L1spline: L1 spline. Function to compute the L1 spline, by use of the...

In helenecharlotte/L1splines: Robust splines

Usage

L1spline(t, y, gfun, y0 = 0, kappa = 1, lambda = 1e-04, x = Inf,
  niter = 200, sw = NULL, t0 = NULL, CI = 0, niterk = 5,
  stopiter = 1e-10, constr = -Inf, sign = 1, mp = 100, b = 0)

Arguments

t	Numeric vector of measurement times.
y	Numeric vector of values to be fitted.
gfun	Choice of Greens function (1-10).
y0	Numeric value to move y-values up or down. Per default y0 = 0.
kappa	Asymmetry parameter. Kappa is estimated if kappa=NULL is specified. Default is kappa = 1, corresponding to no skewness.
lambda	Smoothness parameter.
x	Numeric vector. If specified, spline is computed in this set of points.
niter	Number of iterations.
sw	Integer value corresponding to number of anchor points used in sigma estimation. Default is sw=NULL, corresponding to no sigma estimation.
t0	Numeric vector of anchor points used in sigma estimation. Default is t0=NULL, so that sigma estimation is only performed when sw is not NULL. If t0 is specified, this overrules sw.
CI	Option to compute confidence intervals. 0: Nothing is computed. 1: sigmahat is obtain from kappa(hat) and thetahat, and CI is computed.
niterk	Number of iterations for the kappa estimation.
stopiter	Stopping criterion.
constr	Integer value in (-Inf, 0, 1, 2). Choice of constraint. -Inf: No constraint. 0: Positivity contraint. 1: Monotonicity constraint. 2: Convexity/concavity constraint.
sign	Numeric value in (-1, 1). 1: Increasing, convexity. -1: Decreasing, concavity.
mp	Numeric value. Number of points to evaluate spline in.
b	Numeric nonnegative value, corresponding to truncation window. If b=0, the non-adapted spline is fitted. This is default.

Examples

t = generate.t(m = 25); y = generate.y()(t) + generate.noise(t, type = 2)
plot(t, y, col = "blue")
lines(t, L1spline(t, y, 9, lambda = 0.0001))

helenecharlotte/L1splines documentation built on May 17, 2019, 3:24 p.m.