Are variance based sensitivity analysis methods healthy for your models?
A combination of causes has hindered so far the effective use of sensitivity analysis (SA) for the appreciation of the sources of uncertainty affecting the prediction of a model including epistemological factors, cultural dominance factors, and practical factors. The epistemological reason is that many different things are meant by sensitivity analysis in different user communities. For an engineer, SA could be the process of moving or changing components in the design of a plant to investigate how a fault tree analysis for that plant would change. For an economist, it is most likely the derivative of the output function with respect to the input factors. For a software engineer, SA will be related to the robustness and reliability of the software with respect to different assumptions. There is no doubt that it will be extremely hard to convince those potential users of the benefit of new SA tools unless this epistemological barrier is overcome.
Bibliographic Reference: Paper presented: SAMO 98, Venice (IT), April 19-22, 1998
Availability: Available from (1) as Paper EN 41269 ORA
Record Number: 199810474 / Last updated on: 1998-05-05
Original language: en
Available languages: en