Towards Effective Structure Learning for Large Bayesian Networks
Prashant Doshi, Lloyd G. Greenwald, John Clarke · 2002
The effectiveness of Bayesian network construction is a func-tion of the predictive ability and speed of inference of the re-sulting network, the effort and expertise required to construct the network, and the effort and expertise required to under-stand the network. We empirically evaluate three alternative methods for constructing Bayesian networks from data, con-sidering both objective performance measures and subjective construction cost measures. Empirical results are obtained for a large real-world medical database. We provide results comparing network structure laboriously elicited from a do-main expert to structure automatically induced by two alter-native structure learning algorithms. The parameters of the Bayesian network produced by each method are induced us-ing the Bayesian MAP approach. Additionally, we introduce the use of classification paths as an aggregation technique to reduce the size and structural complexity of Bayesian net-works. Empirical results are obtained both with and without this complexity reduction technique.