Figure 3
From: Machine learning random forest for predicting oncosomatic variant NGS analysis
ROC curve for the validation set. Thirty cross-validations were assessed using 7 functional features: chromosome (Chr), position (POS), exon, variant allele frequency (Freq), minor allele frequency (MAF), coverage and amino acid change (protdesc) for the model machine learning random forest (MLRF) for oncosomatic variants. Model performance was evaluated using the receiver operating characteristic (ROC) curve. The area under the curve is denoted AUC. AUC obtained was 0.99. A large area under the curve was observed. The model could then determine whether a variant was benign or pathogenic with a minimal error rate.
