Fig. 4From: Machine learning approach for Migraine Aura Complexity Score prediction based on magnetic resonance imaging dataPredicted vs. real average MACS for SVM algorithm and wrapper feature selection subset. The x-axis shows predicted average MACS and the y-axis real average MACS scores. The prediction is performed using the SVM algorithm and features that were derived with the wrapper feature selection method. Each black dot represents one subject. This model achieved R2 = 0.8046, MAE = 0.871, and RMSE = 1.0827Back to article page