Skip to content

Feature/pcegpe

Maria Fernanda Morales requested to merge feature/pcegpe into develop

Create a new surrogate class that combines PCE+GPE by calling the existing PCE and GPE class

  • Build metamodel function (MF)
  • Fit (train) function (calling PCE and then GPE) (MF)
  • eval (evaluation) function (MF)
  • PyTests (Alina)
  • update changelog (Alina and MF)
  • merge with calculatemoments branch
  • add simple example (analytical example) (Alina and MF)
  • Check PostProcessing with new class (Alina)

On a fork

  • add example comparing PCE+GPE with standard PCE (same solver)
  • add example comparing PCE+GPE to other Bayesian PCE approaches (BayesianRegression, FastARD)

Notes: For the sequential_training() and PostProcessing functions to work with no errors, this branch must first be merged with the calculatemoments branch.

Edited by Alina Lacheim

Merge request reports