Table 4 Result of argument annotation using deep learning model with attention mechanism (FastText Word Embedding)
From: Argument annotation and analysis using deep learning with attention mechanism in Bahasa Indonesia
No | Model name | Accuracy (%) | Recall (%) | Precision (%) | F1 macro (%) |
|---|---|---|---|---|---|
1 | CNN | 76.50 ± 1.21 | 61.45 ± 1.92 | 65.68 ± 1.71 | 62.10 ± 1.44 |
2 | CNN + Att | 74.44 ± 2.11 | 52.31 ± 1.56 | 57.50 ± 12.65 | 49.37 ± 3.30 |
3 | LSTM | 75.76 ± 0.82 | 54.91 ± 3.07 | 65.19 ± 4.07 | 53.56 ± 5.03 |
4 | LSTM + Att | 75.87 ± 0.51 | 52.02 ± 1.75 | 62.03 ± 11.35 | 48.78 ± 3.62 |
5 | GRU | 75.74 ± 0.85 | 56.95 ± 4.17 | 64.00 ± 8.51 | 56.21 ± 5.60 |
6 | GRU + Att | 76.17 ± 0.56 | 56.32 ± 2.14 | 66.28 ± 1.42 | 56.18 ± 3.10 |
7 | BiLSTM | 75.94 ± 0.24 | 52.44 ± 2.85 | 50.35 ± 8.25 | 48.25 ± 4.25 |
8 | BiLSTM + Att | 75.88 ± 0.31 | 52.79 ± 2.98 | 63.88 ± 13.77 | 49.42 ± 5.27 |
9 | BiGRU | 76.12 ± 0.30 | 51.74 ± 1.16 | 64.30 ± 8.52 | 48.14 ± 2.27 |
10 | BiGRU + Att | 76.41 ± 0.63 | 55.49 ± 4.28 | 58.67 ± 8.98 | 53.38 ± 6.40 |