Global sensitivity measures from given data

摘要: Simulation models support managers in the solution of complex problems. International agencies recommend uncertainty and global sensitivity methods as best practice audit, validation application scientific codes. However, numerical complexity, especially presence a high number factors, induces analysts to employ less informative but numerically cheaper methods. This work introduces design for estimating indices from given data (including simulation input–output data), at minimum computational cost. We address problem starting with statistic based on L1-norm. A formal definition estimators is provided corresponding consistency theorems are proved. The determination confidence intervals through bias-reducing bootstrap estimator investigated. strategy applied identification key drivers computer code developed National Aeronautics Space Administration (NASA) assessing risk lunar space missions. also introduce symmetry result that enables estimation measures datasets produced outside conventional functional framework.