Persalys can realize sensitivity analyses to determine which input parameters are influential on the outputs.

Two types of indices are offered:

• Sobol indices: these indices are defined from the model ANOVA (analysis of variance). If several outputs are analysed, aggregated Sobol indices are also provided.
• SRC indices (Standard Regression Coefficient)these indices are retrieved by approximating the real model by a first-order linear model.

These indices (included between 0 and 1) allows to determine which input parameters are, or not, affecting the output. The closer the index is to 1, the more it is influential. In the case of the SRC indices, a positive or negative index allows to determine if the input variable has a growing or diminishing effect on the output.

For more information on how to create a model, refer to the following video: