Impulse parametric sensitivity analysis

Abstract

Complexity often limits human intuition in understanding system behavior, motivating the creation and analysis of mathematical models to understand these systems. One of the most commonly used analyses is the classical parametric sensitivity analysis (PSA), which maps out the parametric dependence of system behavior. In this article, we present a novel sensitivity analysis, called impulse parametric sensitivity analyses (iPSA), which is based on impulse perturbations to model parameters. Hence, the iPSA gives a step-by-step mechanistic insight on how a particular output behavior is accomplished and complementarily how (an impulse) perturbation propagates through the network. More specifically iPSA coefficients can reveal which parameters are essential and when they become important. The efficacy of this analysis is demonstrated on examples drawn from biology: the Fas-induced cell death model of human Jurkat T-cells.