A decision support system for herd health management for dairy farms

Abstract

Industrial dairy farms boast highly advanced health monitoring and disease diagnosis systems. But without easily accessible, user-friendly web platforms for real-time decision-making, most dairy farmers cannot proactively manage herd health management and optimize treatments based on disease prediction and prevention. To bridge this gap, we have developed a web application of a Decision support system (DSS) for dairy health management based on machine learning. The system architecture combines a Flask backend with a React frontend and scalable cloud data storage and includes preprocessing, data integration, predictive modelling, and cost analysis. DSS forecasts herd diseases with an accuracy 6.66 mean absolute error and 2.35 median absolute deviation across predictions. Its core predictive capabilities rely on long short-term memory (LSTM) neural networks to forecast disease progression from historical records and on a linear trend model to project cuts in treatment costs. The system calculates medication dosages and cost per disease, streamlines supplier selection, and simulates various treatment scenarios, thereby identifying high-cost diseases with potential savings. In other words, this DSS application processes disease and treatment data by incorporating veterinary records into advanced data analytics and neural networks, thereby predicting diseases, optimizing disease prevention and treatment strategies, and reducing costs. As such, this DSS application provides dairy farmers with a tool for strategic decision-making, veterinary treatment planning, and cost-effective disease management towards improving animal welfare and increasing milk yield.

Keywords

• dairy cows
• disease monitoring
• neural networks, predictive analysis
• treatment optimisation
• web applications