Bayesian Calibration of Complex Computer Codes
Assembles matrices blockwise into a block diagonal matrix
Matrix of distances from D1 to D2
tools:::Rd_package_title("calibrator")
beta1 estimator
estimator for beta2
Expectation of beta, given theta, phi and d
Covariance function for posterior distribution of z
Create new toy datasets
Function to join x.star to t.vec to give matrix D1
Augments observation points with parameters
Distance between two points
Expectation and variance with respect to theta
Posterior mean of K
Expectation of computer output
Extracts lat/long matrix and theta matrix from D2.
Expectation of z given y, beta2, phi
Expectation as per equation 10 of KOH2001
H function
Basis functions
Basis functions for D1 and D2
Toy example of hbar (section 4.2)
Is a matrix positive definite?
Very basic implementation of the Metropolis-Hastings algorithm
Apostiori probability of psi1
A postiori probability of hyperparameters
A postiori probability of hyperparameters
Functions to create or change hyperparameters
A priori probability of psi1, psi2, and theta
Reality
Stage 1,2 and 3 optimization on toy dataset
Symmetrize an upper triangular matrix
Auxiliary functions for equation 9 of the supplement
Integrals needed in KOH2001
Variance matrix for observations
Distance matrix
distance between observation points
Variance matrix for d
covariance matrix for beta
Variance matrix for beta1hat
variance matrix for beta2
Performs Bayesian calibration of computer models as per Kennedy and O'Hagan 2001. The package includes routines to find the hyperparameters and parameters; see the help page for stage1() for a worked example using the toy dataset. A tutorial is provided in the calex.Rnw vignette; and a suite of especially simple one dimensional examples appears in inst/doc/one.dim/.
Useful links