Interpolate time-varying parameters with a spline

Build a smooth time series from values specified at spline knots. This is useful when a model parameter is allowed to vary through time but is represented by a smaller set of control values to be estimated.

Usage

spline_par(par, n, knots = NULL, periodic = FALSE, period = NULL)

# S3 method for class 'spline.par'
plot(x, ...)

Arguments

par: Numeric vector of parameter values at the knots.
n: Number of output time points.
knots: Optional numeric vector giving the knot positions associated with par. If NULL, equally spaced knots are used.
periodic: Logical. Should the spline be treated as periodic?
period: Optional period of the cycle when periodic = TRUE and custom knots are supplied.
x: A spline.par object returned by spline_par().
...: Further graphical arguments passed to graphics::plot().

Value

An object of class spline.par: a list containing the interpolated values (x), the corresponding time points (time), the knot locations, the knot values, and auxiliary information used by the plotting method.

Details

The function rescales knot positions internally to the unit interval before building the spline. If periodic = TRUE, the series is extended so that the interpolation wraps around smoothly at the boundaries.

The returned object is designed to be plotted with plot() and to provide both the discrete interpolated values used in a model and a smoother curve for visual inspection.

Examples

sp = spline_par(par = c(0.2, 0.8, 0.4, 0.6), n = 12)
head(sp$x)
#> [1] 0.384375 0.643750 0.778125 0.812500 0.771875 0.681250
plot(sp)