Research Article
TASG: An Augmented Classification Method for Impersonal HAR
Table 9
Performance comparison with and without TASG on the HASC-PAC2016 dataset.
| ā | Accuracy | Precision | Recall | F1 score |
| SVM | 0.5981 | 0.5966 | 0.5937 | 0.5951 | TASG-SVM | 0.6345 | 0.6285 | 0.6292 | 0.6288 |
| RF | 0.7139 | 0.7087 | 0.7105 | 0.7096 | TASG-RF | 0.7441 | 0.7416 | 0.7410 | 0.7413 |
| KNN | 0.5321 | 0.5352 | 0.5270 | 0.5311 | TASG-KNN | 0.5823 | 0.5680 | 0.5748 | 0.5714 |
| RNN | 0.6561 | 0.6511 | 0.6535 | 0.6523 | TASG-RNN | 0.6816 | 0.6760 | 0.6786 | 0.6773 |
|
|