Skip to contents

AdaReFun: Adaptive Regional Functional Regression

Source: R/AdaReFun-package.R
AdaReFun-package.Rd

The package name AdaReFun is a mnemonic for Adaptive Regional Functional Regression. It fits smooth and locally sparse scalar-on-function linear regression models using functional MCP (fMCP), functional SCAD (fSCAD), and the adaptive regional folded-concave (ARFC) family. One ARFC specification fixes a resolution structure (adaptive single resolution, ASR, or adaptive multi resolution, AMR) and a penalty family (MCP or SCAD). arfc() fits one such specification, whereas arfc_select() fits the four ASR-MCP, ASR-SCAD, AMR-MCP, and AMR-SCAD candidates and selects exactly one with an explicitly supplied validation sample; it is not an ensemble. B-spline basis construction and user-facing methods are implemented in R; matrix-intensive fitting is implemented in C++ with RcppArmadillo.

Details

The package is developed by Bahadır Yüzbaşı and Jiguo Cao. Bahadır Yüzbaşı is the maintainer and corresponding author for the accompanying work. During peer review, the versioned package source and replication archive are available from the corresponding author. A public repository URL and archival DOI will be added only after those resources are released.

Author

Maintainer: Bahadır Yüzbaşı bahadir.yuzbasi@inonu.edu.tr (ORCID)

Authors: