Une journée d'introduction au package RGeostats.
Course Outline
This paragraph gives a coarse descritpion of the program covered during this day.
- Introduction to RGeotats basic concepts.
- The Variogram and its extension (variogram cloud, variogram map) and the Model selection (fitting procedures)
- Estimation using an optimal linear predictor: Kriging (simple, ordinary, intrinsic) and its extension (cross-validation). Neighborhood specification.
- Multivariate Geostatistics. Bivariate statistics (covariance, linear model ...) and multivariate spatial tools (cross-variogram and cross-covariance). Linear Model of Coregionalization. Multivariate estimation technique (Co-Kriging) and its extensions (collocated Co-Kriging, External Drift concept …). Factorial Kriging Analysis.
- Simulations. Spatial law and Gaussian random function (anamorphosis). Conditional laws. Gibbs technique; Sequential Gaussian Simulation (SGS) algorithms; Turning bands algorithm; Stochastic Partial Derivative Simulations.
- Simulations of (one or several) categorical variables. Principles and inference of the parameters of the spatial characteristics. Using one or several underlying Gaussian Random Functions (PluriGaussian). Conditioning.
- New avenues: SPDE
What should be done before attending
I recall that you must bring your own laptop on which you will use the RGeostats package.
In order to have a good startup, please be sure to install:
- the R package
- the RStudio package
- the RGeostats package
For RGeostats, you can refer to the documentation provided at http://rgeostats.free.fr/forum/viewtopic.php?f=2&t=48.
As mentionned there, you can download the current version (11.2.9 minimum) for your favorite Operating System from this site.
During the course, a cloud will be used to download some documents, scripts and slide shows. Note that the last software version is also available from this cloud (in Software directory).
Instructors
- Didier RENARD
- Nicolas DESASSIS
The course will be held on Wednesday Septh 18th, from 9h30 to 17h30