Representation and analysis of causal knowledge in medicine.

Abstract

In medicine, we seek to discover and represent the causal relationships between variables of interest. In order to represent causal knowledge computationally, it is necessary to uses to directed graphs. There are two fundamental techniques: Bayesian networks and diffuse cognitive maps. In the present work both techniques are compared and the advantage of diffuse cognitive maps in the differential diagnosis is shown by calculating the scoring of each alternative using multi-criteria methods.