Related Experiment Video
Updated: Jul 12, 2026
07:38
Milk Collection in the Rat Using Capillary Tubes and Estimation of Milk Fat Content by Creamatocrit
Published on: December 16, 2015
A computerized mastitis decision aid using farm-based records: an artificial neural network approach.
1Department of Dairy and Animal Science, The Pennsylvania State University, University Park 16802, USA. CWH3@psu.edu
Journal of Dairy Science
|May 3, 2000
Summary
A new computer model uses artificial neural networks to classify bacterial causes of mastitis in dairy cows. This tool aids in interpreting Dairy Herd Improvement Association (DHIA) data for better herd management.
Area of Science:
- Veterinary Science
- Computer Science
- Animal Science
Background:
- Mastitis is a significant concern in dairy herds, impacting animal health and milk production.
- Accurate identification of bacterial pathogens causing mastitis is crucial for effective treatment and control strategies.
- Dairy Herd Improvement Association (DHIA) data offers a valuable resource for herd health monitoring but requires sophisticated analysis for detailed insights.
Purpose of the Study:
- To develop and evaluate a computer module utilizing artificial neural networks (ANNs) for classifying bacterial causes of mastitis in dairy herds.
- To assess the potential of this diagnostic module in aiding the interpretation of DHIA data for herd consultants and record processing centers.
- To compare the performance of the ANN model against linear discriminant analysis in predicting mastitis etiology.
Main Methods:
- Collected field survey data, herd management practices, quarter milk samples, and monthly DHIA data from Pennsylvania dairy herds.
- Developed an artificial neural network model trained on this data to discriminate between four categories of mastitis-causing bacteria.
- Validated the trained ANN model using new DHIA and management data from selected cow groups and untested herds.
Main Results:
- The ANN model achieved diagnostic probabilities ranging from 57% to 71% for classifying the bacteriologic status of mastitis.
- Model performance was optimal in herds with a higher prevalence of minor and contagious mastitis pathogens.
- Linear discriminant analysis yielded less successful prediction results, with a range of 42% to 57%.
Conclusions:
- The developed artificial neural network computer module shows promise for classifying bacterial causes of mastitis using DHIA and management data.
- This diagnostic tool has the potential to enhance the interpretation of DHIA data, supporting herd health management decisions.
- ANNs offer a superior approach compared to linear discriminant analysis for this specific diagnostic task in dairy herds.
