Compute total and individual sensitivity indices, significant components, and confidence intervals for the total sensitivity indices.
The total and individual sensitivity indices are calculated using a method based on the VIP of the PLS regression, proposed by J.P. Gauchi. The significant components are calculated by the SIMCA software rule and by the Lazraq & Cl<U+00E9>roux test. The confidence intervals for the total sensitivity indices are determined by the bootstrap method. A fast algorithm is implemented which allows big datasets analysis.
Gauchi, J.-P. and Lehuta, S. and Mah<U+00E9>vas, S. 2010. Optimal sensitivity analysis under constraints: Application to fisheries. In Procedia - Social and Behavioral Sciences, vol. 2. Elsevier, pp. 7658-7659. Sixth International Conference on Sensitivity Analysis of Model Output (Milan, Italy), 19-22 July; co-organized by the ELEUSI research center of Bocconi University and by the Joint Research Center of the European Commission.
Gauchi, J.-P. 2012. Global Sensitivity Analysis: The SIVIP method (SAS/IML language). Rapport technique 2012-3. INRA, UR1404, F-78350 Jouy-en-Josas, France.
Gauchi, J.-P. 2015. A practical method of global sensitivity analysis under constraints. Rapport technique 2015-1. INRA, UR1404, F-78350 Jouy-en-Josas, France.
Lazraq, A. and Cl<U+00E9>roux, R. 2001. The PLS multivariate regression model: testing the significance of successive PLS components. Journal of Chemometrics. Vol. 15(6), pp 523-536.
Lazraq, A. and Cl<U+00E9>roux, R. and Gauchi, J.-P. 2003. Selecting both latent and explanatory variables in the PLS1 regression model. Chemometrics and Intelligent Laboratory Systems, Vol. 66(2), pp 117-126. Elsevier. doi:10.1016/S0169-7439(03)00027-3.
SIMCA Software. http://www.umetrics.com/products/simca.