Learning Bayesian-Network Topologies in Realistic Medical Domains

In recent years, a number of algorithms have been developed for learning the structure of Bayesian networks from data. In this paper we apply some of these algorithms to a realistic medical domain-stroke. Basically, the domain of stroke is taken as a typical example of a medical domain where much data are available concerning a few hundred patients. Learning the structure of a Bayesian network is known to be hard under these conditions. In this paper, two different structure learning algorithms are compared to each other. A causal model which was constructed with the help of an expert clinician is adopted as the gold standard. The advantages and limitations of various structure-learning algorithms are discussed in the context of the experimental results obtained.

Authors and Affiliations

1. Dept. of Computing Science, University of Aberdeen, Aberdeen, AB24 3UE, UK

   Xiaofeng Wu, Peter Lucas & Roelf Dijkhuizen

2. Stroke Unit, Aberdeen Royal Infirmary, Aberdeen, AB25 2ZD, UK

   Susan Kerr

Editors and Affiliations

1. Polytechnical University of Madrid Artificial Intelligence Laboratory, Medical Informatics Group, 28660, Boadilla del Monte (Madrid), Spain

   Jose Crespo  & Victor Maojo  & 

2. Health Bioinformatics Unit, Institute of Health Carlos III, Ctra. de Majadahonda-Pozuelo Km.2, 28220, Majadahonda (Madrid), Spain

   Fernando Martin

Reprints and permissions

© 2001 Springer-Verlag Berlin Heidelberg

Policies and ethics