Title: Accelerated failure time model based on nonparametric Gaussian scale mixtures
Authors: Byungtae Seo - Sungkyunkwan University (Korea, South) [presenting]
Sangwook Kang - Yonsei University (Korea, South)
Abstract: When some parametric error distributions, such as normal for the accelerated failure time model, are assumed, estimators typically suffer from misspecification problems. To relax this problem, we propose a nonparametric Gaussian scale mixture model to flexibly specify the error distribution. Unlike existing non- or semi-parametric estimation methods such as rank-based procedures, the proposed method enables us to use an explicit likelihood function while avoiding potential misspecification problems. We present this model with specific estimating algorithms and some numerical examples.