Table 3 Experimental results of the tenfold cross-validation yielded by SKCNN on Cdataset
From: Predicting drug−disease associations via sigmoid kernel-based convolutional neural networks
Test set | Acc. (%) | Pre. (%) | Recall (%) | F1-score (%) |
|---|---|---|---|---|
1 | 90.35 | 92.18 | 88.19 | 90.14 |
2 | 93.11 | 95.82 | 90.16 | 92.90 |
3 | 89.13 | 90.91 | 86.96 | 88.89 |
4 | 92.09 | 95.32 | 88.54 | 91.80 |
5 | 89.53 | 91.32 | 87.35 | 89.29 |
6 | 91.50 | 91.67 | 91.30 | 91.49 |
7 | 91.30 | 91.97 | 90.51 | 91.24 |
8 | 91.50 | 93.03 | 89.72 | 91.35 |
9 | 93.87 | 93.02 | 94.86 | 93.93 |
10 | 91.50 | 91.67 | 91.30 | 91.49 |
Average | 91.38 ± 1.39 | 92.69 ± 1.58 | 89.89 ± 2.21 | 91.25 ± 1.45 |