Author(s)

Fulbert Bembamba¹, Frédéric T. Ouédraogo¹, Soudré Albert², Amadou Traoré³

¹Université Norbert Zongo, Laboratoire Mathématiques, Informatique et Applications (LAMIA), Koudougou, Burkina Faso.
²Université Norbert Zongo, Koudougou, Burkina Faso.
³Institut de l’Environnement et de Recherches Agricoles (INERA), Ouagadougou, Burkina Faso.

Unlike zebus, taurine cattle have the natural ability to resist trypanosomosis, a parasitic disease endemic to the humid areas of West Africa. However, repeated crossbreeding between zebus and taurine cattle is jeopardizing the genetic heritage of the Taurines and their ability to resist trypanosomosis. To strengthen protection and conservation efforts, it is essential to accurately distinguish purebred taurines from crossbreds. In this study, five Machine Learning models were built using morphological data collected from 1968 cattle. These models were trained to determine whether a given individual is purebred taurine or not. The classifiers yielded promising results. The random forest model and RBF Kernel SVM performed the best with up to 86% and 85% accuracy respectively. Moreover, the study of the correlation coefficients and the feature importance scores allowed us to define the most discriminating morphological traits.

Machine Learning, Trypanosomosis, Purebred Taurine, Accuracy, Model

Bembamba, F. , Ouédraogo, F. , Albert, S. and Traoré, A. (2022) Toward an Intelligent System for Taurine Cattle Recognition. Journal of Intelligent Learning Systems and Applications, 14, 1-13. doi: 10.4236/jilsa.2022.141001.