PLEASE NOTE: THIS PACKAGE WAS ONLY EVER EXPERIMENTAL. IT ONLY EVER WORKED WITH THE "GAUSSIAN" FAMILY. IT WAS NEVER EXTENSIVELY TESTED. AND I NO LONGER ACTIVELY MAINTAIN THIS CODE. I DON'T EVEN WRITE R CODE ANYMORE.
The aim of
mcmcglmm is to provide some added functionality to the MCMCglmm
package by facilitating cross validation through both data preparation and
new-data prediction, by implementing default priors, and by providing a few
tools to evaluate parameters and model fit. You can track and contribute to
mcmcglmm at https://github.com/schaunwheeler/mcmcglmm.
Package (little) mcmcglmm
SplitDatatakes a data frame and splits it into a large subset, to be used for model training, and a small subset, to be used for cross validation. The function checks the small subset to make sure it does not contain variable options not included in the large subset, thus ensuring that cross-valiation checks will be possible (since it's not very easy to predict based on variables that weren't included in the original model).
mcmcglmmis a wrapper for the MCMCglmm() function in the
MCMCglmmpackage developed by Jerrod Hadfield (http://cran.r-project.org/web/packages/MCMCglmm/ index.html). The wrapper function allows for two variants of two defualt priors on the covariance matrices. The two defaults are InvW for an inverse- Wishart prior, which sets the degrees of freedom parameter equal to the dimension of each covariance matrix, and InvG for an inverse-Gamma prior, which sets the degrees of freedom parameter to 0.002 more than one less than the dimensions of the covariance matrix. "-pe" can be added to the call for either of these priors to use parameter-expanded priors. The function also saves the levels for each variables stored as a character or factor, which faclitates the PredictNew() function. Unlike the MCMCglmm function, mcmcglmm saves the random effects values as a default.
QuickSummaryprovides some slightly-more-than-basic measures for evaluating an 'mcmcglmm' output. Given the output, the function calculates the posterior mean, the highest posterior density intervals for a given probability (set through the "prob" option), the "type S" error (probability that the estimate actually is of the opposite sign of the posterior mean), and the "type M" error (probability that the estimate is the same sign but substantially smaller than the posterior mean - defaults to measuring the probability that the estimate is less than one half the size of the mean). The function also allows for rounding of the output for convenience - defaults to four decimal places.
PredictNewis a modified version of predict.MCMCglmm() that allows for prediction based on new data. Additionally, PredictNew() differs from predict.MCMCglmm() in that by default it marginalizes none of the random effects, whereas predict.MCMCglmm() marginalizes all random effects by default, and defaults to predicting values in post-link-function (Gaussian) scale, rather than on their original scale. This function also allows a vector to be passed to an "index" option, which will append that index to the output.