marcher (version 0.0-2)

getTau: Compute time scale parameters

Description

A mostly internal function that takes the "residuals" of a range-shift process and estimates $$\tau_z$$ and, if necessary, $$\tau_v$$.

Usage

getTau(Z.res, T = T, model = c("wn", "ou", "ouf")[1], tau0 = NULL,
  CI = FALSE, method = c("like", "ar")[1])

Arguments

Z.res

complex vector of isotropic Gaussian, possibly autocorrelated time series of points

T

time vector

model

one of "wn" (white noise), "ou" or "ouf" (case insensitive), denoting, respectively, no autocorrelation, position autocorrelation, or velocity and position autocorrelation. If model = NULL and method = "ar", the algorithm will select a model using AIC comparisons of the three. If the selected model iswhite noise, the function will return 0's for both parameters.

tau0

initial values of parameter estimates - a named vector: c(tau.z = tau0[1], tau.v = tau0[2])

CI

whether or not to compute the confidence intervals (temporarily only available for like method).

method

either "like" or "ar". The former refers to the likelihood method - it is most general (i.e. works with irregular sampling). The latter refers to the auto-rgressive model equivalence, which is faster but only works with regular sampling.