摘要：The aim of this research was to define methods for early prediction (based on I. milk control record) of lactation flow in Holstein heifers as well as to choose optimal one in terms of prediction fit and application simplicity. Total of 304,569 daily yield records automatically recorded on a 1,136 first lactation Holstein cows, from March 2003 till August 2008., were included in analysis. According to the test date, calving date, the age at first calving, lactation stage when I. milk control occurred and to the average milk yield in first 25th, T1 (and 25th-45th, T2) lactation days, measuring monthcalving month-age-production-time-period subgroups were formed. The parameters of analysed nonlinear and linear methods were estimated for each defined subgroup. As models evaluation measures, adjusted coefficient of determination, and average and standard deviation of error were used. Considering obtained results, in terms of total variance explanation (R2 adj), the nonlinear Wood’s method showed superiority above the linear ones (Wilmink’s, Ali-Schaeffer’s and Guo-Swalve’s method) in both time-period subgroups (T1 - 97.5 % of explained variability; T2 - 98.1 % of explained variability). Regarding the evaluation measures based on prediction error amount (eavg±eSD), the lowest average error of daily milk yield prediction (less than 0.005 kg/day), as well as of lactation milk yield prediction (less than 50 kg/lactation (T1 time-period subgroup) and less than 30 kg/lactation (T2 time-period subgroup)); were determined when Wood’s nonlinear prediction method were applied. Obtained results indicate that estimated Wood’s regression parameters could be used in routine work for early prediction of Holstein heifer’s lactation flow.
