Table 7. Experiment results of Word2Vec embedding model with deep learning algorithms for independent dataset.
The best accuracy percentages are represented in bold letters.
| MODELS | ACC | SEN | SPE | MCC | AUC |
|---|---|---|---|---|---|
| WS+CNN | 57.81 | 55.29 | 60.33 | 14.17 | 57.98 |
| WC+CNN | 84.25 | 80.10 | 68.36 | 52.47 | 83.05 |
| WS+LSTM | 88.75 | 89.25 | 88.31 | 77.48 | 93.65 |
| WC+LSTM | 86.51 | 82.62 | 90.74 | 72.83 | 91.25 |
| WS+BiLSTM | 89.06 | 88.88 | 89.37 | 77.65 | 94.05 |
| WC+BiLSTM | 92.31 | 93.79 | 91.52 | 84.62 | 96.55 |