findMaxLams_step: findMaxLams

In aplantin/MCL: Fits multilevel compositional lasso

Description

Finds maximum lambda value (smallest lambda such that all coefficients are zero) for both lam1 and lam2

Usage

findMaxLams_step(Z, W, y, facZ, facW, step, groups, groupIdx, mu, maxit, thresh)

Arguments

Z	OTU matrix (group level)
W	OTU matrix (within-group level)
y	Vector of outcomes
facZ	Scaling factors for Z
facW	Scaling factors for W
step	Step-size
groups	Vector indicating group membership for each feature
groupIdx	List indicating feature indices in each group
mu	Lagrange scaling parameter
maxit	Maximum number of iterations allowed
thresh	Parameter defining convergence

Value

vector with maxlam1, maxlam2

aplantin/MCL documentation built on May 25, 2019, 2:26 p.m.