Table 11.

Two-sided Wilcoxon paired-samples signed ranks test comparing the performance of ISMB with other algorithms. For each performance measure the number on top is the Z statistic and the number at the bottom is the corresponding p-value. The Z statistic is negative when ISMB has a lower score on a performance measure than the competing algorithm. On all measures, a negative Z statistic indicates better performance by ISMB. Underlined results indicate p-values of 0.05 or smaller.

Performance measure	ISMB-MS	NISMB	NB	DT	LR	NN	kNN	LBR	AB
Misclassification error	−2.338	−0.776	−2.121	−2.070	−1.181	−3.861	−0.825	−0.368	−1.720
	0.019	0.438	0.034	0.038	0.238	<0.001	0.409	0.713	0.085
AUC	−2.085	−1.257	−1.511	−4.457	−2.197	−4.029	−4.203	−0.927	−3.198
	0.037	0.209	0.131	0.001	0.028	0.001	0.001	0.354	0.001
Logarithmic loss	−3.595	−2.426	−4.280	−4.457	−3.340	−4.254	−4.026	−4.051	−3.215
	0.001	0.015	0.001	0.001	0.001	0.001	0.001	0.001	0.001
Squared error	−3.608	−2.313	−3.975	−3.24	−2.121	−4.127	−3.518	−3.213	−2.839
	0.001	0.021	0.001	0.001	0.034	0.001	0.001	0.001	0.005
CAL score	−2.032	−1.867	−4.026	−2.806	−0.063	−4.076	−1.892	−3.543	−1.443
	0.042	0.062	0.001	0.005	0.949	0.001	0.058	0.001	0.149