Global sensitivity analysis of nonlinear models : Importance measures and Sobol sensitivity indices
The existing relevant literature on the variance-based sensitivity analysis methods is reviewed, focusing on Sobol' sensitivity indices. The relationship among the various methods is investigated, and the relations of Sobol' indices to other similar measures is discussed. Sobol' indices are tested on a few analytical and computational models. The computation of the indices themselves is ameliorated and finally a version of the indices based on rank transformation of the data is discussed. The computational novelty of this study is the introduction of the 'total effect' parameter index. This provides a measure of the global effect of a parameter, including all the possible synergetic terms arising from the coupling of that parameter with all the others. Rank transformation of the data are introduced in order to increase the reproducibility of the method. The main conclusion of this work is the identification of a sensitivity analysis methodology which is both flexible, accurate and informative, and which can be achieved at reasonable computational cost.
Bibliographic Reference: EUR 16052 EN (1994) 39pp., FS, free of charge
Availability: Available from the Public Relations and Publications Unit, JRC Ispra, I-21020 Ispra (IT)
Record Number: 199510011 / Last updated on: 1995-03-21
Original language: en
Available languages: en