A logic and computer algebra based expert system for diagnosis and treatment of migraine

Laura Laita, B. López-Bravo, Eugenio Roanes–Lozano, Laura Laita, Luis de Ledesma · 2007

A Rule-Based Expert System, dealing with both diagnosis and treatment of migraine is described in this chapter. The aim of the chapter is first to propose a middle- sized expert system that could help family doctors and health care personnel in the process of guess a diagnosis that may suggest them sending the patient to a specialist. A second purpose, not less important, of the chapter is to propose to teams of researchers in both medicine and computer science, a reliable methodology of construction of expert systems that might, if the information they contain is larger and deeper than the one this chapter expert system contains, help specialists to study and treat migraine automatically. The system is built up from three sub-systems that deal, respectively, with a diagnosis about whether or not a person suffers chronic migraine, with the type of migraine the patient may suffer and with the treatment. Each subsystem has a knowledge base containing the necessary information, and an inference engine, which enables both to verify consistency and to extract consequences from the information which the Knowledge Base of each sub-system contains. These knowledge bases can, be improved or changed by specialists now, or in the future, because, although in the second subsystem some attention is paid to internal factors that lead to suffering chronic migraine, they almost only consider external factors. Nevertheless, although changes could be made in the knowledge bases, the inference engines, Grobner Bases based, would not need essential changes. The Grobner bases approach to Rule Based Expert System to be applied in this chapter has been applied by the authors in other different fields. It is able to deal with multi-valued and modal logics, allowing in this way to deal with uncertainty. © 2007 Nova Science Publishers, Inc. All rights reserved.

Read the paper · More papers on PaperTik