calcCutOff: Calculate multiple-test cut-off

In PuttickMacroevolution/motmot: Models of Trait Macroevolution on Trees

Description

Calculate the log-likelihood, AIC, or AICc cut-off necessary for type-one error to reach acceptable levels

Usage

calcCutOff(
  phy,
  n = 1000,
  mc.cores = 1,
  model,
  measure = "AICc",
  alpha.error = 0.05,
  ...
)

Arguments

phy	An object of class phylo (see ape).
n	Number of simulations
mc.cores	Number of cores for parallel processing for linux-type systems (not applicable to Windows)
model	Evolutionary model, typically "tm1", "tm2", or "timeSlice", which is used to test the empirical data
measure	Measure used to summarise the model. One of "lnL" (log-likelihood), "AIC", or "AICc"
alpha.error	Target for the desired type-one error rate for the model (default 0.05)
...	Arguments to be passed to transformPhylo.ML which should be identical to the model applied to empirical data

Value

The cut-off requred to produce an type-one error rate equal to quantile.cut.off (default = 0.05) when data are simulated under Brownian motion, and these data are analysed under the appropriate model.

Author(s)

Mark Puttick

See Also

transformPhylo.ML, transformPhylo.ll, transformPhylo, transformPhylo.MCMC

Examples

data(anolis.tree)
set.seed(393)
# calculated necessary AICc cut-off to reduce type-one error to 5% 
# for a timeSlice model with a split at 30Ma (only 5 simulations used,
# it's recommend to use 1000 for analyses)
calcCutOff(anolis.tree, n=5, model="timeSlice", splitTime=30)

PuttickMacroevolution/motmot documentation built on June 5, 2020, 7 p.m.