Cattle Pinkeye Disease Classification using Machine Learning

Abstract

Pinkeye (infectious bovine keratoconjunctivitis, or IBK) is a bacterial infection of the cattle eye that causes inflammation and, in severe cases, temporary or permanent blindness. It is a painful, debilitating condition that can severely affect animal productivity. Due to lower weight gain, lower milk production, and higher medical costs, the cattle industry could experience large losses. Previous studies tried to classify livestock diseases using machine learning, but there has been a lack of studies conducted on pinkeye disease classification. The proposed study aims to design a classification model to classify whether the infected cattle have pinkeye or not at an early stage by analyzing a set of attributes. The study collected data from the Wolaita Sodo Kenido Koyisha Wereda Livestock and Fishery Office. The significance of this study is to prevent the expansion of disease among the cattle with an early detection for taking precautionary measures. The researchers used the percentage splits 80/20, 70/30, 60/40, and 90/10 to build classification models. Based on the results of the experiments, the researcher chose the 70/30 split due to the better performance obtained. The study trained four different models, including Random Forest, AdaBoost, Artificial Neural Network, and Extreme Gradient Boost algorithms. These models are selected based on an exhaustive study conducted. To assess the algorithm's performance, confusion matrix, accuracy, precision, recall, and f1-score have all been utilized. By a 99.15% accuracy, the Artificial Neural Network outperforms the other algorithms by all the metrics except recall.

Keywords: Machine learning, Infectious bovine keratoconjunctivitis, IBK, Cattle pinkeye, Pinkeye disease classification