AdaptFitOS (version 0.69)

aspHetero: Estimate varying residual variance

Description

Estimates a varying residual variance on basis of an asp object. Resulting object can be plotted with simultaneous confidence bands corrected for heteroscedasticity.

Usage

aspHetero(object, xx, nknots=5, knots=NULL, basis="os", 
		 degree=c(3,2), tol=1e-8, niter=100, niter.var=250)

Value

An object of class asp with varying variances, with additional element sigmax including information on the spline of the varying variance.

Arguments

object

an asp object.

xx

the covariate.

nknots

the number of knots. Does not apply when knots are given.

knots

the knots. Does not apply if basis=="os". Otherwise, if NULL nknots equidistant knots are used.

basis

the spline basis: "os" (default), "trunc.poly" or "tps".

degree

the spline degree (and penalty order in case of B-splines). Defaults to c(3,2).

tol

tolerance for the convergence criterion. Default is 1e-8.

niter

a maximum number of iterations for residual variance function estimation, default is 100.

niter.var

a maximum number of iterations for the variance of random effects estimation within the residual variance function estimation routine, default is 250.

References

Wiesenfarth, M., Krivobokova, T., Klasen, S., Sperlich, S. (2012).
Direct Simultaneous Inference in Additive Models and its Application to Model Undernutrition. Journal of the American Statistical Association, 107(500): 1286-1296.

Examples

attach(mcycle)

y=accel
kn1 <- default.knots(times,20)
# fit model with constant residual variance
  fit= asp2(accel~f(times,basis="os",degree=3,knots=kn1,adap=FALSE),
  								niter = 20, niter.var = 200)


# fit model with varying residual variance
fith=aspHetero(fit,times,tol=1e-8)
op <- par(mfrow = c(1,3))
plot(fit);plot(fith)
#sigma() returns the fitted varying residual variance
plot(sort(times),sigma(fith)[order(times)],type="l")
par(op)