New paper published in International Journal for Uncertainty Quantification
Together with O. Roustant, F. Gamboa and B. Iooss from the Institut de Mathématiques de Toulouse in France, N. Lüthen, S. Marelli, and B. Sudret published a new paper about global sensitivity analysis using Poincaré basis functions.
In the paper entitled Global sensitivity analysis using derivative-based sparse Poincaré chaos expansions, we explore a new type of spectral expansion for surrogate modeling and sensitivity analysis. The multivariate Poincaré basis is an orthonormal basis with the unique and defining property that partial derivatives of the basis functions are again orthogonal with respect to the same probability measure. It is therefore ideally suited for including model derivatives into the surrogate modeling process, and allows the analytical computation of Sobol' indices and derivative-based global sensitivity indices (DGSM) from the expansion coefficients.
For more information, please follow this external pagelink for the publication and this link for the associated report on our internal archive.