Target and Conditional Sensitivity Analysis with Emphasis on Dependence Measures

In the context of sensitivity analysis of complex phenomena in presence of uncertainty, we motivate and precise the idea of orienting the analysis towards a critical domain of the studied phenomenon. We make a brief history of related approaches in the literature, and propose a more general and systematic approach. Nonparametric measures of dependence being well-suited to this approach, we also make a review of available methods and of their use for sensitivity analysis, and clarify some of their properties. As a byproduct, we notably describe a new way of computing correlation ratios for Sobol' indices, which does not require specific experience plans nor rely on independence of the input factors. Finally, we show on synthetic numerical experiments both the interest of target and conditional sensitivity analysis, and the relevance of the dependence measures.