Robust parameter estimation techniques for stochastic within-host macroparasite models (Q2177452)

This paper studies robust parameter estimation techniques for stochastic within-host macro-parasite models. The model is Markovian with triple of random variables \(I(t)\), \(L(t)\) and \(M(t)\), where \(I\) is the hosts experience of infection, \(L\) is the number of three-stage larvae and \(M\) is the number of mature parasites. The paper explores different approaches for estimating the parameters of simple non-linear stochastic dynamical models. Influences studied using simulation including a stochastic goodness-of-fit statistic, maximization of goodness-of-fit and confidence bounds and goodness-of-fit \(p\) values. Moment equations and moment closure assumptions have been studied. Robust estimation techniques are shown to give greater biological insight. The methods developed here are tested using data from a lymphatic filariasis animal model.