Reduced sampling schedule for the glucose minimal model: importance of Bayesian estimation