Animal Molecular Breeding 2024, Vol.14, No.5, 335-344 http://animalscipublisher.com/index.php/amb 338 indicate health issues or changes in welfare. For instance, wearable sensors and barn monitoring systems can track feeding and drinking patterns, which are essential for early detection of diseases and optimizing feeding strategies (Figure 1) (Eckelkamp, 2019; Bloch et al., 2023). Additionally, social behaviors can be monitored to understand herd dynamics and identify potential stressors or welfare concerns (Sharma et al., 2021). Environmental factors such as temperature, humidity, and air quality can significantly impact animal performance and health. Integrating these factors with behavioral and physiological data can provide a comprehensive understanding of how environmental conditions affect dairy cows. Machine learning models can analyze these integrated datasets to predict outcomes such as milk yield, disease risk, and overall animal performance (Cockburn, 2020; Lasser et al., 2021). This holistic approach enables farmers to make data-driven decisions to optimize the barn environment and improve animal welfare and productivity (Cabrera et al., 2019). The integration of diverse data sources, including sensor technology, genomic data, and behavioral and environmental data, is revolutionizing precision dairy farming. Machine learning models leveraging these data sources can provide valuable insights for improving herd management, disease detection, and breeding programs. As technology continues to advance, the potential for more accurate and comprehensive predictive analytics in dairy farming will only grow, leading to more sustainable and efficient farming practices. Figure 1 Component of the location and acceleration measuring system installed in a barn (Adopted from Bloch et al., 2023) Image caption: RuuviTag inside a protecting plastic box (a), tag on the cow collar (b) and receiving station installed on a barn structure (c) marked by red circles (Adopted from Bloch et al., 2023) 4 Machine Learning Models and Techniques Used 4.1 Supervised learning in dairy farming Regression models are widely used in precision dairy farming to predict various continuous outcomes such as milk yield and quality. These models help in understanding the relationship between different farm parameters and the output variables. For instance, regression-based algorithms have been applied to predict milk production and quality by analyzing factors like milking parameters and milk properties (Slob et al., 2020). Additionally, regression methods are utilized to forecast yield by integrating agrarian factors, which include soil, climate, and water regime (Elavarasan et al., 2018; Yang, 2024). Classification algorithms are extensively used to predict health-related issues and yield outcomes in dairy farming. Decision tree-based algorithms are the most commonly used, followed by artificial neural networks (Slob et al., 2020). These algorithms help in early disease detection, which is crucial for maintaining the health and productivity of dairy cows. For example, supervised ensemble classifiers have been employed to classify cattle behavior patterns, aiding in the early detection of health issues such as lameness (Dutta et al., 2015). Moreover, classification techniques are used to predict various health conditions and milk yield, enhancing the decision-making process for dairy farmers (Cockburn, 2020). 4.2 Unsupervised learning techniques Clustering techniques are used to group animals based on their health and productivity metrics. These unsupervised learning methods help in identifying patterns and anomalies that are not apparent through traditional
RkJQdWJsaXNoZXIy MjQ4ODYzNA==