A model for expert reasoning: fuzzy ganglionar lattices

This work presents an elemental model for expert\´s reasoning with multiple antecedents of different relative importance. The model is essentially obtained by decomposing a global reasoning into basic ones, flowing from concepts with useful meaning to experts. Such a decomposition results in a lattice of basic reasonings which enables a suitable propagation of the antecedents\´ characteristics; such as intensity. The proposed model is adjusted and tested for the diagnosis of "cardiac insufficiency with hypertrophic left ventricle", which is a part of an expert system to aid in the organic characterization of hypertense patients. From this test, capacities and limitations of the model are inferred, and a discussion about possible extensions of the model to include the expert\´s approximate reasoning, explicative capacity, and the uncertainty of antecedents, is performed.