MACHINE LEARNING FOR SPOILAGE PREDICTION IN DAIRY SUPPLY CHAINS: A REVIEW

Muskan Siddiqua, Pratiksha Yadav, Samina A Amaravati, Sharath Kumar N, Nalini Chitta

The dairy supply chain is essential for ensuring the delivery of safe and high-quality milk products, but it faces considerable challenges due to milk’s perishable nature and its vulnerability to factors such as temperature fluctuations, humidity, and transportation delays. Conventional monitoring approaches mainly depend on physical sensors and typically offer only real-time or past data, without the capability to forecast spoilage or estimate remaining shelf life. This review examines the use of advanced machine learning techniques—such as Artificial Neural Networks (ANN), Random Forest (RF), Support Vector Machines (SVM), and Extreme Gradient Boosting (XGBoost)—for intelligent milk quality assessment. These models are effective in identifying complex patterns within sensor data, thereby enhancing prediction accuracy and enabling proactive decision-making. Among them, XGBoost stands out due to its high accuracy, computational efficiency, and robustness, making it well-suited for real-time applications in dairy supply chains. In addition, the study discusses the integration of Internet of Things (IoT), Digital Twin systems, and virtual sensing to improve monitoring and support predictive analytics. It also highlights current research gaps and suggests a framework for building a smart, data-driven milk quality monitoring system. Overall, this work provides valuable insights into minimizing spoilage, improving operational efficiency, and maintaining quality standards in dairy supply chains through machine learning-driven solutions.

Keywords –Artificial Neural Network (ANN), Dairy Supply Chain, Digital Twin, IoT, Machine Learning, Milk Quality Monitoring, Predictive Analytics, Random Forest, Shelf Life Prediction, Spoilage Detection, Support Vector Machine (SVM), XGBoost.