Table 2.
Feature | Model | Dim | 10-Fold Cross-Validation | Independent Test | ||||||||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
ACC | MCC | Sn | Sp | F1 | auPRC | auROC | ACC | MCC | Sn | Sp | F1 | auPRC | auROC | |||
SSA + UniRep b |
SVM c | 2021 | 0.861 | 0.723 | 0.875 a | 0.848 | 0.863 | 0.929 | 0.927 | 0.867 | 0.734 | 0.859 | 0.875 | 0.866 | 0.954 | 0.952 |
LGBM c | 0.840 | 0.680 | 0.848 | 0.832 | 0.841 | 0.933 | 0.924 | 0.859 | 0.719 | 0.859 | 0.859 | 0.859 | 0.960 | 0.958 | ||
RF c | 0.838 | 0.676 | 0.840 | 0.836 | 0.838 | 0.923 | 0.917 | 0.867 | 0.735 | 0.844 | 0.891 | 0.864 | 0.955 | 0.954 | ||
SSA + BiLSTM b |
SVM c | 3726 | 0.836 | 0.672 | 0.848 | 0.824 | 0.838 | 0.915 | 0.917 | 0.883 | 0.766 | 0.859 | 0.906 | 0.880 | 0.943 | 0.947 |
LGBM c | 0.848 | 0.696 | 0.859 | 0.836 | 0.849 | 0.927 | 0.927 | 0.875 | 0.751 | 0.906 | 0.844 | 0.879 | 0.961 | 0.957 | ||
RF c | 0.824 | 0.649 | 0.832 | 0.816 | 0.826 | 0.906 | 0.911 | 0.898 | 0.797 | 0.891 | 0.906 | 0.898 | 0.959 | 0.951 | ||
UniRep + BiLSTM b |
SVM c | 5505 | 0.844 | 0.688 | 0.859 | 0.828 | 0.846 | 0.921 | 0.926 | 0.891 | 0.783 | 0.922 | 0.859 | 0.894 | 0.966 | 0.962 |
LGBM c | 0.863 | 0.727 | 0.871 | 0.855 | 0.864 | 0.932 | 0.935 | 0.870 | 0.737 | 0.859 | 0.886 | 0.887 | 0.972 | 0.958 | ||
RF c | 0.832 | 0.664 | 0.844 | 0.820 | 0.834 | 0.932 | 0.930 | 0.875 | 0.750 | 0.859 | 0.891 | 0.873 | 0.963 | 0.960 | ||
SSA + UniRep + BiLSTM b |
SVM c | 5626 | 0.871 | 0.742 | 0.863 | 0.879 | 0.870 | 0.943 | 0.941 | 0.891 | 0.783 | 0.922 | 0.859 | 0.894 | 0.940 | 0.943 |
LGBM c | 0.855 | 0.711 | 0.844 | 0.867 | 0.854 | 0.945 | 0.942 | 0.898 | 0.797 | 0.891 | 0.906 | 0.898 | 0.971 | 0.971 | ||
RF c | 0.840 | 0.680 | 0.848 | 0.832 | 0.841 | 0.926 | 0.925 | 0.898 | 0.799 | 0.859 | 0.937 | 0.894 | 0.963 | 0.957 |
a Values representing the best performance values are in bold and are underlined. b SSA + UniRep: SSA features combined with UniRep features; SSA + BiLSTM: SSA features combined with BiLSTM features; UniRep + BiLSTM: UniRep features combined with BiLSTM features; SSA + UniRep + BiLSTM: all the above features are combined. c SVM: support vector machine; LGBM: light gradient boosting machine; RF: random forest.