MilkBot lactation API (1.4.0)

This service uses the MilkBot fitting engine to fit the MilkBot model to lactation data. The returned parameter set can be substituted into the MilkBot equation

((1 - E^((offset - t)/ramp)/2)*scale)/E^(decay X t)

to generate predicted milk production at any time in the lactation, where t is time since calving (in days, DIM=0 at calving) and E is Euler's number.

Simple formulas described at the support URL can calculate persistence, M305, or cumulative milk between dates. For a fuller description of the model see Quantifying inter-group variability in lactation curve shape and magnitude with the MilkBot® lactation model

To use this service, a valid API key must be in a X-API-Key header of each request. For a free API Key, contact Jim Ehrlich jehrlich@MilkBot.com

The fitting process uses Bayesian prior expectations as the expected means and standar deviation of parameter values as well as expected sd of fitting error. If priors are provided, they will be used. Otherwise default values based on breed and parity will be used. Use GET /priorsTree to see default priors. The PriorsTree uses a discriminator at each node to value lactations to an appropriate Priors.

• Priors It is easy to generate a customized PriorsTree by fitting a training set of lactations, then using the averages from that set as the priors for future fittings.
• Options The optional Options object may be included with a FittingJob to control what data is returned.
• Milk Units Milk values may be in Kg (the default) or pounds (by specifying "milkUnit": "Lb")

• Fitting Jobs may exceed the maximum memory available to the fitting function. If that becomes a problem please contact us and we will consider increasing the memory pool. A Workaround is to fit in small batches or as individual lactations.