Table 10 Performance of LDA-GARB with \(k=64\) and different N on two datasets under \(CV_1\).
From: Predicting lncRNA and disease associations with graph autoencoder and noise robust gradient boosting
| Â | Dataset | \(k=64,\,N=1\) | \(k=64,\,N=2\) | \(k=64,\,N=3\) | \(k=64,\,N=4\) | \(k=64,\,N=5\) |
|---|---|---|---|---|---|---|
Precision | Dataset 1 | 0.8636 ± 0.0450 | 0.8502 ± 0.0510 | 0.8513 ± 0.0391 | 0.8697 ± 0.0422 | 0.8764 ± 0.0419 |
Dataset 2 | 0.9344 ± 0.0147 | 0.9342 ± 0.0155 | 0.9277 ± 0.0186 | 0.9315 ± 0.0164 | 0.9403 ± 0.0183 | |
Recall | Dataset 1 | 0.7682 ± 0.0452 | 0.7716 ± 0.0501 | 0.7690 ± 0.0570 | 0.7469 ± 0.0642 | 0.7552 ± 0.0677 |
Dataset 2 | 0.8680 ± 0.0429 | 0.8633 ± 0.0439 | 0.8698 ± 0.0374 | 0.8620 ± 0.0430 | 0.8459 ± 0.0498 | |
Accuracy | Dataset 1 | 0.8282 ± 0.0338 | 0.8241 ± 0.0298 | 0.8218 ± 0.0325 | 0.8234 ± 0.0394 | 0.8281 ± 0.0350 |
Dataset 2 | 0.9036 ± 0.0281 | 0.9010 ± 0.0264 | 0.9011 ± 0.0257 | 0.8990 ± 0.0259 | 0.8957 ± 0.0289 | |
F1-score | Dataset 1 | 0.8117 ± 0.0312 | 0.8077 ± 0.0394 | 0.8066 ± 0.0370 | 0.8017 ± 0.0424 | 0.8088 ± 0.0392 |
Dataset 2 | 0.8995 ± 0.0266 | 0.8967 ± 0.0244 | 0.8974 ± 0.0230 | 0.8948 ± 0.0251 | 0.8897 ± 0.0280 | |
AUC | Dataset 1 | 0.9180 ± 0.0219 | 0.9167 ± 0.0244 | 0.9144 ± 0.0249 | 0.8988 ± 0.0324 | 0.8998 ± 0.0269 |
Dataset 2 | 0.9716 ± 0.0134 | 0.9709 ± 0.0121 | 0.9714 ± 0.0123 | 0.9692 ± 0.0120 | 0.9669 ± 0.0110 | |
AUPR | Dataset 1 | 0.9160 ± 0.0286 | 0.9048 ± 0.0395 | 0.9119 ± 0.0303 | 0.9014 ± 0.0323 | 0.9007 ± 0.0324 |
Dataset 2 | 0.9723 ± 0.0101 | 0.9706 ± 0.0096 | 0.9730 ± 0.0090 | 0.9701 ± 0.0095 | 0.9660 ± 0.0113 |
