Hi !
This is just a message to share two pointers towards discussions on sensitivity analysis.
I have asked a question about the use of SRC versus Pearson correlation at Stackexchange that might be interested for those who use linear regression for sensitivity analysis:
There was an interesting question on UQWorld about Sobol’ indices:
The thing is that the sum of the estimate of first order Sobol’ indices has 50% chance of being larger than 1 for a linear model, a result which might be surprising at first.
Regards,
Michaël