Prediction of body condition score in dairy cows using routine electronic data capture
Body condition scoring is a widely accepted practical herd management technique used for assessing the impact of changes in energy status of adult dairy cows. Achieving optimum cow condition scores throughout lactation is necessary to maximise milk production, reproductive performance and health sta...
| Main Author: | |
|---|---|
| Format: | Thesis (University of Nottingham only) |
| Language: | English |
| Published: |
2025
|
| Subjects: | |
| Online Access: | https://eprints.nottingham.ac.uk/80776/ |
| _version_ | 1848801270911991808 |
|---|---|
| author | Nelson, Rebecca |
| author_facet | Nelson, Rebecca |
| author_sort | Nelson, Rebecca |
| building | Nottingham Research Data Repository |
| collection | Online Access |
| description | Body condition scoring is a widely accepted practical herd management technique used for assessing the impact of changes in energy status of adult dairy cows. Achieving optimum cow condition scores throughout lactation is necessary to maximise milk production, reproductive performance and health status and indirectly optimising overall farm profitability. However, the traditional visual body condition scoring system is subjective and can result in substantial inter and intra scorer variability. An objective and automated method of body condition scoring would be a valuable tool for the management of modern dairy herds.
The objective of this study was to predict cow body condition score from measurements of body weight and body stature, taking into account confounding factors such as rumen fill and gestation stage and days in milk. Body weight, body condition score and body stature measurements of Holstein-Friesian dairy cows (n = 68) were obtained 2-4 weeks prior to expected calving date and repeat measurements of body weight and body condition score taken monthly thereafter. Body stature measurements included heart girth, belly girth, diagonal and horizontal length, hip width at the level of the tuber coxae and the pin bones, leg circumference and withers height. These were not repeated and assumed to stay the same throughout for all cows. Data analysis comprised fitting multiple predictive algorithms and conducting cross validation to identify the best model. The Support Vector Machine model with a polynomial kernel was the best performing model, with the mean absolute error (MAE) in cross validation being 0.3. When body condition scores were rounded up to the nearest 0.5 and compared with the observed conditions score, the model allowed us to differentiate between 2, 3, 3.5 and 4 however there was no differentiation between 2.5 and 3. The results suggest that the relationship between cow body weight, stature and body condition score is not straightforward or sufficiently similar between cows to allow a generalised prediction to be made. Additional research is needed to explain and understand these results, including exploring the ratio of subcutaneous and visceral fat and the relationship to body condition score in different animals. |
| first_indexed | 2025-11-14T21:04:48Z |
| format | Thesis (University of Nottingham only) |
| id | nottingham-80776 |
| institution | University of Nottingham Malaysia Campus |
| institution_category | Local University |
| language | English |
| last_indexed | 2025-11-14T21:04:48Z |
| publishDate | 2025 |
| recordtype | eprints |
| repository_type | Digital Repository |
| spelling | nottingham-807762025-07-24T04:40:13Z https://eprints.nottingham.ac.uk/80776/ Prediction of body condition score in dairy cows using routine electronic data capture Nelson, Rebecca Body condition scoring is a widely accepted practical herd management technique used for assessing the impact of changes in energy status of adult dairy cows. Achieving optimum cow condition scores throughout lactation is necessary to maximise milk production, reproductive performance and health status and indirectly optimising overall farm profitability. However, the traditional visual body condition scoring system is subjective and can result in substantial inter and intra scorer variability. An objective and automated method of body condition scoring would be a valuable tool for the management of modern dairy herds. The objective of this study was to predict cow body condition score from measurements of body weight and body stature, taking into account confounding factors such as rumen fill and gestation stage and days in milk. Body weight, body condition score and body stature measurements of Holstein-Friesian dairy cows (n = 68) were obtained 2-4 weeks prior to expected calving date and repeat measurements of body weight and body condition score taken monthly thereafter. Body stature measurements included heart girth, belly girth, diagonal and horizontal length, hip width at the level of the tuber coxae and the pin bones, leg circumference and withers height. These were not repeated and assumed to stay the same throughout for all cows. Data analysis comprised fitting multiple predictive algorithms and conducting cross validation to identify the best model. The Support Vector Machine model with a polynomial kernel was the best performing model, with the mean absolute error (MAE) in cross validation being 0.3. When body condition scores were rounded up to the nearest 0.5 and compared with the observed conditions score, the model allowed us to differentiate between 2, 3, 3.5 and 4 however there was no differentiation between 2.5 and 3. The results suggest that the relationship between cow body weight, stature and body condition score is not straightforward or sufficiently similar between cows to allow a generalised prediction to be made. Additional research is needed to explain and understand these results, including exploring the ratio of subcutaneous and visceral fat and the relationship to body condition score in different animals. 2025-07-24 Thesis (University of Nottingham only) NonPeerReviewed application/pdf en cc_by https://eprints.nottingham.ac.uk/80776/1/Nelson%2C%20Rebecca%2C%2020128304%2C%20Corrections.pdf Nelson, Rebecca (2025) Prediction of body condition score in dairy cows using routine electronic data capture. MVM thesis, University of Nottingham. Cow condition; Body condition scoring; Herd management technique; Automated methods |
| spellingShingle | Cow condition; Body condition scoring; Herd management technique; Automated methods Nelson, Rebecca Prediction of body condition score in dairy cows using routine electronic data capture |
| title | Prediction of body condition score in dairy cows using routine electronic data capture |
| title_full | Prediction of body condition score in dairy cows using routine electronic data capture |
| title_fullStr | Prediction of body condition score in dairy cows using routine electronic data capture |
| title_full_unstemmed | Prediction of body condition score in dairy cows using routine electronic data capture |
| title_short | Prediction of body condition score in dairy cows using routine electronic data capture |
| title_sort | prediction of body condition score in dairy cows using routine electronic data capture |
| topic | Cow condition; Body condition scoring; Herd management technique; Automated methods |
| url | https://eprints.nottingham.ac.uk/80776/ |