A Variable Selection Method for Spatial Additive Models with Applications

• Activity Number: 379
• Type: Topic Contributed
• Date/Time: Tuesday, August 5, 2014 : 2:00 PM to 3:50 PM
• Sponsor: Section on Statistics and the Environment
• Author(s): Siddhartha Nandy*+ and Chae Young Lim and Tapabrata Maiti
• Companies: Michigan State University and Michigan State University and Michigan State University
• Keywords: Variable Selection ; Adaptive Group LASSO ; Spatial Additive Models

Abstract

We develop a variable selection technique, specifically adaptive group LASSO type of selection in additive models with spatially dependent Gaussian random error. We also consider the problem of consistently estimating non-zero components under the same model. We allow the number of components to be 'large' but the number of non-zero components is 'small' compared to the number of observations. To address both selection and estimation, we use adaptive group Lasso technique, where we first use a group Lasso method to reduce the dimension and then apply an adaptive group Lasso method to select the number of non-zero components. We validate the proposed method by simulation studies and real data examples.