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
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
|
\name{ltMatrices}
\alias{ltMatrices}
\alias{syMatrices}
\alias{Tcrossprod}
\alias{Crossprod}
\alias{diagonals}
\alias{Mult}
\alias{Mult.ltMatrices}
\alias{Mult.syMatrices}
\alias{solve.ltMatrices}
\alias{logdet}
\alias{as.array.ltMatrices}
\alias{as.array.syMatrices}
\alias{diagonals.ltMatrices}
\alias{diagonals.syMatrices}
\alias{diagonals.matrix}
\alias{diagonals<-}
\alias{diagonals<-.ltMatrices}
\alias{diagonals<-.syMatrices}
\alias{diagonals.integer}
\alias{is.ltMatrices}
\alias{is.syMatrices}
\alias{as.ltMatrices}
\alias{as.ltMatrices.ltMatrices}
\alias{as.ltMatrices.syMatrices}
\alias{as.syMatrices}
\alias{is.chol}
\alias{is.invchol}
\alias{as.chol}
\alias{as.invchol}
\alias{Lower_tri}
\alias{chol.syMatrices}
\alias{aperm.chol}
\alias{aperm.invchol}
\alias{aperm.ltMatrices}
\alias{aperm.syMatrices}
\alias{deperma}
\alias{adddiag}
\alias{chol2cov}
\alias{invchol2chol}
\alias{chol2invchol}
\alias{invchol2cov}
\alias{invchol2pre}
\alias{chol2pre}
\alias{Dchol}
\alias{invcholD}
\alias{chol2cor}
\alias{invchol2cor}
\alias{invchol2pc}
\alias{chol2pc}
\alias{vectrick}
\alias{standardize}
\alias{destandardize}
\title{
Multiple Lower Triangular or Symmetric Matrices
}
\description{
A class representing multiple lower triangular or symmetric matrices and some methods.
}
\usage{
ltMatrices(object, diag = FALSE, byrow = FALSE, names = TRUE)
syMatrices(object, diag = FALSE, byrow = FALSE, names = TRUE)
\S3method{as.array}{ltMatrices}(x, symmetric = FALSE, \dots)
\S3method{as.array}{syMatrices}(x, \dots)
\S3method{diagonals}{ltMatrices}(x, \dots)
\S3method{diagonals}{syMatrices}(x, \dots)
\S3method{diagonals}{matrix}(x, \dots)
\S3method{diagonals}{integer}(x, \dots)
diagonals(x) <- value
\S3method{diagonals}{ltMatrices}(x) <- value
\S3method{diagonals}{syMatrices}(x) <- value
\S3method{solve}{ltMatrices}(a, b, transpose = FALSE, \dots)
\S3method{chol}{syMatrices}(x, \dots)
\S3method{aperm}{chol}(a, perm, \dots)
\S3method{aperm}{invchol}(a, perm, \dots)
\S3method{aperm}{ltMatrices}(a, perm, \dots)
\S3method{aperm}{syMatrices}(a, perm, \dots)
deperma(chol = solve(invchol), permuted_chol = solve(permuted_invchol),
invchol, permuted_invchol, perm, score_schol)
\S3method{Mult}{ltMatrices}(x, y, transpose = FALSE, \dots)
\S3method{Mult}{syMatrices}(x, y, \dots)
Tcrossprod(x, diag_only = FALSE)
Crossprod(x, diag_only = FALSE)
logdet(x)
Lower_tri(x, diag = FALSE, byrow = attr(x, "byrow"))
is.ltMatrices(x)
is.syMatrices(x)
as.ltMatrices(x)
\S3method{as.ltMatrices}{ltMatrices}(x)
\S3method{as.ltMatrices}{syMatrices}(x)
as.syMatrices(x)
is.chol(x)
is.invchol(x)
as.chol(x)
as.invchol(x)
chol2cov(x)
invchol2chol(x)
chol2invchol(x)
invchol2cov(x)
invchol2pre(x)
chol2pre(x)
Dchol(x, D = 1 / sqrt(Tcrossprod(x, diag_only = TRUE)))
invcholD(x, D = sqrt(Tcrossprod(solve(x), diag_only = TRUE)))
chol2cor(x)
invchol2cor(x)
chol2pc(x)
invchol2pc(x)
vectrick(C, S, A, transpose = c(TRUE, TRUE))
standardize(chol, invchol)
destandardize(chol = solve(invchol), invchol, score_schol)
as.ltMatrices(x)
}
\arguments{
\item{object}{a \code{matrix} representing the lower triagular elements of
\eqn{N} lower triangular matrix, each of dimension \eqn{J \times
J}. Dimensions of \code{object} depend on \code{diag}: With
diagonal elements, \code{object} is a \eqn{J(J+1)/2 \times N} matrix,
otherwise, the number of rows is \eqn{J(J - 1) / 2}.
