paleoTS (version 0.3-1)

mle.GRW: Maximum likelihood parameter estimators

Description

These functions calculate maximum likelihood estimators for the general random walk (mle.GRW), unbiased random walk (mle.URW) and stasis (mle.Stasis) models.

Usage

mle.GRW(y)
mle.URW(y)
mle.Stasis(y)

Arguments

y

a paleoTS object

Value

• a vector of parameter estimates, either c(mstep, vstep), or c(theta, omega)

Details

These functions return maximum likelihood estimators for the general random walk mle.GRW, unbiased random walk mle.URW and stasis mle.Stasis models, but only under a restriced set of circumstances are these valid! For these estimators to be valid, the sampling error must be the same in all samples, which generally means equal sample size and variances in all samples. For the random walk models, it is also assumed that samples are evenly spaced in time. Because these assumptions usually do not hold for paleontological data, almost all users should instead use the numerical optimization functions (see opt.GRW). The main purpose for the present functions is to provide starting estimates for numerical optimization.

References

Hunt, G. 2006. Fitting and comparing models of phyletic evolution: random walks and beyond. Paleobiology32:578--601.

See Also

opt.GRW, logL.GRW

Examples

## Warning: better to use opt.RW() for real data
  y <- sim.GRW(ns=20, ms=0, vs=1)
  w <- mle.GRW(y)