Salmonella Dublin (S. Dublin) is a costly infection for dairy cows, potentially lethal to humans. Surveillance is based on bulk tank milk (BTM) antibody measurements, taken each quarter of the year. Herds are classified as Status 1- likely free of S. Dublin, or Status 2 – likely infected with S. Dublin, based on present /recent characteristics, but not actual S. Dublin detection. We develop a predictive model based on characteristics from last quarter, using on registry data for 2001-2007 for 9387 herds in Denmark . Only 2004-2007 data modeled due to data contamination.
Main Research Area:
24th Nordic Conference in Mathematical Statistics, 2012