Adaptively Weighted Group Lasso for Semiparametric Quantile Regression Models
Internal function: Quantile Regression with Adaptively Group Lasso wit...
Internal function: Quantile Regression with Adaptively Group Lasso wit...
Internal Function: Validate Parameters for Prediction with a `qrglasso...
Orthogonalized B-splines
Internal Function: Plot BIC Results w.r.t. lambda
Internal Function: Plot Coefficient Function
Internal Function: Plot Sequentially
Display Predicted Coefficient Functions from qrglasso
Display BIC Results from qrglasso
Predict Top-k Coefficient Functions
Adaptively Weighted Group Lasso
Implements an adaptively weighted group Lasso procedure for simultaneous variable selection and structure identification in varying coefficient quantile regression models and additive quantile regression models with ultra-high dimensional covariates. The methodology, grounded in a strong sparsity condition, establishes selection consistency under certain weight conditions. To address the challenge of tuning parameter selection in practice, a BIC-type criterion named high-dimensional information criterion (HDIC) is proposed. The Lasso procedure, guided by HDIC-determined tuning parameters, maintains selection consistency. Theoretical findings are strongly supported by simulation studies. (Toshio Honda, Ching-Kang Ing, Wei-Ying Wu, 2019, <DOI:10.3150/18-BEJ1091>).
Useful links