Usage Examples
The Python scripts and the corresponding notebooks of the examples are located in
\examples directory. In addition, a full emulation and calibration example with
Gaussian process models can be found at surmise Jupyter notebook.
In addition, for a gentle introduction of emulation and calibration using Gaussian processes, visit surmise Jupyter notebook.
Examples linked below require matplotlib as an additional plotting package to visualize results, which can be installed via
$ pip install matplotlib
Example 1 (Emulation of falling ball example)
To illustrate how the surmise’s emulator object works in practice, we
use the falling ball example.
Example 2 (Calibration of falling ball example)
To illustrate how the surmise’s calibrator object works in practice, we
use Example 1’s falling ball example.
Example 3 (Acceleration due to gravity)
This example is discussed in Chapter 8 in Gramacy, 2020.
It demonstrates how to use surmise’s emulator and calibrator objects.
Example 4 (Emulation and calibration of epidemic model)
This example illustrates the Bayesian parameter inference of Susceptible-Infected
Recovered (SIR) type epidemic model via surmise’s emulator and calibrator objects.
Although there are many model parameters, we estimated most of them based on the epidemiological studies of COVID-19, and infer only 10 influential parameters in this example.
Example 5 (Emulation of nuclear physics simulation)
This example illustrates the usage of Principal Component Stochastic Kriging model (PCSK) on simulation data from a
Viscous Anisotropic Hydrodynamic model via surmise’s emulator object.