Figure 7From: What variables are important in predicting bovine viral diarrhea virus? A random forest approach Evaluation of models performance for independent test data. A) RF, SVM and GBM were also compared using an independent test set, which corresponds to the 20% portion of data that was not used in the training and cross-validation procedures. According to the ROC curves, the GBM classifier outperforms RF and SVM. B) Relative to the logistic regression, a traditional statistical approach used for the analysis of risk factors, RF achieved a more robust performance.Back to article page