AI and Data Analytics in the Dairy Farms: A Scoping Review.

Area of Science:

• Agricultural Technology
• Data Science
• Animal Science

Background:

• Rising global population necessitates increased bovine milk production.
• Advanced technologies like data analytics, big data, and sensors are crucial for efficient milk yield enhancement.
• A review of modern technologies and data analytics for milk prediction is timely to support dairy farm decision-making.

Purpose of the Study:

• To conduct a scoping review of modern technologies and data analytics methods for bovine milk production prediction.
• To identify trends in research scope, data types, analytical methods, and predictive models.
• To inform future research directions for enhanced dairy farm decision-making tools.

Main Methods:

• Scoping review methodology applied to 151 relevant articles.
• Analysis of publication trends, study scope, data input types, and analytical techniques.
• Evaluation of machine learning models, particularly artificial neural networks and convolutional neural networks.

Main Results:

• Studies are recent (average 5.95 years old), focusing on milk prediction (29%), lameness detection (26%), and mastitis detection (13%).
• Predictive analytics dominate (87%), with limited use of prescriptive analytics (3%). Historical data (70%) is preferred over real-time data (25%).
• Artificial neural networks (47%) and convolutional neural networks (24%) are prevalent, showing good accuracy but high variability. Simulation tools (4%) and stochastic models (5%) are underutilized.

Conclusions:

• Artificial neural networks offer significant potential for data analytics in dairy farming.
• Integrating optimization-simulation methods with artificial neural networks can enhance future decision-making tools for bovine milk production.
• Further research should explore the combined power of advanced analytics and simulation for dairy farm management.