On Mixture Regression Shrinkage and Selection via the MR-LASSO
Ronghua Luo, Hansheng Wang, Chih‐Ling Tsai · 2015
In finite mixture regression models, we generalize the application of the least absolute shrinkage and selection operator (LASSO) to obtain MR-Lasso, which incorporates both mixture and regression penalties. Because MR-Lasso jointly penalizes both regression coefficients and mixture components, it enables simul-taneous identification of significant variables and determination of important mixture components. Simulation studies indicate that MR-Lasso outperforms LASSO. Extensions to mixture non-Gaussian and mixture time series models are briefly described.