outerbase is an R package providing regression approaches designed for creating emulators for high-accuracy simulations. The package creates high-dimensional approximations (near-interpolators) using a unique outer product basis function structure. The advantages over other similar approaches are efficiency and robustness. It can be used to construct predictors that
-
are stable and consistent,
-
remain accurate under massive data,
-
leverage large parallel computing resources, and
-
accommodate flexible data generation.
The software is open source, which can be found on Github, and it is licensed under the MIT license. For details on installation and references to papers, see the webpage docs.
A CRAN package is on the way, but Github will be a reliable way to
try out the code. The code can be pulled directly using {devtools}.
devtools::install_github("mattplumlee/outerbase")The project was originated and is actively maintained by Matthew Plumlee ( mplumlee@northwestern.edu)