Improved and accelerated sensitivity analysis using State Dependent Parameter models

In this paper we apply State Dependent Parameter (SDP) modelling (a model estimation approach, based on recursive filtering and smoothing estimation) to estimate the variance based main effect sensitivity indices of computational models (a popular method for quantitative sensitivity analysis). The method is conceptually simple and all measures of interest are computed using a single set of model runs. It is flexible because in principle it can be applied with any available type of Monte Carlo sample. It is extremely efficient, especially when coupled with efficient sampling methods, such as the quasi-random LP-tau sequence, it allows for a strong reduction in the cost of the analysis. Moreover, the method also allows us to estimate the first order terms of the High Dimensional Model Representation of the model under analysis.