Covxtreme: Matlab Software for Non-Stationary Penalised Piecewise Constant Marginal and Conditional Extreme Value Models

• DOI: 10.2139/ssrn.4633238
• Publication Date: 2023-11-14T21:28:13Z
• Authors: Ross Towe; Emma Ross; David Randell; Philip Jonathan
• Keywords: Constant (computer programming); Value (mathematics)

covXtreme is MATLAB software for estimation of marginal and conditional extreme value models, non-stationary with respect to covariates, environmental design contours. Generalised Pareto models peaks over threshold are estimated, using a piecewise-constant representation the variation scale parameters on covariate domain interest. The one or more associated variates, given large single conditioning variate, described extremes model Heffernan Tawn (2004). Optimal smoothness covariates estimated cross-validated roughness-penalised maximum likelihood estimation. Uncertainties in parameter estimates due choice, sample size, quantified bootstrap resampling scheme. Estimates contours various schemes calculated under fitted models. Software user guide can be downloaded from GitHub.