An optimized graph-based structure for single-cell RNA-seq cell-type classification based on non-linear dimension reduction

Table 3 Simulation result for all compared methods

	SCEA		SC3		scGAC		scGNN		Seurat		Desc
Dataset	ARI	NMI	ARI	NMI	ARI	NMI	ARI	NMI	ARI	NMI	ARI	NMI
Klein	0.843	0.869	0.831	0.859	0.835	0.856	0.601	0.680	0.836	0.856	0.808	0.784
Zeisel	0.562	0.616	0.822	0.076	0.520	0.624	0.495	0.635	0.494	0.661	0.473	0.593
Romanov	0.600	0.535	0.551	0.530	0.569	0.553	0.283	0.330	0.426	0.641	0.403	0.589
Chung	0.317	0.490	0.025	0.310	0.303	0.446	0.298	0.469	0.209	0.470	0.215	0.493
Pbmc	0.661	0.781	0.643	0.737	0.613	0.760	0.562	0.683	0.623	0.746	0.687	0.757
Mouse	0.446	0.672	0.616	0.757	0.393	0.664	0.362	0.635	0.571	0.762	0.593	0.752
Biase	1.00	1.00	0.948	0.929	1.00	1.00	0.330	0.443	1.00	1.00	0.594	0.641
Petropoulos	0.594	0.654	0.538	0.627	0.439	0.583	0.282	0.384	0.322	0.523	0.318	0.555
Neuron	0.408	0.461	0.327	0.315	0.385	0.427	0.236	0.425	0.249	0.324	0.239	0.364
Mouse Brain	0.786	0.746	0.754	0.696	0.744	0.708	0.306	0.486	0.410	0.579	0.535	0.588

ISSN: 1471-2164