compare_pic_stat: Compare observed and simulated test statistics

In mwpennell/arbutus: Evaluate the adequacy of continuous trait models

Compare observed and simulated test statistics

Description

Calculates two–tailed p-values from comparing observed to simulated test statistics in order to evaluate model adequacy for phylogenetic models of continuous character evolution.

Usage

compare_pic_stat(obs, sim)

Arguments

obs	a data.frame containing the observed test statistics with the column names denoting the names of the statistics used.
sim	a data.frame containing the simulated test statistics with the column names denoting the names of the statistics used.

Details

The column names must be the same for the observed and simulated test statistics (though not necessarily in the same order). Both obs and sim are designed to be obtained from the function calculate_pic_stat.

If there is only 1 row in obs (such as the case if ML estimate of model is used), then the function will compare the observed test statistic(s) to the distribution of simulated simulated test statistics.

If there is more than 1 row in obs (such as the case if we have a posterior distribution of parameter estimates or the test statistics are calculated across many trees), then the function requires that the number of rows in obs and sim be equal. This is because the function calculates the p-values by performing pairwise comparisons. the function will automatically check that this condition is met and produce an error if it is not.

In both cases, the p-values that are returned are two–tailed p-values as the function considers the probability of the observed statistics being larger or smaller than the simulated statistics.

The distribution of simulated test statistics along with the observed test statistic using a generic plot function (see plot.arbutus) for more details. For the purposes of plotting, the observed and simulated test statistics used as arguments are returned with the p-values.

The p-values can be extracted from the output of compare_pic_stat using the function pvalue_arbutus

Value

an object of class arbutus which is a list containing the following objects:

p.values

Two–tailed p-values for all test statistics

obs

The input data frame of observed test statistics

sim

The input data frame of simulated test statistics

See Also

calculate_pic_stat, plot.arbutus, pvalue_arbutus, arbutus

Examples

data(finch)
phy <- finch$phy
dat <- finch$data[,"wingL"]
unit.tree <- make_unit_tree(phy, data=dat)

## calculate default test stats on observed data
obs <- calculate_pic_stat(unit.tree, stats=NULL)

## simulate data on unit.tree
simdat <- simulate_char_unit(unit.tree, nsim=10)

## calculate default test stats on simulated data
sim <- calculate_pic_stat(simdat, stats=NULL)

## compare simulated to observed test statistics
res <- compare_pic_stat(obs, sim)

## extract p-values
pvalue_arbutus(res)

## plot results
plot(res)

mwpennell/arbutus documentation built on Oct. 6, 2022, 10 a.m.