cline_summary: Summarize cline model results

In tjthurman/BAHZ: Bayesian Analysis of Hybrid Zones

Description

Summarizes the results of the cline model by providing the posterior mean and median, credible intervals, and diagnostics for the model parameters. To save computation time, this function uses the memoise package to save past results and avoid recalculations. To clear the cache of saved results, use clear.cache = T or run the clear_bahz_cache function.

Usage

cline_summary(stanfit, prob = 0.95, method = "HPDI", show.all = F,
  clear.cache = F)

Arguments

stanfit	A stanfit object holding your model results.
prob	The probability interval to return. Default is .95. Numeric, between 0 and 1.
method	The method for calculating credible intervals. Either "ET" for equal-tail probability intervals, or "HPDI" for highest posterior density intervals. Default is "HPDI".
show.all	By default, function returns summaries only for the main cline parameters. Set show.all to T to see summaries for matrix and vector parameters.
clear.cache	Clear the cache of saved results to ensure recalculation of summary? TRUE or FALSE, default FALSE.

Details

Uses the rstan summary method on class stanfit objects for calculating posterior means and medians, SEM, standard deviations, equal tail probability intervals, and diagnostics. Uses the HPDinterval function in coda to calculate HPDI intervals, the default.

Value

A summary data frame with the columns:

• param: The model parameter.

• mean: The mean of the posterior distribution.

• median: The median of the posterior distribution.

• se_mean: The standard error of the mean of the posterior distribution.

• sd: The standard deviation of the posterior distribution.

• lower: The lower limit of the credible interval. The column name will include the probability value and method used for calculating the CI.

• upper: The upper limit of the credible interval. The column name will include the probability value and method used for calculating the CI.

• n_eff: The effective number of samples from the posterior distribution.

• Rhat: The Gelman-Rubin convergence diagnostic. Should be 1, any other value is indicative of a problem during sampling.

Examples

## Not run: 
cline_summary(stanfit)

## End(Not run)

tjthurman/BAHZ documentation built on May 30, 2020, 8:28 a.m.