Variable Selection for Expanding Covariate Space and High-Dimensional Discrete Spatial Data

• Activity Number: 235 - Spatio-Temporal Theory and Methods
• Type: Contributed
• Date/Time: Monday, July 31, 2017 : 2:00 PM to 3:50 PM
• Sponsor: Section on Statistics and the Environment
• Author(s): Abdhi Sarkar* and Chae Young Lim and Tapabrata Maiti
• Companies: Michigan State University and Seoul National University and Michigan State University
• Keywords: Spatial ; Mixing conditions ; GEE

Abstract

To study real world applications of discrete data on a geographical domain we still face fundamental issues such as not being able to express the likelihood of correlated multivariate data. We circumvent this by assuming a parametric structure on the moments of a multivariate random variable and use a quasi-likelihood approach. In this talk, I propose a method that is able to select relevant variables from an expanding dimension covariate space and estimate their corresponding coefficients simultaneously. Under increasing domain asymptotics after introducing a misspecified working correlation matrix that satisfies a certain mixing condition we show that this estimator possess the" oracle" for the non-convex SCAD penalty. Several simulation results and a real data example are provided to illustrate the performance of our proposed estimator.