Package {impala}

Type:	Package
Title:	Bayesian Model Calibration
Version:	0.1.0
Description:	Package provides tools for modular Bayesian model calibration. these tools allow for posterior exploration with sampling methods including tempering and adaptive Markov Chain Monte Carlo (MCMC). Allows for pooled calibration or hierarchal calibration of parameters. For more information see Francom et al., 2025 <doi:10.1137/24M1644092>.
License:	MIT + file LICENSE
Encoding:	UTF-8
Imports:	cli, methods, progress, einsum,
Suggests:	fdasrvf, BASS, BayesPPR, mvBayes, coda, bayesplot, gridExtra, hexbin, knitr, rmarkdown
Config/roxygen2/version:	8.0.0
URL:	https://github.com/sandialabs/rImpala
BugReports:	https://github.com/sandialabs/rImpala/issues
VignetteBuilder:	knitr
NeedsCompilation:	no
Packaged:	2026-06-12 14:31:34 UTC; jdtuck
Author:	J. Derek Tucker [aut, cre], Sandia National Laboratories [cph, fnd]
Maintainer:	J. Derek Tucker <jdtuck@sandia.gov>
Repository:	CRAN
Date/Publication:	2026-06-19 12:00:36 UTC

Pooled Bayesian Model Calibration

Description

This function setups up calibration object

Usage

CalibSetup(bounds, constraint_func)

Arguments

Value

An object of class 'CalibSetup' which is a list with the following components:

- 'nexp': number of experiments

PCA Based Model Emulator using BASS

Description

This function setups up emulator object

Usage

ModelBassPca_func(bmod, input_names, exp_ind = NULL, s2 = "MH")

Arguments

Value

An object of class 'ModelBassPca_func'

mvBayes Emulator for Functional Outputs (can use different BASS/BPPR type emulators)

Description

ModelmvBayes Handles larger-dimensional functional responses (e.g., on large spatial fields) using various inversion tricks. We require any other covariance e.g., from discrepancy, measurement error, and basis truncation error) to be diagonal. This function setups up emulator object.

Usage

ModelmvBayes(bmod, input_names, exp_ind = NULL, s2 = "MH", h = FALSE)

Arguments

Value

An object of class 'ModelmvBayes'

mvBayes Emulator for Functional Outputs (GP)

Description

ModelmvBayes_GP Handles larger-dimensional functional responses (e.g., on large spatial fields) using various inversion tricks. We require any other covariance e.g., from discrepancy, measurement error, and basis truncation error) to be diagonal. Built for a GP emulator.

Usage

ModelmvBayes_GP(bmod, input_names, exp_ind = NULL, s2 = "MH", h = FALSE)

Arguments

Details

This function setups up emulator object.

Value

An object of class 'ModelmvBayes_GP'

mvBayes elastic Emulator for Functional Outputs (can use different BASS/BPPR type emulators)

Description

ModelmvBayes_elastic Handles larger-dimensional functional responses (e.g., on large spatial fields) using various inversion tricks. We require any other covariance e.g., from discrepancy, measurement error, and basis truncation error) to be diagonal. Contains a change on the likelihood

Usage

ModelmvBayes_elastic(
  bmod,
  bmod_warp,
  input_names,
  exp_ind = NULL,
  s2 = "MH",
  h = FALSE
)

Arguments

Details

This function setups up emulator object.

Value

An object of class 'ModelmvBayes_elastic'

mvBayes elastic Emulator for Functional Outputs (can use different BASS/BPPR type emulators)

Description

mvBayes_elastic Handles larger-dimensional functional responses (e.g., on large spatial fields) using various inversion tricks. We require any other covariance e.g., from discrepancy, measurement error, and basis truncation error) to be diagonal. Contains a change on the likelihood and built for a GP emulator. This function setups up emulator object.

