INTRODUCTION
Worldwide, domestic animals are severely hampered by cattle disease, which also significantly
increases animal mortality [1]. The field of livestock management is currently undergoing a
transition to a high-throughput digital era. Based on quantitative and qualitative analytical
results, actionable decisions are made on a daily and hourly basis, utilizing vast amounts of data
collected by systems of electro-optical, acoustical, mechanical, and biosensors [2]. Animal
husbandry and human health are greatly threatened by major animal illnesses. Big data analysis
and prediction of animal diseases is becoming more and more crucial due to the acceleration of
globalization and the wealth of available data [3]. Due to farmers' lack of knowledge about early
health disorder prevention and detection methods, the majority of subclinical diseases go
undiagnosed until they manifest clear clinical symptoms [4]. This prolongs the detrimental
effects on the health and productivity of dairy animals while makes treatment challenging and
costly [2]. Therefore, information obtained before a diagnosis that is capable of predicting an
animal’s risk or identifying a disease more accurately than clinical indicators would naturally be
more helpful for early identification and intervention. Within the domain of Artificial
Intelligence (AI), Machine Learning (ML) employs statistical methods to identify or forecast
disease occurrences or cattle performance through the use of extensive datasets [5], [6]. ML
algorithms aims to teach computers how to learn from data and use that knowledge to analysis
and prediction [4], [7], [8]. This study utilized ML algorithms for early analysis of disease
prediction and classification in livestock. Using random forest, naïve Bayes, and eXtreme
Gradient boosting, neural networks, decision-tree induction, and logistic generalized linear
mixed models, researchers have employed ML techniques to identify or forecast a number of
health conditions, including clinical mastitis [9], [10]. Through efficient use of readily available
various medical and anthropometric data, researchers can also develop extremely accurate
prognostic and diagnostic models for many additional disorders, such as heart disease, metabolic
syndrome in humans, animal diseases [7], [11]–[13]. This is made possible by the development
of modern information processing techniques, particularly those based in the domains of AI, ML
and healthcare. ML is an algorithmbased data analysis technology having strong data analysis
capabilities to deal with intricate interconnections between variables [14]. In light of these
factors, an ML based risk prediction model for cattle animal disease diagnosis was created for
this investigation. The effectiveness of several ML classifiers, such as Support Vector Machines
(SVM), Random Forests (RF), K nearest neighbour (KNN), Extreme boosting (XGBoost), Light
gradient boosting (LGB), Categorical boosting (Catboost), for disease prediction and
classification has been used and investigated in this work.