Research on Improved Dual Channel Medical Short Text Intention Recognition Algorithm
, and
Mar 16, 2024
About this article
Published Online: Mar 16, 2024
Page range: 82 - 89
DOI: https://doi.org/10.2478/ijanmc-2023-0079
Keywords
© 2023 Chao Wang et al., published by Sciendo
This work is licensed under the Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.
Figure 1.
Figure 2.
Figure 3.
Figure 4.
Figure 5.
Figure 6.
Figure 7.
comparison of experimental results
| Model | Acc% | Pre% | Recall% | F1% |
|---|---|---|---|---|
| SAttBiGRU | 96.16 | 96.20 | 96.16 | 96.17 |
| Self-Attention-CNN | 94.85 | 94.89 | 94.85 | 94.85 |
| BiGRU-MCNN | 95.43 | 95.45 | 95.43 | 95.43 |
| MC-AttCNN-AttBiGRU | 95.93 | 95.98 | 95.93 | 95.93 |
experimental dataset
| Name | Training Set | Test Set | Validation Set | Category | Total |
|---|---|---|---|---|---|
| KUAKE-QIC | 6931 | 1994 | 1955 | 11 | 10880 |
| THUCNews_Title | 180000 | 10000 | 10000 | 10 | 200000 |
results of ablation experiment
| Model | Acc% | Pre% | Recall% | F1% |
|---|---|---|---|---|
| TextCNN | 89.96 | 89.90 | 89.96 | 89.90 |
| Improved TextCNN | 94.85 | 94.89 | 94.85 | 94.85 |
| BiGRU-att | 94.00 | 94.17 | 94.00 | 94.90 |
| AB-CNN-BGRU-att | 96.68 | 96.68 | 96.67 | 96.67 |
comparison between bigru-att and bilstm-att
| Network Layer | Average Duration | Total Duration | Acc% | F1% |
|---|---|---|---|---|
| BiGRU-att | 2286.9s | 45738s | 90.83 | 90.64 |
| BiLSTM-att | 2422.55s | 48451s | 90.45 | 90.41 |