Utilization of boosted classification trees for the detection of cows with conception difficulties

The present study is planned to apply boosted classification trees for the detection of cows with potential conception problems. Nine hundred and eighteen artificial insemination records from Polish Holstein-Friesian Black-and-White cows were included. Each record consisted of nine predictor variables. The output variable was a conception difficulty class (good or poor). Sensitivity, specificity and accuracy on the test set were 82.93%, 84.46% and 83.91%, respectively. The most influential predictors of conception difficulty included calving interval, gestation length, body condition score, fat and protein content and age. Boosted classification trees could be applied as an on-farm decision support tool to identify cows with conception problems.