Ensemble Methods with Voting Protocols Exhibit Superior Performance for Predicting Cancer Clinical Endpoints and Providing More Complete Coverage of Disease-Related Genes

Table 8

Coverage of the datasets from miRNA.

Disease	ShareNum	Subdata scale
		10	30	50	100
BRCA	5	1	2	2	8
	4	3	8	17	47
	3	3	11	23	31
	2	3	9	12	26
	1	18	57	79	127
	Total	28	87	133	239
	Coverage fraction	35.7%	34.5%	40.6%	46.9%
KIRC	5	1	14	16	35
	4	3	6	13	53
	3	5	7	19	12
	2	2	7	11	9
	1	14	21	39	59
	Total	25	55	98	168
	Coverage fraction	44%	61.8%	60.2%	64.9%
OV	5	0	5	17	83
	4	3	15	27	0
	3	7	10	6	0
	2	3	5	12	0
	1	11	25	15	0
	Total	24	60	77	83
	Coverage fraction	54.2%	58.3%	80.5%	100%

The coverage fraction means the number of the genes which are used for more than 1 times in all of the genes, and accordingly, the formula is . The ShareNum is the number of a gene that is used for the subdatasets. For example, in Table 8, the value in the OV-miRNA group with ShareNum 3 and data scale 10 is 7; it means that there are 7 genes which are used by 3 subdatasets and each subdataset has 10 microRNAs as the attributes. The subdatasets from the OV-miRNA group have at most 83 microRNAs, and thus, the scale “100” of OV-miRNA means 83.