Some error occured while loading the Quick View. Please close the Quick View and try reloading the page.
Given developments in sensor technologies and the huge amounts of data that are becoming available, milk yield monitoring is now broadening its potential to provide holistic data-driven decision support in dairy herd management, including in managing mastitis. This chapter first introduces mastitis, milk quality and the drivers behind automatic detection, and then describes the main sensor technologies used during milking. It subsequently presents a framework that separates filtering, transformation and classification steps, and uses this framework to discuss both more classical model-based approaches and newer data-driven and machine learning methods in data analysis. Later sections address practical aspects such as defining mastitis events, evaluating model performance and implementing detection in commercial systems, before turning to economic and welfare implications and future trends. Throughout the chapter highlights recurring methodological challenges and areas where standardization or more transparent reporting would be particularly helpful.
- 1 Introduction
- 2 Clinical and subclinical mastitis: effects on milk yield and quality
- 3 Detecting mastitis: current challenges
- 4 In-line and on-line sensors used for mastitis detection
- 5 Characteristics of milking sensor data streams
- 6 Data processing: filtering, transformation and classification
- 7 Classical approaches to data processing and analysis
- 8 Data-driven and machine learning approaches
- 9 Data filtering and transformation techniques
- 10 Other issues: confirming mastitis, reference windows, performance metrics and study design
- 11 Applying sensor-based mastitis detection in commercial systems
- 12 Conclusion and future trends in research
- 13 Acknowledgements and declaration of interests
- 14 References