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
|
% Generated by roxygen2: do not edit by hand
% Please edit documentation in R/rowAvgsPerColSet.R
\name{rowAvgsPerColSet}
\alias{rowAvgsPerColSet}
\alias{colAvgsPerRowSet}
\title{Applies a row-by-row (column-by-column) averaging function to equally-sized
subsets of matrix columns (rows)}
\usage{
rowAvgsPerColSet(X, W = NULL, rows = NULL, S, FUN = rowMeans, ...,
na.rm = NA, tFUN = FALSE)
colAvgsPerRowSet(X, W = NULL, cols = NULL, S, FUN = colMeans, ...,
na.rm = NA, tFUN = FALSE)
}
\arguments{
\item{X}{A \code{\link[base]{numeric}} NxM \code{\link[base]{matrix}}.}
\item{W}{An optional \code{\link[base]{numeric}} NxM
\code{\link[base]{matrix}} of weights.}
\item{rows}{A \code{\link[base]{vector}} indicating subset of rows to
operate over. If \code{\link[base]{NULL}}, no subsetting is done.}
\item{S}{An \code{\link[base]{integer}} KxJ \code{\link[base]{matrix}}
specifying the J subsets. Each column holds K column (row) indices for the
corresponding subset.}
\item{FUN}{The row-by-row (column-by-column) \code{\link[base]{function}}
used to average over each subset of \code{X}. This function must accept a
\code{\link[base]{numeric}} NxK (KxM) \code{\link[base]{matrix}} and the
\code{\link[base]{logical}} argument \code{na.rm}, and return a
\code{\link[base]{numeric}} \code{\link[base]{vector}} of length N (M).}
\item{...}{Additional arguments passed to then \code{FUN}
\code{\link[base]{function}}.}
\item{na.rm}{(logical) Argument passed to \code{FUN()} as
\code{na.rm = na.rm}. If \code{\link[base:logical]{NA}} (default), then
\code{na.rm = TRUE} is used if \code{X} or \code{S} holds missing values,
otherwise \code{na.rm = FALSE}.}
\item{tFUN}{If \code{\link[base:logical]{TRUE}}, the NxK (KxM)
\code{\link[base]{matrix}} passed to \code{FUN()} is transposed first.}
\item{cols}{A \code{\link[base]{vector}} indicating subset of columns to
operate over. If \code{\link[base]{NULL}}, no subsetting is done.}
}
\value{
Returns a \code{\link[base]{numeric}} JxN (MxJ)
\code{\link[base]{matrix}}, where row names equal \code{rownames(X)}
(\code{colnames(S)}) and column names \code{colnames(S)}
(\code{colnames(X)}).
}
\description{
Applies a row-by-row (column-by-column) averaging function to equally-sized
subsets of matrix columns (rows). Each subset is averaged independently of
the others.
}
\details{
If argument \code{S} is a single column vector with indices \code{1:N}, then
\code{rowAvgsPerColSet(X, S = S, FUN = rowMeans)} gives the same result as
\code{rowMeans(X)}. Analogously, for \code{colAvgsPerRowSet()}.
}
\examples{
X <- matrix(rnorm(20 * 6), nrow = 20, ncol = 6)
rownames(X) <- LETTERS[1:nrow(X)]
colnames(X) <- letters[1:ncol(X)]
print(X)
# - - - - - - - - - - - - - - - - - - - - - - - - - -
# Apply rowMeans() for 3 sets of 2 columns
# - - - - - - - - - - - - - - - - - - - - - - - - - -
nbr_of_sets <- 3
S <- matrix(1:ncol(X), ncol = nbr_of_sets)
colnames(S) <- sprintf("s\%d", 1:nbr_of_sets)
print(S)
Z <- rowAvgsPerColSet(X, S = S)
print(Z)
# Validation
Z0 <- cbind(s1 = rowMeans(X[, 1:2]),
s2 = rowMeans(X[, 3:4]),
s3 = rowMeans(X[, 5:6]))
stopifnot(identical(drop(Z), Z0))
# - - - - - - - - - - - - - - - - - - - - - - - - - -
# Apply colMeans() for 5 sets of 4 rows
# - - - - - - - - - - - - - - - - - - - - - - - - - -
nbr_of_sets <- 5
S <- matrix(1:nrow(X), ncol = nbr_of_sets)
colnames(S) <- sprintf("s\%d", 1:nbr_of_sets)
print(S)
Z <- colAvgsPerRowSet(X, S = S)
print(Z)
# Validation
Z0 <- rbind(s1 = colMeans(X[ 1:4, ]),
s2 = colMeans(X[ 5:8, ]),
s3 = colMeans(X[ 9:12, ]),
s4 = colMeans(X[13:16, ]),
s5 = colMeans(X[17:20, ]))
stopifnot(identical(drop(Z), Z0))
# - - - - - - - - - - - - - - - - - - - - - - - - - -
# When there is only one "complete" set
# - - - - - - - - - - - - - - - - - - - - - - - - - -
nbr_of_sets <- 1
S <- matrix(1:ncol(X), ncol = nbr_of_sets)
colnames(S) <- sprintf("s\%d", 1:nbr_of_sets)
print(S)
Z <- rowAvgsPerColSet(X, S = S, FUN = rowMeans)
print(Z)
Z0 <- rowMeans(X)
stopifnot(identical(drop(Z), Z0))
nbr_of_sets <- 1
S <- matrix(1:nrow(X), ncol = nbr_of_sets)
colnames(S) <- sprintf("s\%d", 1:nbr_of_sets)
print(S)
Z <- colAvgsPerRowSet(X, S = S, FUN = colMeans)
print(Z)
Z0 <- colMeans(X)
stopifnot(identical(drop(Z), Z0))
}
\author{
Henrik Bengtsson
}
\keyword{internal}
\keyword{utilities}
|