}
\item{diag}{logical, \code{object} contains diagonal elements if
\code{TRUE}, otherwise unit diagonal elements are assumed.
}
\item{byrow}{logical, \code{object} represents matrices in row-major order
if \code{TRUE} or, otherwise, in column-major order.
}
\item{names}{logical or character vector of length \eqn{J}.
}
\item{symmetric}{logical, object is interpreted as a symmetric matrix if
\code{TRUE}.
}
\item{diag_only}{logical, compute diagonal elements of crossproduct only
if \code{TRUE}.
}
\item{x,chol,invchol,permuted_chol,permuted_invchol}{object of class \code{ltMatrices} or \code{syMatrices} (for \code{chol}).
}
\item{value}{a matrix of diagonal elements to be assigned (of dimension \eqn{J \times N}).
}
\item{a}{object of class \code{ltMatrices}.
}
\item{perm}{a permutation of the covariance matrix corresponding to \code{a}.
}
\item{D}{a matrix (of dimension \eqn{J \times N}) of diagonal elements to be multiplied with.
}
\item{y}{matrix with \eqn{J} rows.
}
\item{b}{matrix with \eqn{J} rows.
}
\item{C}{an object of class \code{ltMatrices}.}
\item{S}{an object of class \code{ltMatrices} or a matrix with \eqn{J^2} rows
representing multiple \eqn{J x J} matrices (columns of vec operators).}
\item{A}{an object of class \code{ltMatrices}.}
\item{transpose}{a logical of length two indicating if \code{A} or \code{B} shall be transposed in \code{vectrick}.
For \code{solve}, this argument being true computes
\code{solve(t(a), b)} (in absence of a \code{t()} method for \code{ltMatrices}
objects).}
\item{score_schol}{score matrix for a standardized \code{chol} object.}
\item{\dots}{additional arguments, currently ignored.
}
}
\details{
\code{ltMatrices} interprets a matrix as lower triangular elements of
multiple lower triangular matrices. The corresponding class can be used to
store such matrices efficiently. Matrix multiplications, solutions to linear
systems, explicite inverses, and crossproducts can be computed based on such
objects. Details can be found in the \code{lmvnorm_src} package vignette.
\code{syMatrices} only store the lower triangular parts of multiple
symmetric matrices.
}
\value{
The constructor \code{ltMatrices} returns objects of class \code{ltMatrices}
with corresponding methods. The constructor \code{syMatrices} returns objects of class
\code{syMatrices} with a reduced set of methods.
}
\seealso{\code{vignette("lmvnorm_src", package = "mvtnorm")}}
\examples{
J <- 4L
N <- 2L
dm <- paste0("d", 1:J)
xm <- paste0("x", 1:N)
(C <- ltMatrices(matrix(runif(N * J * (J + 1) / 2),
ncol = N, dimnames = list(NULL, xm)),
diag = TRUE, names = dm))
## dimensions and names
dim(C)
dimnames(C)
names(C)
## subset
C[,2:3]
## multiplication
y <- matrix(runif(N * J), nrow = J)
Mult(C, y)
## solve
solve(C)
solve(C, y)
## tcrossprod
Tcrossprod(C)
## convert to matrix
as.array(solve(C[1,]))[,,1]
}
\keyword{matrix}
|
