AN EMPIRICAL-COMPARISON OF EXPERT-DERIVED AND DATA-DERIVED CLASSIFICATION TREES

Classification trees provide an attractively transparent discriminatio n technique, and may be derived from both expert opinion and from data analysis, We consider a real and complex problem concerning the diagn osis of babies with suspected critical congenital heart disease into o ne of 27 classes, A full loss matrix for all possible misclassificatio ns was obtained from clinical assessments. A tree derived from expert opinion was compared with those derived from analysis of 571 past case s, both for the full problem and for a subset of 6 diseases. Automatic methods for tree creation and pruning were found to have problems for rare diseases, and hand-pruning was carried out. Inclusion of costs l ed to much improved clinical performance, even for trees that had orig inally been constructed to minimize classification errors. The expert tree showed a specific building strategy that could not be reproduced automatically. The expert tree generally outperformed those derived fr om data, particularly in the ability to identify important composite f eatures.