models with Life-Time Data
Author(s): Tony Sit*
Companies: The Chinese University of Hong Kong
Keywords: Survival Analysis ; Semiparametric models ; Variable selection

Technology advances facilitate collection of high-dimensional covariate information including microarray, proteomic and SNP data. Challenges have frequently been encountered when scientists attempt to understand the association, if any, between the high-dimensional covariates of the subjects and the survival time, which is complicated due to censoring. In this work, we develop a new rank-based approach that enables us to tackle the variable selection problem for various semiparametric models for life-time data via the novel pseudo value method. While there has only been sure independence screening (SIS) for ultra-high dimensional data modelled by Cox proportional hazards models in literature, our methodology can handle a much broader class of semiparametric models including general transformation models and the accelerated failure time model. Numerical studies have demonstrated promising performance that is comparable to the (iterative) sure independence screening (SIS). Our method was also applied to analyse Diffuse large-B-cell lymphoma data, which discovered potential genes that can be influential.

This is a joint work with Ming Gao Gu and Yongze Xu

Authors who are presenting talks have a * after their name.

Back to the full JSM 2016 program

Copyright © American Statistical Association