Learning and discovery from a clinical database: an incremental concept formation approach

V W Soo; J S Wang; S P Wang

doi:10.1016/0933-3657(94)90065-5

Learning and discovery from a clinical database: an incremental concept formation approach

Artif Intell Med. 1994 Jun;6(3):249-61. doi: 10.1016/0933-3657(94)90065-5.

Authors

V W Soo¹, J S Wang, S P Wang

Affiliation

¹ Department of Computer Science, National Tsing-Hua University, Hsin-Chu, Taiwan.

PMID: 7920969
DOI: 10.1016/0933-3657(94)90065-5

Abstract

The main interest of this research is to discover clinical implications from a large PTCA (Percutaneous Transluminal Coronary Angioplasty) database. A case-based concept formation model D-UNIMEM, a modified version of Lebowitz's UNIMEM, is proposed for this purpose. In this model, we integrated two kinds of class memberships: the feature-disjunction class membership and the index-conjunction class membership. The former is a polythetic clustering approach and serves at the early stage of concept formation. The latter allows only relevant instances to be placed in the same cluster and serves as the later stage of concept formation. D-UNIMEM could extract interesting correlations among features from the learned concept hierarchy.

Publication types

Research Support, Non-U.S. Gov't

MeSH terms

Abstracting and Indexing
Angina Pectoris
Angioplasty, Balloon, Coronary* / adverse effects
Artificial Intelligence*
Cardiac Catheterization
Coronary Artery Bypass
Diabetes Mellitus
Electrocardiography
Exercise Test
Gated Blood-Pool Imaging
Humans
Hypertension
Information Systems*
Learning*
Neural Networks, Computer
Risk Factors
Smoking