Chicken and Chips: How Chickens can Benefit from Machine Learning
11th October 2011, 16:00, ALT
Prof. Kenton Morgan
Dept Musculoskeletal Biology
Institute of Ageing and Chronic Diseases
University of Liverpool
In collaboration with DIMACS, Rutgers University New Jersey we have been exploring the epidemiological application of Support Vector Machine learning to improve the health of broiler chickens.
Epidemiology is the study of the frequency, distribution and determinants of disease (and health) in populations. Observational epidemiology analyses the patterns of a disease to identify the variables associated with its occurrence (risk factors), in time and space. Logisitic regression models are the industry standard for doing this.
In this BBSRC funded project Support Vector Machine (SVM) is applied to broiler flock management data to identify the risk factors for hock burn - the chicken equivalent of bed sores.
This is the first time that machine learning has been applied to disease prevention in any species. It has the potential to improve the health and welfare of broiler chickens worldwide and also to be applied to human healthcare.
We are looking for collaborators in Computer Science at Liverpool to progress this with our partners at DIMACS.