Improvement of sensitivity analysis

[HvanderStok]

• [x] Add more methods from the SALib
• [x] Some information is lost in the results of the senitivity analysis

[github-actions[bot]]

Branch 43-Improvement-of-sensitivity-analysis created!

[HvanderStok]

Current status

I checked the first boxes of this issue, where I implemented in branch 43-Improvement-of-sensitivity-analysis:

• [x] Add more methods from the SALib:
  • [x] FAST method
  • [x] PWAN method
• [x] Some information is lost in the results of the sensitivity analysis:
  • [x] Option to get verbose information, where also the sensitivity of the single target values in combined goals are calculated
  • [x] Option to get all analysis variables which are calculated with the methods of the SALib

New issues to this topic

In the process, some additional bugs and options for improvements occurred, on which I am currently working:

• [x] The sampler of the Sobol method must be changed according to the SALib
• [x] Automatic visualization of the sensitivity results
• [x] Optimization of memory usage during sensitivity analysis with large data sets
• [x] Use same simulations for sensitivity analysis for different calibration classes on the same data
• [x] Save samples and corresponding statistical measures for later use in the calibration process
• [x] Option to get the same verbose information in the validation as in the sensitivity analysis. This could give the possibility to make a connection between the validation results the results of the sensitivity analysis.