R: M-Spline Basis for Polynomial Splines

mSpline {splines2}

R Documentation

M-Spline Basis for Polynomial Splines

Description

Generates the basis matrix of regular M-spline, periodic M-spline, and the corresponding integrals and derivatives.

Usage

mSpline(
  x,
  df = NULL,
  knots = NULL,
  degree = 3L,
  intercept = FALSE,
  Boundary.knots = NULL,
  periodic = FALSE,
  derivs = 0L,
  integral = FALSE,
  ...
)

Arguments

`x`	The predictor variable. Missing values are allowed and will be returned as they are.
`df`	Degree of freedom that equals to the column number of the returned matrix. One can specify `df` rather than `knots`. For M-splines, the function chooses `df - degree - as.integer(intercept)` internal knots at suitable quantiles of `x` ignoring missing values and those `x` outside of the boundary. For periodic spline based on M-spline (`periodic = TRUE`), `df - as.integer(intercept)` internal knots will be chosen instead and the number of internal knots must be greater or equal to the specified `degree - 1`. If internal knots are specified via `knots`, the specified `df` will be ignored.
`knots`	The internal breakpoints that define the splines. The default is `NULL`, which results in a basis for ordinary polynomial regression. Typical values are the mean or median for one knot, quantiles for more knots. For periodic splines (`periodic = TRUE`), the number of knots must be greater or equal to the specified `degree - 1`.
`degree`	A nonnegative integer specifying the degree of the piecewise polynomial. The default value is `3` for cubic splines. Zero degree is allowed for piecewise constant basis functions.
`intercept`	If `TRUE`, the complete basis matrix will be returned. Otherwise, the first basis will be excluded from the output.
`Boundary.knots`	Boundary points at which to anchor the splines. By default, they are the range of `x` excluding `NA`. If both `knots` and `Boundary.knots` are supplied, the basis parameters do not depend on `x`. Data can extend beyond `Boundary.knots`. For periodic splines (`periodic = TRUE`), the specified boundary knots define the cyclic period.
`periodic`	A logical value. If `TRUE`, the periodic splines will be returned instead of regular M-splines. The default value is `FALSE`.
`derivs`	A non-negative integer specifying the order of derivatives of M-splines. The default value is `0L` for M-spline basis functions.
`integral`	A logical value. If `TRUE`, the corresponding integrals of spline basis functions will be returned. The default value is `FALSE`. For periodic splines, the integral of each basis is integrated from the left boundary knot.
`...`	Optional arguments that are not used.

Details

This function contains an implementation of the close form M-spline basis based on the recursion formula given by Ramsay (1988) or periodic M-spline basis following the procedure producing periodic B-splines given in Piegl and Tiller (1997). For monotone regression, one can use I-splines (see iSpline) instead of M-splines.

Value

A numeric matrix of length(x) rows and df columns if df is specified. If knots are specified instead, the output matrix will consist of length(knots) + degree + as.integer(intercept) columns if periodic = FALSE, or length(knots) + as.integer(intercept) columns if periodic = TRUE. Attributes that correspond to the arguments specified are returned for usage of other functions in this package.

References

Ramsay, J. O. (1988). Monotone regression splines in action. Statistical science, 3(4), 425–441.

Piegl, L., & Tiller, W. (1997). The NURBS book. Springer Science \& Business Media.

Examples

library(splines2)

## Example given in the reference paper by Ramsay (1988)
x <- seq.int(0, 1, 0.01)
knots <- c(0.3, 0.5, 0.6)
msMat <- mSpline(x, knots = knots, degree = 2, intercept = TRUE)

op <- par(mar = c(2.5, 2.5, 0.2, 0.1), mgp = c(1.5, 0.5, 0))
matplot(x, msMat, type = "l", ylab = "y")
abline(v = knots, lty = 2, col = "gray")

## derivatives of M-splines
dmsMat <- mSpline(x, knots = knots, degree = 2,
                  intercept = TRUE, derivs = 1)

## or using the deriv method
dmsMat1 <- deriv(msMat)
stopifnot(all.equal(dmsMat, dmsMat1, check.attributes = FALSE))

## periodic M-spline basis
x <- seq.int(0, 3, 0.01)
knots <- c(0.3, 0.5, 0.6)
bknots <- c(0, 1)
pMat <- mSpline(x, knots = knots, degree = 3, intercept = TRUE,
                Boundary.knots = bknots, periodic = TRUE)
## integrals
iMat <- mSpline(x, knots = knots, degree = 3, intercept = TRUE,
                Boundary.knots = bknots, periodic = TRUE, integral = TRUE)
## first derivatives by "derivs = 1"
dMat1 <- mSpline(x, knots = knots, degree = 3, intercept = TRUE,
                 Boundary.knots = bknots, periodic = TRUE, derivs = 1)
## first derivatives by using the deriv() method
dMat2 <- deriv(pMat)

par(mfrow = c(2, 2))
matplot(x, pMat, type = "l", ylab = "Periodic Basis")
abline(v = seq.int(0, max(x)), lty = 2, col = "grey")
matplot(x, iMat, type = "l", ylab = "Integrals from 0")
abline(v = seq.int(0, max(x)), h = seq.int(0, max(x)), lty = 2, col = "grey")
matplot(x, dMat1, type = "l", ylab = "1st derivatives by 'derivs=1'")
abline(v = seq.int(0, max(x)), lty = 2, col = "grey")
matplot(x, dMat2, type = "l", ylab = "1st derivatives by 'deriv()'")
abline(v = seq.int(0, max(x)), lty = 2, col = "grey")
## reset to previous plotting settings
par(op)

[Package splines2 version 0.4.2 Index]