Variable Selection for Penalized Threshold Regression

• Activity Number: 553
• Type: Topic Contributed
• Date/Time: Wednesday, August 1, 2012 : 2:00 PM to 3:50 PM
• Sponsor: ENAR
• Author(s): Xin He*+ and Mei-Ling Ting Lee
• Companies: University of Maryland and University of Maryland
• Keywords: Generalized cross-validation ; Penalized likelihood ; Survival analysis ; Threshold regression ; Wiener process ; Variable selection

Abstract

As an alternative approach to the Cox proportional hazards model, threshold regression (TR) is a relatively new methodology that does not require a proportional hazard assumption to analyzing time-to-event data. In this article, penalized likelihood approaches are proposed to handle the variable selection problem in the context of threshold regression analysis. The proposed methods simultaneously select significant variables and estimate unknown regression coefficients. An algorithm is presented for this process. Simulation studies are conducted for assessing the performance of the proposed approach, and the methodology is applied to a motivating study of osteoporotic fractures.