Research Article
Speech Deception Detection Based on EMD and Temporal Neural Network
Table 2
Deception detection results of different speech sampling rates.
| | | 4 kHz | 8 kHz | 16 kHz | | Acc (%) | F1 (%) | Acc (%) | F1 (%) | Acc (%) | F1 (%) |
| | KNN | 75.1 | 74.3 | 77.3 | 77.0 | 78.4 | 78.3 | | SVM | 74.2 | 74.0 | 77.5 | 77.2 | 79.5 | 79.5 | | Decision tree | 68.6 | 68.1 | 70.0 | 70.7 | 73.7 | 73.9 | | Random forest | 75.5 | 76.0 | 78.6 | 79.0 | 79.5 | 79.3 | | Naïve Bayes | 66.6 | 66.5 | 68.8 | 69.1 | 69.0 | 69.2 | | AdaBoost | 70.1 | 71.0 | 72.8 | 73.1 | 73.6 | 73.5 | | Ensemble learning | 77.2 | 77.0 | 79.2 | 79.5 | 79.4 | 79.4 | | MLP | 70.9 | 70.2 | 72.1 | 72.3 | 73.1 | 72.7 | | RNN | 58.2 | 59.0 | 64.7 | 64.2 | 64.3 | 65.1 | | Autoencoder | 65.1 | 65.5 | 67.4 | 68.1 | 68.9 | 68.8 |
|
|
