Causal probabilistic networks for solving parameters in compartmental models

Abstract

Using Causal Probabilistic Networks for solving parameters in compartmental models was evaluated with the Minimal Model (R.N. Bergman and C. Cobelli1980) of glucose kinetics and the High Level Specification Language developed by the authors. The networks were evaluated using data from a previous study (D. R. Owens et al 1996) where the parameters for the model had been determined. 32 models were developed to find the best representation of the minimal model. The predictions were compared to the results obtained in the previous study to measure the performance of the Causal Probabilistic Network and the network generator.