Classification of pig stress condition using paraconsistent logic
Jonas Silva, Jair Minoro Abe, Irenilza de Alencar Nääs, Alexandra Cordeiro · 2019 Boston, Massachusetts July 7- July 10, 2019 · 2019
Abstract. Pork is the most consumed meat worldwide, and there is the need for producer countries to relay on complying with the animal well-being international norms. The present study aimed to develop a software that predicts stress in the piglet using the vocal calls emitted during stressful conditions (cold/heat, pain, hunger, thirst). The stress-free piglet was considered the baseline (normal). Vocal signal intensities were extracted from 40 piglets exposed to the stress. The stress conditions considered were heat stress, pain, thirst, and hunger. The database was organized, and the paraconsistent logic was applied to solve the uncertainties generated with the overlap of the intensity of the vocal signals. Results indicate that the most accurate prediction was for the pain (93.1%). The less accurate prediction was for the sound piglet (normal). Although using the solution for resolving most of the uncertainties and overlapping, only the stress by pain was readily detected as the vocalization due to pain has a high intensity and a long duration. Further research connecting the vocal signal and other recorded pattern are needed to improve the accuracy of the stress predicting process.