An uncertainty quantification framework for a Zika virus epidemic model

Abstract

Uncertainty quantification is an important procedure when dealing with errors and discrepancies that are present in any modeling effort. This work presents a consistent uncertainty quantification framework for an epidemiological dynamical system, which is able to construct robust descriptions given a calibrated model. Since arbitrary choices of distributions for the input parameters can provide biased estimates and results, the maximum entropy principle is employed in the construction of the stochastic model to infer the most possibly unbiased probability density functions affected by the lack of information. The framework is applied on a SEIR-SEI compartmental system for the Brazilian Zika virus outbreak to study a stochastic scenario.