Development to Ensure of the Result Reliability of Production Indicators in the Milk Recording During its Computerization

Abstract

Milk recording (MR) is an essential breeder measure. Results are important for inheritance check. The occurrence of errors in the data may compromise the efficiency of breeding of dairy cows. The aim was possibility to reduce the incidence of MR database errors. Analyses of frequency distribution of MR data deviations from different sources and estimations of limits of difference acceptability in milk recording were performed. The results of MR control days of flowmeter in parlor (DMY) were paired to the AVG7 results (average for 7 days) from the same flowmeter (n = 16,247, original recordings of complete lactations). The individual differences in milk yield indicators were calculated between successive MR control days (DMY – R, monthly interval, the reference value (R) = previous DMY) for MR data file. A statistically significant correlation coefficient (AVG7 and DMY) was 0.935 (P < 0.001) and was higher in comparison to the previous assessment under AMS conditions (automatic milking system; 0.898; P < 0.001). This means that 87.3% of the variability in the ​​milk yield values for MR (DMY) can be explained by variations in the AVG7 values and vice versa. Difference tests confirmed significant differences (P < 0.001) 0.76 and 0.55 kg between DMY (in MR) and AVG7 for original and also refined data file. Mentioned differences, although statistically significant, correspond only to 2.96 and 2.15% relatively. The use of multi–day milk yield average from the electronic flowmeter is an equivalent alternative to the use of record from one MR control day. Results are used in MR practice.