derivModelMatrix: Derivatives of Model Matrix

In gek: Gradient-Enhanced Kriging

Derivatives of Model Matrix

Description

Determine the derivatives of a model matrix.

Usage

derivModelMatrix(object, ...)

## Default S3 method:
derivModelMatrix(object, data, ...)
## S3 method for class 'gekm'
derivModelMatrix(object, ...)
## S3 method for class 'gekm'
model.matrix(object, ...)

Arguments

object	an object of an appropriate class. For the default method, a formula defining the regression functions.
data	a data.frame with named columns.
...	further arguments, yet not used.

Details

derivModelMatrix makes use of the function deriv. Accordingly, the calculation of derivatives is only possible for functions that are contained in the derivatives table of deriv.

Note, in contrast to model.matrix, factors are not supported.

Value

The derivatives of the model (or design) matrix.

As in model.matrix there is an attribute "assign".

Author(s)

Carmen van Meegen

See Also

deriv for more details on supported arithmetic operators and functions.

model.matrix for construction of a design (or model) matrix.

Examples

## Several examples for the derivatives of a model matrix

dat <- data.frame(x1 = seq(-2, 2, length.out = 5))

model.matrix(~ 1, dat)
derivModelMatrix(~ 1, dat)

model.matrix(~ ., dat)
derivModelMatrix(~ ., dat)

model.matrix(~ . - 1, dat)
derivModelMatrix(~ . - 1, dat)

model.matrix(~ sin(x1) + I(x1^2), dat)
derivModelMatrix(~ sin(x1) + I(x1^2), dat)

dat <- cbind(dat, x2 = seq(1, 5, length.out = 5))

model.matrix(~ 1, dat)
derivModelMatrix(~ 1, dat)

model.matrix(~ .^2, dat)
derivModelMatrix(~ .^2, dat)

model.matrix(~ log(x2), dat)
derivModelMatrix(~ log(x2), dat)

model.matrix(~ x1:x2, dat)
derivModelMatrix(~ x1:x2, dat)

model.matrix(~ I(x1^2) * I(x2^3), dat)
derivModelMatrix(~ I(x1^2) * I(x2^3), dat)

model.matrix(~ sin(x1) + cos(x2) + atan(x1 * x2), dat)
derivModelMatrix(~ sin(x1) + cos(x2) + atan(x1 * x2), dat)

gek documentation built on April 4, 2025, 12:35 a.m.