Modeling of lactation milk yield with spline regression method

In this thesis, modeling of lactation milk yields by means of the Spline Regression method was examined comparatively with a classical model (Wood). The thesis data were obtained from the Izmir Cattle Breeders' Association. 24960 milk yield record of 832 Holstein-Friesian dairy cows was used. From the available data, both general averages and individual modeling were performed separately for the animals in lactates 1, 2 and 3. 7 different Cubic Spline Regression models were compared with Wood model. According to the results obtained, the Spline Regression method showed successful results in individual curve modeling. In addition, it is concluded that lactation milk yields do not form a standard cubic curve and can be a successful modeling for curves with undulating yields.