What is mastitis, and why do we need a better way to predict it?
Mastitis is an inflammation of the udder that can be caused by a range of pathogens or injuries. It’s painful for the cow and also a persistent, costly and widespread challenge for dairy farmers. It impacts farm profitability, animal welfare and sustainable production.
Although there are in-machine predictors and a lot of papers on mastitis prediction, they often focus on herd-level statistics. They emphasize electrical conductivity of the milk as a main predictive feature since mastitis causes vascular permeability that lets sodium ions and chloride leak from the blood into the milk, which raises conductivity. That is problematic because that type of reading can be finicky to maintain. Another way to predict subclinical cases is somatic cell count testing, which can be expensive for the farmer.
How will your predictive model get around these limitations?
Our implementation takes a combination of milk parlor data (such as the flow of milk over time or the amount of milk produced in total), weather data (such as the maximum temperature per day or the maximum humidity), diagnostic and herd health data (such as the frequency of a cow’s mastitis infections in the past or what pathogen caused previous mastitis episodes), and geospatial data that identifies the cow’s location within the barn at a given point in time.
All this information feeds into our model to help us make meaningful predictions and then take that data and make powerful insights.
Your data comes from CAST, which is creating an ecosystem of networked technologies and techniques for modern farms. What kind of data did they give you?
We're using data gathered by CAST from their main farm, the Cornell Dairy Research and Education Center (CDREC). There’s a lot of different kinds. We have herd health data from Dairy Comp, a dairy herd management software. We also have diagnostic data from the Cornell Animal Health Diagnostic Center for CDREC. Then, we have milk parlor data from Delpro, which is a dairy and milking management platform. Delpro gives us all kinds of measurements of milk output and composition. And we have geospatial ear-tag data, which shows the location of each cow within the barn at a millisecond level.
Source : cornell.edu