Monotone Regression for Rows or Columns in a Matrix

Monotone (isotone) regression for rows (monoreg.rowwise) or columns (monoreg.colwise) in a matrix.

Usage

monoreg.rowwise(yM, wM)

monoreg.colwise(yM, wM)

Arguments

yM

Matrix with dependent variable for the regression. Values are assumed to be sorted.

wM

Matrix with weights for every entry in the yM matrix.

Value

Matrix with fitted values

Author

Alexander Robitzsch

The monoreg function from the fdrtool package is simply extended to handle matrix input.

Note

This function is used for fitting the ISOP model (see isop.dich).

See also

See also the monoreg function from the fdrtool package.

Examples

y <- c(22.5, 23.33, 20.83, 24.25 )
w <- c( 3,3,3,2)
# define matrix input
yM <- matrix( 0, nrow=2, ncol=4 )
wM <- yM
yM[1,] <- yM[2,] <- y
wM[1,] <- w
wM[2,] <- c(1,3,4, 3 )

# fit rowwise monotone regression
monoreg.rowwise( yM, wM )
# compare results with monoreg function from fdrtool package
if (FALSE) {
miceadds::library_install("fdrtool")
fdrtool::monoreg(x=yM[1,], w=wM[1,])$yf
fdrtool::monoreg(x=yM[2,], w=wM[2,])$yf
}