CalculateModelAverages: Calculate Model Averages

In bomeara/phrapl: Phylogeography using Approximate Likelihood

Description

This function outputs model averaged parameter estimates for a set of models.

Usage

CalculateModelAverages(totalData, averageAcrossAllModels = TRUE, parmStartCol = 9, 
keep.na = FALSE)

Arguments

totalData	Path to input file (output of ConcatenateResults)
averageAcrossAllModels	If TRUE, each parameter is model averaged across all models in the dataset, even if that model does not include the given parameter. In these cases where a parameter is absent (i.e., the value is NA) this method simply assumes that the value for that parameter is zero. If FALSE, each parameter is model averaged by only considering those models that include the relevant parameter. The former method may be most appropriate for migration, as models that exclude this parameter are effectively setting migration to zero. One should be careful interpreting model averaged coalescence times (t), however. Keep in mind that when a model excluding a particular coalescent event recieves high support, this can be subject to different interpretations. For example, this could signal that these two populations are so similar that coalescence time between them is effectively zero. However, it could also signal that these two populations are very distinct, but that they coalesce with other populations prior to coalescing with each other.
parmStartCol	This gives the column number in totalData in which the parameter values begin. If using default PHRAPL output, this should be column 9.
keep.na	When TRUE, columns that only contain only NAs are retained in the output.

See Also

ConcatenateResults

Examples

#totalData<-load("/path/totalData.txt")
#modelAverages<-CalculateModelAverages(totalData)

bomeara/phrapl documentation built on Feb. 8, 2021, 6:45 p.m.