8. Discussion and conclusion
The accuracy of disease diagnosis depends on diagnosis method. Our expert system, in first step, the disease
screening, we established the novel model of knowledge representation for inference causes an accurate disease
screening which the accuracy depends on the veterinarian team expertise who defined swine's gender and age range
that indicates the disease. Second step, the disease diagnosis using the symptoms, we established the novel model of
uncertain knowledge representation for inference causes an accurate diagnosis which the accuracy depends on the
veterinarian team expertise who defined a significant weight of each symptom that indicates the disease.
Furthermore, another important part is the completeness of the occurred symptoms and symptom certainty factor
which are specified by the user. From the experiment, the occurred symptoms are incomplete or uncertain, but we
can diagnosis through the significant weight and the certainty factor of occurred symptom. In third step, the disease
diagnosis using the lesion, we established the novel model of knowledge representation for inference causes an
accurate diagnosis which the accuracy depends on the veterinarian team expertise who defined major lesion group
for confirmation of morbidness. In experiment, the occurred lesions are incomplete, but we can confirm morbidness
through major lesion group.