csgl.fit: Fits CSGL

In aplantin/CSGL: Compositional Sparse Group Lasso

Description

Function to fit CSGL. Interior function for csgl (sequence of lambdas), gic.csgl (choosing best lambda by GIC), and cv.csgl (choosing best lambda by k-fold cross-validation).

Usage

csgl.fit(Z, y, n, p, H, mu, lambda, theta, groups, thresh, erel, maxit)

Arguments

Z	Matrix of log OTU proportions
y	Outcome vector
n	Number of subjects
p	Number of OTUs
H	Matrix I_p - (1/p) 1_p 1_p'
mu	Augmented Lagrangian parameter
lambda	Regularization parameter
theta	Defines convex combination between L1 and L2 penalties (theta = 1 is L1 only, theta = 0 is L2 only)
groups	Vector indicating group membership of each OTU
thresh	Threshold for convergence
erel	Internal parameter for ADMM convergence
maxit	Maximum number of iterations

Value

Fitted coefficients beta

aplantin/CSGL documentation built on May 21, 2019, 12:08 p.m.