Table 7 Results of BiLSTM + attention, transformer and GCN.
From: Predictive model of ulcerative colitis syndrome with ensemble learning and interpretability methods
BiLSTM + attention | Transformer | ||||||
|---|---|---|---|---|---|---|---|
Precision | Recall | F1-score | Precision | Recall | F1-score | Support | |
0 | 0.76 | 0.32 | 0.45 | 0.71 | 0.54 | 0.62 | 92 |
1 | 0.22 | 0.78 | 0.35 | 0.53 | 0.27 | 0.35 | 282 |
2 | 0.80 | 0.39 | 0.52 | 0.69 | 0.90 | 0.78 | 955 |
3 | 0.00 | 0.00 | 0.00 | 0.20 | 0.24 | 0.22 | 59 |
4 | 0.03 | 0.01 | 0.02 | 0.36 | 0.05 | 0.09 | 96 |
5 | 0.25 | 0.18 | 0.21 | 0.40 | 0.20 | 0.27 | 132 |
Accuracy | 0.40 | 0.64 | 1616 | ||||
Macro avg | 0.34 | 0.28 | 0.26 | 0.48 | 0.37 | 0.39 | 1616 |
Weighted avg | 0.58 | 0.40 | 0.41 | 0.60 | 0.64 | 0.59 | |
GCN (Precision) | Recall | F1-score | Support | ||||
|---|---|---|---|---|---|---|---|
0 | 0.45 | 0.98 | 0.62 | 92 | |||
1 | 0.00 | 0.00 | 0.00 | 279 | |||
2 | 0.88 | 0.10 | 0.18 | 956 | |||
3 | 0.00 | 0.00 | 0.00 | 58 | |||
4 | 0.06 | 0.71 | 0.12 | 98 | |||
5 | 0.08 | 0.08 | 0.08 | 133 | |||
Macro avg | 0.25 | 0.31 | 0.17 | 1616 | |||
Weighted avg | 0.56 | 0.17 | 0.16 | 1616 |
