Fig. 4From: Role of data measurement characteristics in the accurate detection of Parkinson’s disease symptoms using wearable sensorsFeature set performance and computation time. Combinations of feature categories for the symptom models, ordered by total computation time. Model performance (AUROC) is computed for each model type and feature category combination. Generally, there is a trade-off between feature complexity and model performance, as more comprehensive feature sets improve AUROC but take longer to computeBack to article page