Table 2.

Diversity and quality of multiple clusterings generated by scMCs on benchmark datasets.^a

		CellMix	PBMC_3K	AdBrain
		${\mathcal{C}}_t$	${\mathcal{C}}_t$	${\mathcal{C}}_t$
NMI$\uparrow$	${\mathcal{C}}_1$	0.845	0.695	0.513
	${\mathcal{C}}_2$	0.365	0.204	0.289
JI$\uparrow$	${\mathcal{C}}_1$	0.860	0.378	0.364
	${\mathcal{C}}_2$	0.355	0.197	0.291
SC$\uparrow$	${\mathcal{C}}_1$	0.666	0.644	0.268
	${\mathcal{C}}_2$	0.599	0.826	0.579
DI$\uparrow$	${\mathcal{C}}_1$	0.076	0.071	0.048
	${\mathcal{C}}_2$	0.054	0.040	0.053

^a ${\mathcal{C}}_t$ is the ground truth, while ${\mathcal{C}}_1$ is the clustering similar to ground truth, ${\mathcal{C}}_2$ is the other alternative clustering.