VARS-TOOL Tutorial 1: Sensitivity Analysis (SA) with VARS and G-VARS

Exercise 1: From A to Z on the Ishigami test function https://github.com/vars-tool/vars-tool

Objectives:

Experiment 1 | First, let's run SA on Ishigami via VARS when the inputs are uncorrelated and uniformly distributed - this is the classic SA setting in the literature. This experiment is a benchmark for the rest of this exercise.

Experiment 2 | Now, switch to G-VARS but keep the inputs uncorrelated and uniformly distributed - this experiment is essentially the same as the previous experiment but with a more generalized method, with the objective of testing if G-VARS can re-produce VARS results.

Experiment 3 | Run G-VARS when the inputs are correlated but still uniformly distributed - the objective of this experiment is to assess the role of correlation between inputs on the SA results.

Experiment 4 | Finally, run the previous experiment now with both correlated and non-uniformly distributed inputs - the objective of this module is to see what happens when the inputs follow other distributions such as normal or triangular distribution.