Sistema experto para diagnosticar enfermedades en mascotas caninas para la fundación “CRIAR”.

Abstract

The development of an expert system for the diagnosis of canine diseases responded to the need to provide accurate and timely diagnoses, especially in communities with limited access to veterinary experts. This study focused on the creation of an expert system to diagnose diseases in canine pets for the Criar Foundation, using the SCRUMBAN methodology combining elements of Scrum and Kanban, for agile and adaptive management of the project. This approach integrated seven clearly defined phases, from goal setting to final product deployment, ensuring accurate and comprehensive implementation of the expert system. The results revealed an exceptionally high system accuracy, with an accuracy of 100% in most of the diseases evaluated. However, an opportunity for improvement was identified in the accuracy of the diagnosis of Mite Dermatitis (E05), with 71%. The system's sensitivity and specificity were generally strong, highlighting its ability to detect real cases and avoid false positives. Validation of the system compared to diagnosis by expert veterinarians demonstrated impressive performance, although areas for improvement were identified, especially in the diagnostic accuracy of certain diseases. In conclusion, the expert system represents a valuable tool for veterinary medicine, with the potential for continuous improvement through updates and refinements in its knowledge base and artificial intelligence algorithms.