Stagewise Estimating Equations for Variable Selection with Longitudinal Rate Data

• Activity Number: 215 - Contributed Poster Presentations: Section on Statistical Learning and Data Science
• Type: Contributed
• Date/Time: Tuesday, August 4, 2020 : 10:00 AM to 2:00 PM
• Sponsor: Section on Statistical Learning and Data Science
• Author(s): Gregory Vaughan*
• Companies: Bentley University
• Keywords: GEE; clustered data; model selection; penalized regression; sparsity; stagewise estimation

Abstract

In many fields, there is a need for modeling rate data where response measures are reported as a percentage or proportion and are thus bounded from 0 to 1; i.e. in Psychology, many metrics fall on a bounded scale, which can be scaled to fall between $0$ and $1$, or one may want to model the proportion of subscribers successfully signed up for a service such as cable. Though beta regression has been proposed to model rate data, beta regression is not designed to work with longitudinal or clustered data where the individual observations may have some correlation within clusters and little work has been done to explore model selection in the context of modeling rate data. This work presents a model selection technique utilizing a stagewise approach that easily allows for longitudinal rate data. Numerical studies are presented to demonstrate the efficacy of the proposed approach.