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
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
|
% Copyright (C) 1998
% Berwin A. Turlach <bturlach@stats.adelaide.edu.au>
% Bill Venables <wvenable@stats.adelaide.edu.au>
% $Id: summary.l1ce.d,v 1.2 1998/09/27 02:02:02 bturlach Exp $
% --> ../COPYRIGHT for more details
\name{summary.l1ce}
\alias{summary.l1ce}
\alias{print.summary.l1ce}
\title{Summary Method for ``l1ce'' Objects (Regression with L1 Constraint)}
\description{
Returns a summary list for a regression model with an L1 constraint on
the parameters. A null value will be returned if printing is invoked.
}
\usage{
\method{summary}{l1ce}(object, correlation = TRUE,
type = c("OPT", "Tibshirani"),
gen.inverse.diag = 0, sigma = NULL, \dots)
\method{print}{summary.l1ce}(x, digits = max(3, getOption("digits") - 3), \dots)
}
\arguments{
\item{object}{fitted model of class \code{"l1ce"}.}
\item{correlation}{logical indicating if the correlation matrix for
the coefficients should be included in the summary.}
\item{type}{character string specifying
whether to use the covariance formula of Osborne, Presnell and Turlach
or the formula of Tibshirani.}
\item{gen.inverse.diag}{
if Tibshirani's formula for the covariance matrix is used, this value
is used for the diagonal elements of the generalised inverse that
appears in the formula that corresponds to parameters estimated to be
zero. The default is 0, i.e. use the Moore-Penrose inverse.
Tibshirani's code uses gen.inverse.diag=1e11.}
\item{sigma}{
the residual standard error estimate. If not provided, then it is
estimated by the deviance of the model divided by the error degrees of
freedom.}
\item{x}{an \R object of class \code{summary.l1ce}.}
\item{digits}{number of significant digits to use.}
\item{\dots}{further potential arguments passed to methods.}
}
\value{
an object of class \code{summary.l1ce} (for which there's a
\code{print} method).
It is basically a list with the following components:
\item{correlation}{the computed correlation coefficient matrix for the
coefficients in the model.}
\item{cov.unscaled}{the unscaled covariance matrix; i.e, a matrix such
that multiplying it by an estimate of the error variance produces an
estimated covariance matrix for the coefficients.
}
\item{df}{the number of degrees of freedom for the model and for residuals.}
\item{coefficients}{a matrix with three columns, containing the
coefficients, their standard errors and the corresponding t statistic.}
\item{residuals}{the model residuals. These are the weighted
residuals if weights were given in the model.}
\item{sigma}{the residual standard error estimate.}
\item{terms}{the terms object used in fitting this model.}
\item{call}{the call object used in fitting this model.}
\item{bound}{the bound used in fitting this model.}
\item{relative.bound}{the relative bound used in fitting this model
(may not be present).}
\item{Lagrangian}{the Lagrangian of the model.}
}
\details{
This function is a method for the generic function
\code{\link{summary}()} for class \code{"l1ce"}.
It can be invoked by calling \code{summary(x)} for an
object \code{x} of the appropriate class, or directly by
calling \code{summary.l1ce(x)} regardless of the
class of the object.
}
\seealso{
\code{\link{l1ce}}, \code{\link{l1ce.object}}, \code{\link{summary}}.
}
\examples{%%- or just those in ./l1ce.Rd
data(Prostate)
summary(l1ce(lpsa ~ .,Prostate))
# Produces the following output:
\dontrun{
Call:
l1ce(formula = lpsa ~ ., data = Prostate)
Residuals:
Min 1Q Median 3Q Max
-1.636 -0.4119 0.076 0.452 1.83
Coefficients:
Value Std. Error Z score Pr(>|Z|)
(Intercept) 0.7285 1.3898 0.5242 0.6002
lcavol 0.4937 0.0919 5.3711 0.0000
lweight 0.2682 0.1774 1.5115 0.1307
age 0.0000 0.0111 0.0000 1.0000
lbph 0.0093 0.0587 0.1581 0.8744
svi 0.4551 0.2525 1.8023 0.0715
lcp 0.0000 0.0947 0.0000 1.0000
gleason 0.0000 0.1685 0.0000 1.0000
pgg45 0.0002 0.0046 0.0391 0.9688
Residual standard error: 0.7595 on 88.36 degrees of freedom
The relative L1 bound was : 0.5
The absolute L1 bound was : 0.9219925
The Lagrangian for the bound is: 13.05806
Correlation of Coefficients:
(Intercept) lcavol lweight age lbph svi lcp gleason
lcavol 0.1988
lweight -0.4815 -0.2071
age -0.3938 -0.0603 -0.0974
lbph 0.3629 -0.0201 -0.5165 -0.1303
svi -0.0624 -0.2273 -0.1442 0.0635 0.0648
lcp 0.0457 -0.4153 0.0598 0.0665 0.0632 -0.3779
gleason -0.7666 -0.2009 0.1163 -0.0774 -0.0617 0.1084 -0.0243
pgg45 0.4988 0.0956 -0.0380 -0.0630 -0.1111 -0.1921 -0.2935 -0.6526
}
}
\keyword{regression}
% Converted by Sd2Rd version 1.21.
|
