cvplogistic (version 2.1-0)
Majorization Minimization by Coordinate Descent Algorithm for
Concave Penalized Logistic Regression for High Dimensional Data
Description
The package uses majorization minimization by coordinate
descent (MMCD) algorithm to compute the solution surface for
concave penalized logistic regression models. The SCAD and MCP
(default) are two concave penalties considered in this
implementation. The package provides three types of solutions
surfaces, one computed along the regulation parameter kappa
(default), the one along the penalty parameter lambda, and the
one computed using a hybrid algorithm. The package also
provides three tuning parameter selection methods, one based on
AIC, one based on BIC and one based on k-fold cross-validated
area under ROC curve. Other algorithms such as the adaptive
rescaling approach and local linear approximation approach are
also provided for the MCP penalty as optional choices.