Document Type
Poster
Publication Date
Summer 2024
Abstract
Global Sensitivity Analysis (GSA) is the process of apportioning the variation in a simulator model’s output to each of its respective inputs. There are various approaches to GSA, but Sobol’s variance-based approach is perhaps the most utilized. This approach relies on centered main effect functions to quantify the magnitude of particular inputs' effects on the response's variability. However, this method heavily assumes a hyper-rectangular input space, which is not always the case, as input spaces are often constrained. This poster proposes an alternative method for GSA in the case of a constrained input space.
Recommended Citation
Mayer, Andrew, "Calculating Empirical Sensitivity Metrics for Computer Experiments with Constrained Input Spaces" (2024). Kenyon Summer Science Scholars Program. Paper 729.
https://digital.kenyon.edu/summerscienceprogram/729