Online Program Home
My Program

Abstract Details

Activity Number: 249
Type: Contributed
Date/Time: Monday, August 1, 2016 : 2:00 PM to 3:50 PM
Sponsor: Section on Statistical Learning and Data Science
Abstract #318951
Title: Joint Modeling of Correlated Binary Response and Longitudinal Covariates via Random Forest Applied to Glaucoma Progression Prediction
Author(s): Juanjuan Fan* and Lucie Sharpsten and Xiaogang Su and Shaban Demirel and Richard Levine
Companies: San Diego State University and United Health Group and The University of Texas at El Paso and Devers Eye Institute and San Diego State University
Keywords: random forest ; correlated data ; longitudinal covariates ; joint modeling ; glaucoma progression

Classification tree and random forest methods are extended to correlated binary data with longitudinally collected covariates and applied to the problem of developing objective prognostic classification rules in ophthalmology research. A random effect model is used to model longitudinally collected visual field data, and the robust $Z$ statistic from generalized estimating equation (GEE) is used as the splitting statistic to measure the between-node difference while adjusting for correlation among the fellow eyes for the same patient. The proposed method is assessed through simulations conducted under a variety of model configurations and illustrated using the perimetry and psychophysics in glaucoma (PPIG) study data. In addition to producing rankings of variable importance, the random forest is also able to predict glaucoma progression with much improved sensitivity and specificity for the test sample.

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

Back to the full JSM 2016 program

Copyright © American Statistical Association