COMPARISON OF DIFFERENT METHODS OF PREDICTING 305-DAY MILK-YIELD USING MEANS CALCULATED FROM WITHIN-HERD LACTATION CURVES
COMPARISON OF DIFFERENT METHODS OF PREDICTING 305-DAY MILK-YIELD USING MEANS CALCULATED FROM WITHIN-HERD LACTATION CURVES
复制标题
DOI:
10.1016/0301-6226(87)90049-2
复制
发表时间:
1987-07-01
期刊:
影响因子:
--
通讯作者:
WILMINK, JBM
中科院分区:
文献类型:
--
作者:
WILMINK, JBM
The prediction of the unknown par to fa 305-day lactation is usually based on regression models. For these models the relation between known test-day yields and remaining lactation yield as well as their means within classes of environmental effects need to be known. Owing to environmental changes, means may vary severely. In this study a procedure for estimation of the means is presented and single regression (SR), multiple regression (MR), and factor analysis (FA) models for prediction are compared. Both MR and FA use all known test-day yields as information sources. Method SR uses the last known test-day yield. Lactation curves were estimated per herd on all records, adjusted for age and season of calving. By back-adjustment for age and season for calving, lactation cures were obtained for a group of cows pertaining to a certain class of herd-age-month of calving. Means necessary for prediction were calculated from these lactation curves (Type II). Correlations with means, which were directly calculated in the data set (Type I), ranged from 0.61 to 0.78 for test-days and from 0.51 to 0.69 for remaining lactation milk yield. The average difference between Types I and II means was almost zero. Using Type II means, no differences were found between SR, MR and FA. The correlation between predicted and realized 305-day milk yield was 0.87, 0.88 and 0.88 respectively for parity 1, 2 and > 2, when the last test-day yield was known at 50 days postpartum. The use of Type II means resulted in a more precise adjustment.