demu (version 0.3.0)

Optimal Design Emulators via Point Processes

Description

Implements the Determinantal point process (DPP) based optimal design emulator described in Pratola, Lin and Craigmile (2018) for Gaussian process regression models. See for more information and examples.

Copy Link

Version

Install

install.packages('demu')

Monthly Downloads

176

Version

0.3.0

License

AGPL-3

Maintainer

Matthew Pratola

Last Published

January 13th, 2020

Functions in demu (0.3.0)

sim.dpp.modal.nystrom.kmeans

Subsample an observational dataset using the conditional DPP design emulator with a kmeans-based Nystrom approximation.

sim.dpp.modal.fast.seq

Draw sequential samples from the conditional DPP given previously sampled points already in the design.

sim.dpp.modal.fast

Draw samples from the conditional DPP design emulator.

matern52

Calculate the correlation matrix according to the Matern model with \(\nu=5/2\).

wendland1

Calculate the correlation matrix according to the Wendland1 model.

unscalemat

Unscale a matrix back to its original ranges.

wendland2

Calculate the correlation matrix according to the Wendland2 model.

sim.dpp.modal.seq

Draw sequential samples from the conditional DPP given previously sampled points already in the design.

sim.dpp.modal.nystrom

Draw samples from the conditional DPP design emulator using grid-based Nystrom approximation.

sim.dpp.modal.np

Draw samples from the conditional DPP design emulator using a kmeans-based Nystrom approximation.

getranges

Get variable ranges from a design matrix.

demu-package

demu is an open-source R package implementing a Gaussian process optimal design emulator based on Determinantal point processes.

rhomat

Calculate the correlation matrix according to the squared exponential family of models.

scaledesign

Rescale a design matrix to the [0,1] hypercube.

makedistlist

Make list of distance matrices for calculating GP correlation matrices.

generalized.wendland

Calculate the correlation matrix according to the generalized Wendland model.

matern32

Calculate the correlation matrix according to the Matern model with \(\nu=3/2\).

remove.projections

Identify candidate points making up all marginal subprojections of an existing design.

sim.dpp.modal

Draw samples from the conditional DPP design emulator.