Received: 23 January 2025 / Accepted: 15 April 2025 / Published Online: 04 May 2025
Abstract
Brucellosis is a highly contagious zoonotic disease caused by the bacterium Brucella spp. that distresses mutually animals and humans, especially in underdeveloped countries with poor control programs. In adult cattle, the disease affects mainly the reproductive organs, thus causing major losses in production and reproduction, such as abortion and reduced fertility. This study involves the collection of 460 blood samples from dairy farms, which were analysed for brucellosis infection using the Rose Bengal Test (RBT). Additionally, data on the animals’ history, including placenta (retained), repeat breeding, their age, abortion, and lastly calving, were also recorded. To address the problem of class imbalance between the positive and negative classes, a technique, known as Synthetic Minority Over-sampling Technique (SMOTE) was applied in the research work. A total of five algorithms were used in this paper among them multilayer perceptron (MLP) and weekadeeplearning4j showed the best results for the prediction of brucellosis having 93.59% and 93.94% accuracy, respectively. Besides, risk factors are ranked based on their importance as ordered as retained placenta > repeat breeding > calving > abortion > age, and three association rules are made to understand the correlation of the factors for occurring the disease. By applying this study, early diagnosis of the disease could be possible to mitigate the economic losses.
Keywords: Brucellosis, Zoonotic disease, Dairy cattle, Machine learning, Risk factors, Public health
