cmce-package: 'cmce' is an open-source R package implementing Bayesian...
In cmce: Computer Model Calibration for Deterministic and Stochastic Simulators
Description
Details
Author(s)
References
See Also
Description
cmce implements a Bayesian calibration model based on a dimension-reduction approach via empirical orthogonal functions. The model generalizes the popular SVD-based approach (e.g. Higdon et al. (2008)) to a tensor-variate decomposition, enabling the calibration of both deterministic simulators and stochastic simulators. Currently, cmce is a proof of concept implementation based entirely on R code, leaving open the possibility of future performance improvements if the codebase is moved to C at a later point. It supports both the popular additive discrepancy as well as multiplicative discrepancy. The model is fit using a Metropolis-within-Gibbs MCMC algorithm.
Details
The main model fitting function in the package is calibrate(). This function will return posterior samples of the calibration parameters as well as the predicted field process, predicted discrepancies and predicted calibrated simulator. The model currently does not support predicting off the input grid and assumes that field data and simulator outputs are observed on the same input grid.
Author(s)
Matthew T. Pratola <mpratola@stat.osu.edu> [aut, cre, cph]
References
Pratola, Matthew T. and Chrebtii, Oksana. (2018)
Bayesian Calibration of Multistate Stochastic Simulators.
Statistica Sinica, 28, 693–720.
doi: 10.5705/ss.202016.0403.
Higdon, Dave, Gattiker, James, Williams, Brian and Rightley, Maria. (2008)
Computer model calibration using high-dimensional output.
Journal of the American Statistical Association, 103, 570–583.
doi: 10.1198/016214507000000888.
See Also
cmce documentation built on May 1, 2019, 7:33 p.m.
R Package Documentation
Browse R Packages
We want your feedback!
Note that we can't provide technical support on individual packages. You should contact the package authors for that.
Description Details Author(s) References See Also
Description
cmce implements a Bayesian calibration model based on a dimension-reduction approach via empirical orthogonal functions. The model generalizes the popular SVD-based approach (e.g. Higdon et al. (2008)) to a tensor-variate decomposition, enabling the calibration of both deterministic simulators and stochastic simulators. Currently, cmce is a proof of concept implementation based entirely on R code, leaving open the possibility of future performance improvements if the codebase is moved to C at a later point. It supports both the popular additive discrepancy as well as multiplicative discrepancy. The model is fit using a Metropolis-within-Gibbs MCMC algorithm.
Details
The main model fitting function in the package is calibrate(). This function will return posterior samples of the calibration parameters as well as the predicted field process, predicted discrepancies and predicted calibrated simulator. The model currently does not support predicting off the input grid and assumes that field data and simulator outputs are observed on the same input grid.
Author(s)
Matthew T. Pratola <mpratola@stat.osu.edu> [aut, cre, cph]
References
Pratola, Matthew T. and Chrebtii, Oksana. (2018) Bayesian Calibration of Multistate Stochastic Simulators. Statistica Sinica, 28, 693–720. doi: 10.5705/ss.202016.0403.
Higdon, Dave, Gattiker, James, Williams, Brian and Rightley, Maria. (2008) Computer model calibration using high-dimensional output. Journal of the American Statistical Association, 103, 570–583. doi: 10.1198/016214507000000888.
See Also
R Package Documentation
Browse R Packages
We want your feedback!
Note that we can't provide technical support on individual packages. You should contact the package authors for that.
Embedding an R snippet on your website
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.