Algorithm

Fabs: the Forward and Backward Stagewise (Fabs) algorithm.

Maintainer

Xingjie Shi xingjieshi@njue.edu.cn

Publication

Shi, X., Huang, J., Huang, Y., & Ma, S. (2017). A Forward and Backward Stagewise Algorithm for Nonconvex Loss Functions and Convex Penalties. Manuscript.

Usage

1. Demo a) examples.r provides examples of Fabs on calculating solution paths under the ordinary least square loss and smoothed partial rank loss.

2. Main functions: a) fabs.r --- Fabs algorithm for adaptive lasso under the ordinary least square loss and smoothed partial rank loss b) fabsBrdige.r --- Fabs algorithm for Bridge penalty under the ordinary least square loss and smoothed partial rank loss

3. Functions used in fabs.r and fabsBrdige.r: a) standardize.r --- standardize design matrix b) loss.r --- loss functions c) penalty.r --- penalty functions d) derivative.r --- derivatives of loss functions e) spr.r --- interface functions for calculate smoothed partial rank loss and it derivatives