The RGLUEANN package is now available on GitHub at http://github.com/rogiersbart/RGLUEANN. The package provides an R implementation of the coupling between general likelihood uncertainty estimation (GLUE) and artificial neural networks (ANNs), as presented in our 2012 Mathematical Geosciences paper. It is basically a probabilistic non-linear data-driven modelling tool. The package can be installed using the following R code:
The two examples provided before are available as demos now, after loading the package through library(RGLUEANN):
- demo(“RGLUEANN_training_and_prediction”) shows how to train a GLUE-ANN ensemble, and make predictions with it;
- demo(“RGLUEANN_cross-validation”) provides an example of cross-validation.
Any feedback is welcome!