Fitness of four different mathematical models to the lactation curve of Brown Swiss cows in Konya Province of Turkey

The aim of this study was to investigate the fitness of Incomplete Gamma (WD), Exponential (WIL), Mixed Log (MIL) and Polynomial Regression (AS) models to the lactation curve of Brown Swiss Cows. Data were collected from 143 Brown Swiss cows raised on the Alt?nova State Farm in Konya Province, Turkey. Milk yield was recorded monthly, and milk records were started at the third week of lactation (mean = 16.9 day, SD = 0.7). Total milk yields estimated by the four models were very close to real total milk yield. The models were found to be adequate for estimation of milk yield. The MIL model underestimated the peak yield significantly. The differences between peak yields of the models and real peak yields were not significant and ranged from 27.70 to 29.01 L. All models forecasted peak time earlier than real peak time. The differences for the persistency values of the four models were significant. The AS model's persistency value was nearly equal to the real persistency value (77.56 vs. 77.59%). R2 values of the models changed from 86.05 to 97.95%. The AS model gave the best R2 and the least MSPE values. Consequently, the AS model showed the best fit to the lactation data of Brown Swiss cows and allowed a suitable definition of the lactation curve.