MilkBot(R) will provide enabling technology for a wide variety of methods applicable to dairy research and management. Increasing the sensitivity of our ability to detect changes in milk production may lower the sample size needed for research projects, or improve the ability to detect a response. In field situations the same tools will help in detection and quantification of response to intentional or unintentional changes in management or environment. OBJECTIVES: The objective of this project is to develop a low level statistical modeling framework and tool set that will allow any analysis that sees milk production as a dependent variable to measure response more precisely, both with regard to total production and how that production is distributed over lactations. The methods we are developing are modular and extendable by design. APPROACH: We have developed a novel mechanistic lactation model, nonlinear curve-fitting methods, and computer software for exploration of dairy production data. This package enables, for the first time, a practical method for precise quantification of shape and magnitude of individual lactation curves, and a broad range of sophisticated analytic methods to detect and quantify changes in milk production. Milk production is the most important economic determinant in commercial herds and also a sensitive measure of general cow health, so improved ability to monitor and quantify changes in lactation has great potential to improve almost every aspect of dairying. We will test this tool set in a variety of field situations and against multiple data sets to prove the feasibility of the approach in solving problems in dairy research and management. This will include analyses using a broad sample of dairy herds to identify correlative patterns and normal variability, and initial devlopment of the knowledge base that is necessary before the methodology can be credible or viable commercially
