Fig 1.
The whole process in the present study.
Fig 2.
An illustration of the difference between audio-based prediction and participant-based prediction.
Table 1.
Parameter values in each model.
Table 2.
Descriptive statistics of demographic and vocal data for participants of each group.
Fig 3.
Receiver operating characteristic (ROC) curves in the models for predicting AD.
RF = Random Forest; XGBoost = Extreme Gradient Boosting; LR = Logistic Regression.
Table 3.
Predictive performance of the models built for each audio file for predicting AD.
Table 4.
Predictive performance of the models built for each participant for predicting AD.