PILA, Adriano D.; BORIM, Andréa C. A.; ZANI, Marcelo.; JÚLIO, Márcia R. F. M.
Resumo:
The purpose of this paper is to show how computer system can help the prediction and control
processes in the case of horse colic. Horse colic is an easy treatment problem if diagnosed just in time.
However, early diagnostic is not always easy to get just in time, what takes the animal into a serious pain
situation. In most cases, the exams to detect the colic are invasive and provoke a lot of pain in the animal. In
extreme and not rare cases the animal dies or must be sacrificed. This cause a considerable financial injury
to the horse creators that see their investments being frustrated and without any expectative to get the money
back. This work presents a horse colic prediction system called SiPreCE, which in portuguese stands for
Horse Colic Prediction System. This system uses Artificial Intelligence techniques, actually Machine
Learning, to analyze the information collected with the creators help and later induction of a classification
prediction system capable of predict the animal destiny based on the data collected and compared with the
early real cases. In this way initial symptoms that may lead the animal to death can be treated and solved
quickely. Thus, in this work it is analyzed a dataset with 368 cases about horse colic which proposed system
is capable to point out situations where the animal may live or die.