Mini-minimax uncertainty quantification for emulators

The MiniMiniMaxUQ R package implements the optimization and bounding procedures described in

Jeffrey Regier and Philip Stark. "Mini-Minimax Uncertainty Quantification for Emulators." In: SIAM/ASA Journal on Uncertainty Quantification 3.1 (2015), pp. 686–708.

Installation

To get the current development version from github:

# install.packages("devtools")
devtools::install_github("jeff-regier/MiniMiniMaxUQ")

Example

The code below invokes each function exported by MiniMiniMaxUQ:

library(MiniMiniMaxUQ)

X = expand.grid(1:9, 1:9) / 10
f <- function(x) sin(x[1]) + cos(x[2])
f.X = apply(X, 1, f)

K.hat = find.K.hat(X, f.X)

pointwise.uncertainty(X, f.X, K.hat, c(.55,.33))

lower.bound.computational.burden(X, f.X, K.hat, epsilon=.1)
corners.uncertainty.bound(X, f.X, K.hat)
lower.bound.max.uncertainty(X, f.X, K.hat)
upper.bound.max.uncertainty(X, f.X, K.hat)
branch.and.bound.max.uncertainty(X, f.X, K.hat)

uncertainty.confidence.bounds(X, f.X, K.hat)

Each function is further described in the man pages. Additional examples appear in tests/testthat.

License

MiniMiniMaxUQ is free software, licensed under GPLv3.