¹National University of Life and Environmental Sciences of Ukraine, Faculty of Livestock Raising and Water Bioresources, Department of Animal Feeding and Feed Technology named after Pshenychnyi, 03041 Kyiv, UKRAINE
²National University of Life and Environmental Sciences of Ukraine, Faculty of Design and Engineering, Department of Machines and Equipment Design, 03041 Kyiv, UKRAINE
³National University of Life and Environmental Sciences of Ukraine, Faculty of Livestock Raising and Water Bioresources, Department of Animal Genetics, Breeding and Animal Biotechnology, 03041 Kyiv, UKRAINE DOI : 10.9775/kvfd.2022.28672 Aim of this work was to study effectiveness of prediction of cows" milk productivity (yield, fat and protein content) using artificial neural network (ANN) technology for data sets with missing values. Four variants of data sets, which consist of different numbers of monthly test-day milk records (MTDMRs) were chosen. Calculated milk productivity using Test Interval Method (TIM) was chosen as control value. Obtained results showed that milk productivity can be predicted using ANN even if missing data occurs. Rank correlation coefficients between control and predicted results were 0.918, 0.949, and 0.852 for milk yield, protein, and fat content, respectively Keywords : Artificial Neural Networks, Dairy Productivity Prediction