-
Notifications
You must be signed in to change notification settings - Fork 0
/
extreme_lasso.Rd
51 lines (42 loc) · 1.19 KB
/
extreme_lasso.Rd
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
% Generated by roxygen2: do not edit by hand
% Please edit documentation in R/extreme_lasso.R
\name{extreme_lasso}
\alias{extreme_lasso}
\title{Estimation algorithm of the Extreme Lasso problem.}
\usage{
extreme_lasso(
resp,
pred_mat,
sub_power,
abstol = 0.001,
reltol = 0.001,
init_guess = NULL,
init_step = 1,
shrink = 0.5,
lambda = 1,
maxit = 100
)
}
\arguments{
\item{resp}{n by 1 response matrix.}
\item{pred_mat}{n by p predictor matrix.}
\item{sub_power}{even integer, power of Subbotin distribution.}
\item{abstol}{maximum absolute error.}
\item{reltol}{maximum relative error.}
\item{init_guess}{initialization of parameter estimates.}
\item{init_step}{initial step size scale parameter}
\item{shrink}{backtracking shrinkage parameter}
\item{lambda}{L1 regularization penalty.}
\item{maxit}{maximum iterations of gradient descent.}
}
\value{
A list with the following elements:
\itemize{
\item pars: p by 1 matrix, final parameter estimates
\item sub_power: power of Subbotin distribution.
\item iterations: number of iterations run.
}
}
\description{
Estimates for the Extreme Lasso problem; used for neighborhood selection. Uses proximal gradient descent with backtracking.
}