Usage

ModelmvBayes_elastic_GP(
  bmod,
  bmod_warp,
  input_names,
  exp_ind = NULL,
  s2 = "MH",
  h = FALSE
)

Arguments

Value

An object of class 'ModelmvBayes_elastic_GP'

Add vector experiments

Description

This method adds vector experiments to calibration object

Usage

addVecExperiments(
  obj,
  yobs,
  model,
  sd_est,
  s2_df,
  s2_ind,
  meas_error_cor = NULL,
  theta_ind = NULL,
  D = NULL,
  discrep_tau = 1
)

Arguments

Value

An object of class 'CalibSetup'

Pool Calibration

Description

This function runs pooled Bayesian Model Calibration with adaptive MCMC, tempering, and decorrelation steps. input theta will be normalized to 0-1 and sampled from uniform priors

Usage

calibPool(setup)

Arguments

Value

a list with the following elements - theta: mcmc samples of variables - s2: mcmc samples of error variance - count: number of counts of acceptance - count_s2: number of counts of acceptance on error variance - count_decor: number of times decorrelation occurred - cov_theta_cand: final theta covariance - cov_ls2_cand: final error covariance - pred_curr: current emulator predictions - discrep_vars: discrepancy coefficients - llik: log likelihood - theta_native: mcmc samples of variables in native scale

Examples

  library(impala)
  library(BASS)
  library(mvBayes)
  library(fdasrvf)
  library(bayesplot)

  # generate functions
  f <- function(x) {
    dnorm(seq(0, 1, length.out = 99),
          sin(2 * pi * x[1] ^ 2) / 4 - x[1] / 10 + 0.5,
          0.05) * x[2]
  }

  n = 100
  nt = 99
  p = 3
  x_train = matrix(runif(n * p), n)
  e = rnorm(n * 99)
  y_train = matrix(0, n, nt)
  for (i in 1:n) {
    y_train[i, ] = f(x_train[i, ])
  }

  # generate obs ------------------------------------------------------------
  x_true = c(0.1028, 0.5930)
  ftilde_obs = f(x_true)
  gam_obs = seq(0, 1, length.out = nt)
  vv_obs = gam_to_v(gam_obs)

  tt = seq(0, 1, length.out = nt)
  out = multiple_align_functions(t(y_train), tt, ftilde_obs, 0.01)
  gam_train = out$warping_functions
  vv_train = gam_to_v(gam_train)
  ftilde_train = out$fn
  qtilde_train = out$qn

  # fit emulator ------------------------------------------------------------
  emu_ftilde = mvBayes(bass, x_train, t(ftilde_train), nBasis=2)

  emu_vv = mvBayes(bass, x_train, t(vv_train), nBasis=2)

  # impala ------------------------------------------------------------------
  input_names = c("theta0", "theta1", "theta2")
  bounds = list()
  bounds[['theta0']] = c(0, 1)
  bounds[['theta1']] = c(0, 1)
  bounds[['theta2']] = c(0, 1)

  setup = CalibSetup(bounds, cf_bounds)

  model_ftilde = ModelmvBayes(emu_ftilde, input_names)
  model_vv = ModelmvBayes(emu_vv, input_names)

  setup = addVecExperiments(setup, t(ftilde_obs), model_ftilde, 0.01, 20, rep(1, nt))
  setup = addVecExperiments(setup, t(vv_obs), model_vv, 0.01, 20, rep(1, nt))
  setup = setTemperatureLadder(setup, 1.05 ^ (0:2))
  setup = setMCMC(setup, 1000, 500, 1, 10)
  out_cal = calibPool(setup)

Compare to bounds

Description

This function compares variables to bounds

Usage

cf_bounds(x, bounds)

Arguments

Value

a vector of 'TRUE' or 'FALSE' if the values are within the bounds

evalm constructor

Description

Default constructor for evalm class

Usage

evalm(obj, ...)

Arguments

Value

An object of class 'evalm'

Set MCMC Parameters

Description

This function setups up MCMC parameters for adaptive MCMC, also includes tempering and decorrelation steps

Usage

setMCMC(
  obj,
  nmcmc,
  nburn = 0,
  thin = 1,
  decor = 100,
  start_var_theta = 1e-08,
  start_tau_theta = 0,
  start_var_ls2 = 1e-05,
  start_tau_ls2 = 0,
  start_adapt_iter = 300
)

Arguments

Value

An object of class 'CalibSetup'

Set Temperature Ladder

Description

This function setups up temperature ladder for tempering

Usage

setTemperatureLadder(obj, temperature_ladder, start_temper = 1000)

Arguments

Value

An object of class 'CalibSetup'

Transform parameters

Description

Transforms variables from the scale 0 to 1 to variable provided by bounds

Usage

tran_unif(th, bounds, names)

Arguments

Value

a list of parameters