Table 2.

Performance comparison of extremely randomized trees (ERT)-based ensemble method with ERT-based other classifiers on benchmarking dataset.

Features	Matthews’ correlation coefficient (MCC)	Accuracy	Sensitivity	Specificity	AUC	P-value
Amino acid composition (AAC)	0.367 ± 0.025	0.694 ± 0.034	0.612 ± 0.108	0.743 ± 0.074	0.752 ± 0.015	0.09
Dipeptide composition (DPC)	0.375 ± 0.022	0.686 ± 0.011	0.636 ± 0.022	0.737 ± 0.017	0.757 ± 0.013	0.325
Composition–transition–distribution (CTD)	0.295 ± 0.030	0.647 ± 0.015	0.607 ± 0.019	0.687 ± 0.018	0.694 ± 0.017	0.00002
Physicochemical properties (PCP)	0.313 ± 0.030	0.656 ± 0.015	0.632 ± 0.018	0.680 ± 0.017	0.705 ± 0.013	0.0002
Amino acid index (AAI)	0.371 ± 0.028	0.685 ± 0.014	0.648 ± 0.015	0.722 ± 0.018	0.748 ± 0.013	0.143
Hybrid	0.348 ± 0.022	0.674 ± 0.007	0.645 ± 0.016	0.703 ± 0.006	0.733 ± 0.012	0.02
Ensemble learning (EL)	0.423 ± 0.024	0.712 ± 0.012	0.714 ± 0.014	0.709 ± 0.015	0.775 ± 0.011	–

The first column corresponds to the performance of individual feature group, hybrid feature, and ensemble learning. The column 2–6 respectively represent the MCC, accuracy, sensitivity, specificity, and AUC value, where each value shown as the average ± SD of 10 alternative balanced datasets. The last column represents a pairwise comparison of AUC between EL and the other methods using a two-tailed t-test. P ≤ 0.05 indicates a statistically meaningful difference between EL and the selected composition (shown in bold).