The introduction of Automated Milking Systems (AMSs), or milking robots, represented a significant advance- ment in dairy farming techniques. AMSs enable real-time monitoring of udder health and milk quality during each milking episode, which provides a wealth of data that can be utilized to optimize herd management practices. ML algorithms are well-suited for handling large and multi-dimensional datasets, making them a valuable tool for analyzing the vast amount of data generated by AMSs. This study introduces a novel approach to characterize the milk productivity of Holstein Friesians cows milked by AMSs during individual lactation periods and evaluate their stability over time. Four unsupervised ML clustering algorithms were employed to cluster the cows within each lactation period, and a merging index was proposed to combine the clustering results. The dairy cows were grouped into clusters based on their productivity, and the stability of these Pro- ductivity Groups (PGs) over time was analyzed. The PGs were found to be weakly stable over time, indicating that selecting cows for insemination based solely on their present or past lactation productivity may not be the most effective strategy. In addition, the results revealed that the High Productivity Group exhibited lower levels of protein, fat, and lactose content in the milk. The proposed methodology was demonstrated using data from one farm with dairy cows that exclusively uses the AMS, however, it can be applied to any context and dataset in which a multi-algorithm clustering analysis is suitable, including data from conventional milking parlors. Understanding milk productivity and its factors in future lactation periods is essential for effective herd management. A comprehensive long-term analysis is of significant importance for the zootechnical sector as it could assists farmers in selecting cows for insemination and making decisions on which ones to retain for future lactation periods.
Multi-algorithm clustering analysis for characterizing cow productivity on automatic milking systems over lactation periods
Karina Brotto RebuliFirst
;Laura Ozella
;Mario GiacobiniLast
2023-01-01
Abstract
The introduction of Automated Milking Systems (AMSs), or milking robots, represented a significant advance- ment in dairy farming techniques. AMSs enable real-time monitoring of udder health and milk quality during each milking episode, which provides a wealth of data that can be utilized to optimize herd management practices. ML algorithms are well-suited for handling large and multi-dimensional datasets, making them a valuable tool for analyzing the vast amount of data generated by AMSs. This study introduces a novel approach to characterize the milk productivity of Holstein Friesians cows milked by AMSs during individual lactation periods and evaluate their stability over time. Four unsupervised ML clustering algorithms were employed to cluster the cows within each lactation period, and a merging index was proposed to combine the clustering results. The dairy cows were grouped into clusters based on their productivity, and the stability of these Pro- ductivity Groups (PGs) over time was analyzed. The PGs were found to be weakly stable over time, indicating that selecting cows for insemination based solely on their present or past lactation productivity may not be the most effective strategy. In addition, the results revealed that the High Productivity Group exhibited lower levels of protein, fat, and lactose content in the milk. The proposed methodology was demonstrated using data from one farm with dairy cows that exclusively uses the AMS, however, it can be applied to any context and dataset in which a multi-algorithm clustering analysis is suitable, including data from conventional milking parlors. Understanding milk productivity and its factors in future lactation periods is essential for effective herd management. A comprehensive long-term analysis is of significant importance for the zootechnical sector as it could assists farmers in selecting cows for insemination and making decisions on which ones to retain for future lactation periods.File | Dimensione | Formato | |
---|---|---|---|
Brotto_Rebuli_2023.pdf
Accesso aperto
Tipo di file:
PDF EDITORIALE
Dimensione
1.12 MB
Formato
Adobe PDF
|
1.12 MB | Adobe PDF | Visualizza/Apri |
I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.