Table 1 Results using the original data.
From: Multimodal deep learning for dementia classification using text and audio
Original | Accuracy | Precision | Recall | F1-score | AUROC |
|---|---|---|---|---|---|
Audio | 0.6484 ± 0.008 | 0.593 ± 0.019 | 0.4425 ± 0.063 | 0.5039 ± 0.032 | 0.7085 ± 0.011 |
Text | 0.691 ± 0.034 | 0.6484 ± 0.07 | 0.6299 ± 0.127 | 0.62765 ± 0.027 | 0.7638 ± 0.046 |
Audio \(+\) Time | 0.6123 ± 0.006 | 0.5565 ± 0.022 | 0.3286 ± 0.039 | 0.4115 ± 0.026 | 0.6517 ± 0.011 |
Text \(+\) Time | 0.6909 ± 0.013 | 0.6537 ± 0.036 | 0.5566 ± 0.4463 | 0.5995 ± 0.022 | 0.7647 ± 0.018 |
Audio \(+\) Text | 0.6731 ± 0.04 | 0.5958 ± 0.047 | 0.6852 ± 0.097 | 0.6341 ± 0.048 | 0.7448 ± 0.045 |
Audio \(+\) Text \(+\) Time | 0.6539 ± 0.031 | 0.5874 ± 0.07 | 0.5301 ± 0.138 | 0.55 ± 0.087 | 0.7161 ± 0.043 |
