ExtremalDependence_Unreplicated

This project is the work done to model extremal dependence in unreplicated soil data. It considerds the comparison between an asymptotic, max-stable model (MGPD), and a subasymptotic model of flexible tail dependence (Exponential Factor Copula). The scripts where originally sourced from Castro-Camilo et al (2020) and Kiriliouk et al (2019) for the Exponential Factor Copula and Multivariate GPD respectively.

Script descriptions

Exponential Factor Copula Scripts

1. CensoredLikelihood_nonspatial.R - centralized likelihood function. Calls other likelihood functions.
2. logCopula_nonspatial.R - called by (1.). Computes likelihood of fully-censored data.
3. logLikelihood_nonspatial.R - called by (1.). Computes likelihood of non-censored data.
4. logPartial_nonspatial.R - called by (1.). Computes likelihood of partially-censored data.
5. Tools.R - Common functions for likelihood computation.
6. Fit_NonSpatialExpCopulaModel.R - Fits Exponential Factor Copula model to unreplicated data using (1.). Data is found in pairs_uniform.RData.

Multivariate Generalized Pareto Extremal Dependence Model Scripts

7. MGPD_Dep.R - centralized likelihood function for MGPD dependence model.
8. RevExp_U_Functions.R - Fitting MGPD dependence model to unreplicated data.
9. CommonFunctions.R - commonfunctions for (8.)