rsurface.plot {stilt} | R Documentation |
To produce a response surface plot of the emulator
Description
To produce a response surface plot of the emulator
Usage
rsurface.plot(emul, parind, parvals, tind, n1, n2, zlim = NULL)
Arguments
emul |
A standard emulator object, as output, for example, by the 'emulator' function |
parind |
Vector of parameter indices (x and y) (columns of |
parvals |
A vector of parameter values to use for the rest of the
parameters (that are kept at constant values). Must be the same length as
the number of columns in |
tind |
Time index at which to predict |
n1 |
X direction grid size |
n2 |
Y direction grid size |
zlim |
Vector of z-limits for the filled.contour function. Default is the range of data plotted. |
Details
Produces a response surface plot of the emulator as a function of
selected parameters, while the rest of the parameters are fixed at
their values set by parvals
. Use n1
and n2
to
specify X and Y grid size. Optionally, plot limits can be specified
through zlim
. The code relies on the filled.contour function.
Value
None
Note
Evaluation of this function (especially for large datasets and grid sizes) might take a long time. This is because in current specification, the 'predict.emul' function predicts the entire time-series (or space transect) at once.
References
R. Olson and W. Chang (2013): Mathematical framework for a separable
Gaussian Process Emulator. Tech. Rep., available from
www.scrimhub.org/resources/stilt/Olson_and_Chang_2013_Stilt_Emulator_Technical_Report.pdf.
See Also
Examples
# Plot the SICOPOLIS ice mass loss in year 2500 as a function of Snow
# PDD Factor and Ice PDD Factor, at mid-range values of other parameters
data(emul.Sicopolis)
## Not run:
rsurface.plot(emul=emul.Sicopolis, parind=c(4,5), parvals=c(3, 10, 50, NA, NA),
tind=661, n1=5, n2=5)
## End(Not run)