lb_convex: Fitting Log-Binomial Models by Convex Optimization

In lb: Log-Binomial Regression

Description

Fitting Log-Binomial Models by Convex Optimization

Usage

lb_convex(x, y, method = c("dense", "sparse", "weighted"), tol = 1e-08,
  ceps = 1e-07, control = list(), dry_run = FALSE, solver = "ecos")

Arguments

x	design matrix of dimension n * p.
y	vector of observations of length n.
method	a character string giving the method used to construct the optimization problem, possible
tol	tolerance for the optimizer (default is 1e-8).
ceps	epsilon subtracted from the right hand side (X β ≤q 0 - ceps). Since the inequality X β ≤q 0 is only fullfilled to a certain tolerance we have to substact epsilon to ensure X β ≤q 0.
control	a list containing additional arguments passed to ROI_solve.
dry_run	a logical controling if the problem should be solved (default dry\_run = FALSE) or the optimization problem should be returned.
solver	a character string selecting the solver to be used (default is "ecos").

Value

a vector giving the estimated coefficients.

Examples

d <- simdata_blizhosm2006(500, 1)
lb_convex(d$x, d$y)

lb documentation built on Feb. 19, 2020, 3:01 a.m.