BioGeoBEARS (version 0.2.1)

calc_AICc_column: Calculate AICc values for a list of models

Description

A list of AICc values (second order Akaike Information Criterion) is calculated from two input lists. Lower values of AICc indicate some combination of better fit to the data and more parsimony in the model (fewer free parameters). AICc contains a correction for sample size.

Usage

calc_AICc_column(LnL_vals, nparam_vals, samplesize)

Arguments

LnL_vals

A vector of log-likelihoods (typically negative, but may not be for continuous data).

nparam_vals

A vector of the number of parameters for each model.

samplesize

A single samplesize, or a vector of the samplesizes each model. However, samplesize should always be the same for all comparisons, since maximum likelihood and AIC/AICc model-selection methods are always comparing different models on the same data, not different data on the same mode.

Value

AICc_col A data.frame column of AICc results.

Details

The two input lists are:

1. A list of data likelihoods under a variety of models. 2. A list of the number of free parameters under each model.

samplesize can be a scalar or vector; but see below.

See Burnham et al. (2002) and http://www.brianomeara.info/tutorials/aic for discussion of AIC, AICc and their uses.

References

http://phylo.wikidot.com/matzke-2013-international-biogeography-society-poster http://www.brianomeara.info/tutorials/aic

Burnham_Anderson_2002

Matzke_2012_IBS

See Also

calc_AICc_vals, calc_AIC_column

Examples

LnL_vals = c(-34.5, -20.9)
nparam_vals = c(2, 3)
calc_AICc_column(LnL_vals, nparam_vals, samplesize=20)

LnL_vals = c(-20.9, -20.9, -20.9, -20.9)
nparam_vals = c(3, 4, 5, 6)
calc_AICc_column(LnL_vals, nparam_vals, samplesize=20)