Table 3.

The performance of correctness with the best pair of (d_cn,k_var) for each threshold of the size of the model among training dataset. The best pairs are chosen from the result in Fig. 2 that shows best microAUC score in the same thresholds. We note that 10-fold cross validation is used on the training set

Threshold of the model size	dcn	kvar	Filtered genes		microAUC	Accuracy
			CN	Variants
512	0.17	60	243	265	0.98179	0.83321
1024	0.13	90	358	404	0.98523	0.84317
2048	0.08	270	709	1198	0.98625	0.86827
4096	0.08	550	709	2364	0.98704	0.86974