Figure 5.

Predicting udder health status outcome using random forest model by using a rumen bacteria dataset of 334 dairy cows [24]. (A) Receiver operating characteristic (ROC) curve together with the confusion matrix of the performance of the random forest model with the selected 24 genera. (B) 24 genera selected in the random forest model. AUC, area under the curve; HE, cows with somatic cell counts (SCCs) lower than 500 × 10³ cells/mL in milk or classified as healthy according to the random forest model; MA, cows with SCCs over 500 × 10³ cells/mL in milk or classified as mastitis according to the random forest model; MDA, mean decrease in accuracy.