Table 4. Experimental results for machine learning and deep learning algorithms for Kaggle dataset.
| Kaggle | Machine learning | Deep learning | Proposed technique | ||||||
|---|---|---|---|---|---|---|---|---|---|
| Dataset | NB | LR | SVM | DT | RF | LSTM | Bi-LSTM | GRU | Focal loss |
| Accuracy | 71.2 | 85.2 | 81.2 | 73.6 | 77.1 | 77.1 | 80.5 | 76.4 | 97.00 |
| Precision | 96.6 | 76.6 | 79.7 | 58.1 | 82.4 | 65.6 | 69.5 | 60.4 | 92.72 |
| Recall | 12.3 | 52.4 | 52.9 | 54.4 | 33.4 | 63.9 | 64.7 | 75.3 | 78.88 |
| F1 Score | 18.6 | 62.2 | 63.6 | 65.2 | 47.6 | 64.7 | 67.0 | 67.0 | 85.54 |
| Training time (s) | 0.03 | 2.07 | 12.53 | 7.8 | 23.41 | 4114.3 | 6626.75 | 4163.93 | 614.735 |
| Testing time (s) | 0.01 | 0.23 | 5.16 | 0.20 | 0.46 | 13.53 | 42.09 | 11.52 | 4.238 |
Notes.
The best performing results are shown in bold.