Skip to main content
. 2022 Mar 29;23(3):bbac106. doi: 10.1093/bib/bbac106

Table 7.

Performance scores and validation scheme of the methods involved in this review

Study Algorithms Validation scheme AUROC AUPRC ACC F1 MCC Precision Sensitivity Specificity
Pratapa et al. [22] SVM 10-fold cross-validation 0.796
Wu et al. [71] k-NN 10-fold cross-validation 0.848 0.861 0.764 0.739 0.825 0.670
Pandey et al. [72] MNMC 10-fold cross-validation 0.897
Wu et al. [73] Ensemble learning 5-fold cross-validation 0.871
Li et al. [24] RF 10-fold cross-validation 0.532
Benstead-Hume et al. [25] RF 5-fold cross-validation 0.889
Liu et al. [29] Logistic matrix factorization 5-fold cross-validation 0.848 0.239
Huang et al. [28] Matrix factorization 5-fold cross-validation 0.923
Liany et al. [30] CMF 3-fold cross-validation 0.980 0.980
Wan et al. [41] Neural network 5-fold cross-validation 0.969 0.880 0.959 0.866 0.872 0.903 0.968
Cai et al. [31] GCN 5-fold cross-validation 0.878 0.344 0.552
Long et al. [32] GAT 5-fold cross-validation 0.937 0.948
Hao et al. [33] GAE 5-fold cross-validation 0.917 0.942 0.871
Wang et al. [80] KG 5-fold cross-validation 0.947 0.956 0.887

Notes: AUROC, area under receiver optimizer characteristics curve; AUPRC, area under precision-recall curve; ACC, accuracy; MCC, Matthews correlation coefficient.