Communications Chemistry

Table 1 Performance evaluation on classification datasets from MoleculeNet

From: MolGraph-xLSTM as a graph-based dual-level xLSTM framework for enhanced molecular representation and interpretability

	Sider		Tox21		Clintox		BBBP		BACE		HIV
	AUROC	AUPRC	AUROC	AUPRC	AUROC	AUPRC	AUROC	AUPRC	AUROC	AUPRC	AUROC	AUPRC
FP-GNN	0.661 ± 0.014	0.679 ± 0.026	0.833 ± 0.004	0.459 ± 0.018	0.732 ± 0.068	0.622 ± 0.028	0.892 ± 0.019	0.953 ± 0.007	0.852 ± 0.035	0.740 ± 0.042	0.767 ± 0.039	0.328 ± 0.078
DeeperGCN	0.622 ± 0.031	0.660 ± 0.025	0.840 ± 0.010	0.434 ± 0.021	0.892 ± 0.048	0.741 ± 0.048	0.860 ± 0.014	0.937 ± 0.008	0.830 ± 0.033	0.719 ± 0.039	0.769 ± 0.041	0.300 ± 0.064
DMPNN	0.658 ± 0.032	0.680 ± 0.030	0.849 ± 0.006	0.481 ± 0.026	0.895 ± 0.010	0.727 ± 0.062	0.896 ± 0.014	0.956 ± 0.016	0.851 ± 0.028	0.742 ± 0.043	0.758 ± 0.029	0.278 ± 0.043
HiGNN	0.656 ± 0.024	0.669 ± 0.027	0.844 ± 0.006	0.462 ± 0.018	0.889 ± 0.026	0.735 ± 0.070	0.892 ± 0.014	0.943 ± 0.017	0.836 ± 0.029	0.740 ± 0.048	0.768 ± 0.038	0.310 ± 0.066
TransFoxMol	0.636 ± 0.022	0.686 ± 0.040	0.816 ± 0.011	0.367 ± 0.011	0.830 ± 0.047	0.624 ± 0.036	0.881 ± 0.015	0.947 ± 0.005	0.801 ± 0.054	0.693 ± 0.079	0.727 ± 0.037	0.232 ± 0.063
BiLSTM	0.590 ± 0.008	0.631 ± 0.023	0.807 ± 0.011	0.382 ± 0.008	0.987 ± 0.009	0.947 ± 0.023	0.943 ± 0.020	0.985 ± 0.003	0.826 ± 0.052	0.807 ± 0.076	0.769 ± 0.040	0.305 ± 0.040
AutoML	0.682 ± 0.017	0.699 ± 0.029	0.828 ± 0.004	0.468 ± 0.022	0.875 ± 0.058	0.645 ± 0.018	0.928 ± 0.013	0.969 ± 0.019	0.840 ± 0.013	0.753 ± 0.018	0.772 ± 0.033	0.351 ± 0.078
MolGraph-xLSTM (Ours)	0.697 ± 0.022	0.713 ± 0.032	0.854 ± 0.003	0.487 ± 0.045	0.904 ± 0.032	0.714 ± 0.026	0.959 ± 0.006	0.987 ± 0.002	0.869 ± 0.016	0.784 ± 0.029	0.775 ± 0.027	0.355 ± 0.050

Best results are shown in bold.

Back to article page

Search

Advanced search

Quick links