Prediction of brucellosis disease with ensemble machine learning application

Brucellosis, a global public health challenge caused by Brucella spp., disproportionately affects underdeveloped regions, posing threats to both human and animal populations. This study focuses on the impact of Brucella abortus on cattle in Qazvin, Iran, emphasizing the importance of effective control measures. Serological tests, particularly iELISA, play a crucial role in this regard. Leveraging machine learning techniques, including Vote, AdaBoostM1, LogitBoost, and Bagging, the research aims to predict Brucella infection in cattle using comprehensive data. The study meticulously describes each model, ultimately identifying AdaBoostM1 as the optimal choice. This model demonstrates a robust 75% correlation, low Root Mean Square Error (RMSE) and Mean Absolute Error (MAE), and a commendable Kappa Statistics score of 0.4965. The findings underscore the potential of ensemble machine learning in Brucellosis prediction, showcasing proficiency in handling complex datasets and improving accuracy. AdaBoostM1 emerges as the preferred model, offering valuable insights for disease prediction and contributing to enhanced control strategies. This research contributes to advancing our understanding of Brucellosis and underscores the pivotal role of advanced machine learning in addressing complex public health challenges.