Persalys 20.0 release
30/03/26
The new version of Persalys, version 20.0, is now available. The workflow for some analyses has been modified and the HSIC sensitivity analysis has been added.
In this update, data sets become more important and some analyses can now only be performed on data sets. The logic is now the following: once you have created and evaluated a design of experiments, you should export it as a data set. Then you can perform all the analyses on this data set, such as sensitivity analysis, metamodel creation, etc. This is the continuation of the 19.0 update where we added the possibility to extract the underlying design of experiment from most analyses. The other big novelty in this version is the addition of the HSIC indices for sensitivity analysis. Hilbert-Schmidt Independence Criterion can be used for screening as well as reliability-oriented sensitivity analysis thanks to target and conditional HSIC. We added a button to easily install Python packages inside the Persalys environment. It is now possible to run Monte Carlo central tendency analyses with failed evaluations.
