Description Usage Arguments Details Value Author(s) References
Predicts value and confidence interval at new inputs using Gaussian Process Emulation.
This function should be preceded by the fitEmulator
function.
1 2 3 |
object |
A fit object of class inheriting from |
newdata |
A data matrix of input(s) at which emulation is desired (new inputs).
Must contain at least all parameters given in |
var.cov |
Optionally calculates posterior variance covariance matrix. Default is set to FALSE. For large numbers of training and prediction data, this is quite time consuming. |
sd |
Optionally calculates only the posterior standard deviation. Default is set to |
tol |
The tolerance for capping negative small values of posterior standard deviation to zero. The default is -10^-11. |
... |
Further arguments not used and an error is thrown if provided. |
Note that when using the LMC method, calculating the posterior variance is quite time-consuming.
The function returns a list containting the following components:
posterior.mean | Approximation of the outputs for the given inputs in newdata |
posterior.variance | Variance covariance matrix around this approximation |
standard.deviation | Standard Deviation of the approximation. It equals the square-root
of the diagonal of the posterior.variance
|
When the number of outputs to emulate is more than 1, method = 'separable'
, and object is of class "emulatorFit"
two extra values are returned from this function. These are
correlation.Matrix | A spatial correlation matrix. |
sigmahat | A between outputs covariance matrix. |
Originally written by Jeremy Oakley. Modified by Sajni Malde.
Oakley, J. (1999). Bayesian uncertainty analysis for complex computer codes, Ph.D. thesis, University of Sheffield.
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.