Detection of abnormal recordings in Irish milk recorded data

Abstract

The objective of this study was to detect abnormal recordings of milk yield, fat concentration and protein concentration in Irish milk-recorded data. The data consisted of 14,956 records from both commercial and experimental herds with 92% of the recordings recorded manually and the remainder recorded electronically. The method used in this paper was a modified version of the method employed by the Animal Improvement Programs Laboratory in Maryland, USA and conformed with the guidelines outlined by the International Committee of Animal Recording. The results illustrate the effectiveness of detecting abnormal recordings in Irish milk records. The method described in this paper, defines the upper and lower limits for each production trait and these limits along with the slope parameters were used to determine if a recording was abnormal or not. Three percent of milk yield recordings, 5% of fat concentration recordings and less than 1% of protein concentration recordings were found to be abnormal. The proportion of values declared abnormal in manually recorded and electronically recorded data were examined and found to be significantly different for fat concentration.