Fig. 1: AUC-ROC curves demonstrate XGBoost model performance for detecting cognitive impairment (all patients vs. cognitively unimpaired controls) across various methods. | npj Dementia

Fig. 1: AUC-ROC curves demonstrate XGBoost model performance for detecting cognitive impairment (all patients vs. cognitively unimpaired controls) across various methods.

From: Voiceprints of cognitive impairment: analyzing digital voice for early detection of Alzheimer’s and related dementias

Fig. 1: AUC-ROC curves demonstrate XGBoost model performance for detecting cognitive impairment (all patients vs. cognitively unimpaired controls) across various methods.

The highest numerical value of performance was based on a model using speech and language features, while the model based on demographic data was comparable to one based on a randomized variable.

Back to article page