Simultaneous Variable Selection
Berwin A. Turlach, William Norman Venables, Stephen J. Wright · Technometrics · 2005
We propose a new method for selecting a common subset of explanatory variables where the aim is to model several response variables. The idea is a natural extension of the LASSO technique proposed by Tibshirani (1996) and is based on the (joint) residual sum of squares while constraining the parameter estimates to lie within a suitable polyhedral region. The properties of the resulting convex programming problem are analyzed for the special case of an orthonormal design. For the general case, we develop an efficient interior point algorithm. The method is illustrated on a dataset with infrared spectrometry measurements on 14 qualitatively different but correlated responses using 770 wavelengths. The aim is to select a subset of the wavelengths suitable for use as predictors for as many of the responses as possible.
Cited by 386 works. doi:10.1198/004017005000000139
Authors
References
- Applied Regression Analysis
- Generalized Linear Models
- Regression Shrinkage and Selection Via the Lasso
- Generalized Additive Models.
- Model Selection and Multimodel Inference: A Practical Information-Theoretic Approach
- Atomic Decomposition by Basis Pursuit
- Least angle regression
- Optimization and Nonsmooth Analysis
- Regularization and Variable Selection Via the Elastic Net
- Generalized Linear Models (2nd ed.).