Besides the usual bug fixes there have been improvements to the new sensitivity analysis methods. Higher order Sobol indices are now supported (analytically and using Monte Carlo).
There is also a convenience function for quickly plotting a pie chart of the total Sobol indices of a model.
>> fig = model.plotSobolIndices()
1×6 graphics array
- Added support sensitivity analysis
- Monte carlo Sobol indices and derivatied-based measures
- Analytical derivations are implemented for Kriging, GPML, LS-SVM
- Added SensitivityCrossValidation
- Calculates the error on the sensitivity indices
- Added higher order Sobol indices for LSSVM, Kriging and Gaussian process models
- Added Sobol plot function for the total Sobol indices (as pie chart)
- Fixed export model of the Kriging model for higher dimensions
- Fixed crash with intelligent restart strategy for the model parameters if the Statistics toolbox was missing