Detection of Subclinical Mastitis from On-Line Milking Parlor Data

摘要: Abstract A model, based on automatically collected data, was developed for detection of subclinical mastitis. The logistic regression model the following variables: milk electrical conductivity, production, parity, and DIM. Subclinical mastitis defined as a minimal period 1 wk in which SCC >500 × 10 3 cells/ml. In contrast, periods were healthy if resulting had sensitivity 55% specificity 90% individual milkings. For 14 milkings, 54% 92% when threshold that >6 conductivity signals high SCC. Based these test characteristics, could be used an initial screening tool herd with incidence Cows signal would have higher probability being diseased than total population. such herds. separation from signaled cows might possible management strategy to reduce bulk tank.