I.W. Cavalcanti, F.G. Rodrigues, L.M. Brasil, J.C.C. Rojas, M.T. Britto Filho, and L.J. da Costa (Br
Biomedical Engineering, ArtificialIntelligence, Expert Systems, Odontological Semiology.
This work presents an ES targeted at the aid in the diagnosis in stomatology. Denominated OdontSys, this system has an experimental application in the group of lesions of the Non-Neoplastic Proliferative Processes. OdontSys presents a symbolic architecture, has a mode of electronic report associated to its structure and has a visualization resource of images related to the lesions in study. The Knowledge Base of OdontSys makes use of the Logical Fuzzy as logical formalism. The knowledge is structured on triads that are made up of the term lesion, the sign or symptom and a score. The last one mathematically represents the importance of the sign or symptom to the diagnosis of the lesion. The Inference Machine makes use of a search function with a forward chaining. The answers emitted by the system are suggestions of diagnoses that are presented in three levels of importance. Two experts performed the final tests for the system through evaluation. Fifty tests were performed, in 88% of these, the system emitted response in the first level and in 12% the responses were emitted in the second level. In 33 cases, the system suggested only one lesion as response and in 11 cases the system suggested two lesions. In 6 cases, the system suggested multiple responses.
Important Links:
Go Back
