On Selection of Spatial Linear Models for Lattice Data

• DOI: 10.1111/j.1467-9868.2010.00739.x
• Publication Date: 2010-05-23T10:00:06Z
• Authors: Jun Zhu; Hsin-Cheng Huang; Perla E. Reyes
• Keywords: Lasso; Neighbourhood (mathematics)

Summary Spatial linear models are popular for the analysis of data on a spatial lattice, but statistical techniques selection covariates and neighbourhood structure limited. Here we develop new methodology simultaneous model parameter estimation via penalized maximum likelihood under adaptive lasso. A computationally efficient algorithm is devised obtaining approximate estimates. Asymptotic properties estimates their approximations established. simulation study shows that method proposed has sound finite sample and, illustration, analyse an ecological set in western Canada.