PyApprox provides flexible and efficient tools for credible data-informed decision making. PyApprox implements methods addressing various issues surrounding high-dimensional parameter spaces and limited evaluations of expensive simulation models with the goal of facilitating simulation-aided knowledge discovery, prediction and design. Methods are available for: low-rank tensor-decomposition; Gaussian processes; polynomial chaos expansions; sparse-grids; risk-adverse regression; compressed sensing; Bayesian inference; push-forward based inference; optimal design of computer experiments for interpolation regression and compressed sensing; and risk-adverse optimal experimental design.

Software Website

Software Downloads

Jakeman, John Davis,