Fig. 4
From: Machine Learning for detection of viral sequences in human metagenomic datasets
ROC curves of the metagenomics model for each LOEO cross-validation fold and their micro and macro averages. The grey lines depict the ROC curve in each leave-one-experiment-out cross-validation fold. The red line is the macro-averaged curve and blue line the micro-averaged curve. Green line shows the performance of a random model. The averaged results are clearly above the diagonal, meaning the models perform a much better than a random guess