On selecting the optimal Bayesian Network model construction approach

The construction of a Bayesian Network (BN) model entails two major tasks: realization of the model structure, and the calibration (parameterization) of the model. BN model constructors, ab initio, relied only on domain experts to define both the structure and parameters of a model. Currently, algorithms exist to construct BN models from data. Consequently, there are three BN model construction techniques: total expert-centred, total data-centred, and semi data-centred. We empirically investigated which of these approaches is the optimal approach for the construction of a BN model for our intended application. The investigation yielded some interesting themes.