Hierarchihal Bayesian parameter estimation for HIV dynamic models

Abstract

Most studies on parameter estimation for HIV dynamic models have ignored pre-treatment viral load data hence utilizing only post-treatment viral load data. In this study we utilize pre-treatment viral load data to estimate parameters of the HIV dynamic model in the absence of therapy. By employing hierarchical Bayesian parameter estimation approach, we were able to get reasonably robust estimates of the model parameters. Using simulated data, the parameter estimation was done at both the individual and population levels with the implementation carried out via Markov Chain Monte Carlo methods.