wmatrix {statnet.common}

A data matrix with row weights

Description

A representation of a numeric matrix with row weights, represented on either linear (linwmatrix) or logarithmic (logwmatrix) scale.

Usage

logwmatrix(
  data = NA,
  nrow = 1,
  ncol = 1,
  byrow = FALSE,
  dimnames = NULL,
  w = NULL
)

linwmatrix(
  data = NA,
  nrow = 1,
  ncol = 1,
  byrow = FALSE,
  dimnames = NULL,
  w = NULL
)

is.wmatrix(x)

is.logwmatrix(x)

is.linwmatrix(x)

as.linwmatrix(x, ...)

as.logwmatrix(x, ...)

## S3 method for class 'linwmatrix'
as.linwmatrix(x, ...)

## S3 method for class 'logwmatrix'
as.linwmatrix(x, ...)

## S3 method for class 'logwmatrix'
as.logwmatrix(x, ...)

## S3 method for class 'linwmatrix'
as.logwmatrix(x, ...)

## S3 method for class 'matrix'
as.linwmatrix(x, w = NULL, ...)

## S3 method for class 'matrix'
as.logwmatrix(x, w = NULL, ...)

## S3 method for class 'wmatrix'
print(x, ...)

## S3 method for class 'logwmatrix'
print(x, ...)

## S3 method for class 'linwmatrix'
print(x, ...)

## S3 method for class 'logwmatrix'
compress_rows(x, ...)

## S3 method for class 'linwmatrix'
compress_rows(x, ...)

## S3 method for class 'wmatrix'
decompress_rows(x, target.nrows = NULL, ...)

## S3 method for class 'wmatrix'
x[i, j, ..., drop = FALSE]

## S3 replacement method for class 'wmatrix'
x[i, j, ...] <- value

Arguments

data, nrow, ncol, byrow, dimnames

passed to matrix.

w

row weights on the appropriate scale.

x

an object to be coerced or tested.

...

extra arguments, currently unused.

target.nrows

the approximate number of rows the uncompressed matrix should have; if not achievable exactly while respecting proportionality, a matrix with a slightly different number of rows will be constructed.

i, j, value

rows and columns and values for extraction or replacement; as matrix.

drop

Used for consistency with the generic. Ignored, and always treated as FALSE.

Value

An object of class linwmatrix/logwmatrix and wmatrix, which is a matrix but also has an attribute w containing row weights on the linear or the natural-log-transformed scale.

Note

Note that wmatrix itself is an "abstract" class: you cannot instantiate it.

Note that at this time, wmatrix is designed as, first and foremost, as class for storing compressed data matrices, so most methods that operate on matrices may not handle the weights correctly and may even cause them to be lost.

See Also

rowweights, lrowweights, compress_rows

Examples

(m <- matrix(1:3, 2, 3, byrow=TRUE))
(m <- rbind(m, 3*m, 2*m, m))
(mlog <- as.logwmatrix(m))
(mlin <- as.linwmatrix(m))
(cmlog <- compress_rows(mlog))
(cmlin <- compress_rows(mlin))

stopifnot(all.equal(as.linwmatrix(cmlog),cmlin))

cmlog[2,] <- 1:3
(cmlog <- compress_rows(cmlog))
stopifnot(sum(rowweights(cmlog))==nrow(m))

(m3 <- matrix(c(1:3,(1:3)*2,(1:3)*3), 3, 3, byrow=TRUE))
(rowweights(m3) <- c(4, 2, 2))

stopifnot(all.equal(compress_rows(as.logwmatrix(m)), as.logwmatrix(m3),check.attributes=FALSE))
stopifnot(all.equal(rowweights(compress_rows(as.logwmatrix(m))),
                    rowweights(as.logwmatrix(m3)),check.attributes=FALSE))