Performance comparison of classification algorithms for the diagnosis of mastitis disease in dairy animals

Abstract

Mastitis is a disease that occurs in milk-giving organisms and can reach fatal dimensions especially in dairy animals. This disease, which is usually caused by bacteria, causes significant changes in the physical and chemical structure of milk. Early diagnosis and treatment are very important because the life span of animals is shorter than that of humans. Data mining methods methods are frequently used in early diagnosis systems. Data mining is divided into several sub-branches. Classification is one of these sub-branches. In this study, some classification algorithms like J48, Random Forest, Support Vector Machines, k-nearest Neighbor Algorithm and Naive Bayes Algorithm are used and their performance is compared. These algorithms are applied to the Mastitis data set obtained from the total hundred animals and their performance is given. The results show that J48 algorithm has the best performance with the accuracy rate of 98%. © 2019 IEEE.