Posted Tue, 02 Jun 2015 20:55:19 GMT by Ivar Soares Urdalen
Hi

I am trying to use the Global Sensitivity Analysis experiment to analyze the kinetic and stoichiometric parameter in the ASM2d model

I ran the experiment with 100 Monte Carlo simulations and reasonable ranges for the parameters based on other sensitivity analysis studies.

I have a problem deciding on what statistical value to look at for evaluating the sensitivity of the parameter for the output I am looking at.

Vectors:
LCC: Linear Correlation Coefficient
NRC: Normalised Correlation Coefficient
ORC: Ordinary Regression Coefficient
PCC: Partial Correlation Coefficient
SPC: Semi-partial Correlation Coefficient
SRC: Standardised Regression Coefficient
VIF: Variance Inflation Factor
tLCC: t-statistics of LCC
tPCC: t-statistics of PCC
tSPC: t-statistics of SPC
tSRC: t-statistics of SRC
r: Residuals

What is the best factor to look at for evaluating the sensitivity of the each parameter?
Posted Wed, 03 Jun 2015 09:15:31 GMT by Lorenzo Benedetti
On how to perform a GSA, I suggest to have a look at:
Benedetti, L., Claeys, F., Nopens, I. and Vanrolleghem P.A. (2011) Assessing the convergence of LHS Monte Carlo simulations of wastewater treatment models. Water Science and Technology 63(10), 2219-2224.
About the factors, SRC (and others) is OK, but also look at the R2. How many parameters you have with PDF?

You must be signed in to post in this forum.