Table 15 Performance of LDA-GARB with \(k=128\) and different N on two datasets under \(CV_3\).

From: Predicting lncRNA and disease associations with graph autoencoder and noise robust gradient boosting

 

Dataset

\(k=128,\,N=1\)

\(k=128,\,N=2\)

\(k=128,\,N=3\)

\(k=128,\,N=4\)

\(k=128,\,N=5\)

Precision

Dataset 1

0.8631 ± 0.0282

0.8631 ± 0.0234

0.8626 ± 0.0267

0.8584 ± 0.0291

0.8733 ± 0.0285

Dataset 2

0.9329 ± 0.0117

0.9367 ± 0.0127

0.9293 ± 0.0129

0.9310 ± 0.0130

0.9375 ± 0.0140

Recall

Dataset 1

0.8694 ± 0.0326

0.8653 ± 0.0309

0.8583 ± 0.0336

0.8620 ± 0.0309

0.8750 ± 0.0328

Dataset 2

0.9329 ± 0.0169

0.9378 ± 0.0131

0.9283 ± 0.0135

0.9303 ± 0.0154

0.9357 ± 0.0147

Accuracy

Dataset 1

0.8653 ± 0.0228

0.8636 ± 0.0190

0.8603 ± 0.0212

0.8593 ± 0.0213

0.8735 ± 0.0213

Dataset 2

0.9328 ± 0.0101

0.9371 ± 0.0091

0.9287 ± 0.0091

0.9306 ± 0.0093

0.9366 ± 0.0089

F1-score

Dataset 1

0.8658 ± 0.0229

0.8637 ± 0.0196

0.8599 ± 0.0218

0.8596 ± 0.0211

0.8736 ± 0.0216

Dataset 2

0.9328 ± 0.0103

0.9372 ± 0.0091

0.9287 ± 0.0091

0.9306 ± 0.0094

0.9365 ± 0.0089

AUC

Dataset 1

0.9356 ± 0.0159

0.9370 ± 0.0134

0.9345 ± 0.0140

0.9381 ± 0.0146

0.9317 ± 0.0152

Dataset 2

0.9812 ± 0.0049

0.9817 ± 0.0038

0.9789 ± 0.0042

0.9765 ± 0.0048

0.9790 ± 0.0047

AUPR

Dataset 1

0.9258 ± 0.0222

0.9355 ± 0.0146

0.9300 ± 0.0168

0.9359 ± 0.0184

0.9333 ± 0.0172

Dataset 2

0.9812 ± 0.0056

0.9816 ± 0.0041

0.9796 ± 0.0045

0.9749 ± 0.0065

0.9799 ± 0.0050

  1. The best performance is denoted as bold.
  2. The second-best performance is denoted as underline.
Back to article page