MISSEL: a method to identify a large number of small species-specific genomic subsequences and its application to viruses classification

Table 7 Classification accuracy on training and test set for the 3 genomic regions of Influenza viruses (mean ± standard deviation computed on all solutions)

Genomic region	β	Train [%]	Test [%]
HA	2	95.44 ± 17.14	95.46 ± 17.05
	3	99.99 ± 0.06	99.98 ± 0.08
	4	99.99 ± 0.11	99.96 ± 0.15
	5	100 ± 0.06	99.98 ± 0.09
	6	100	99.99 ± 0.04
	7	100	99.99 ± 0.04
	8	100	99.98 ± 0.05
	9	100	99.98 ± 0.06
	10	100	99.98 ± 0.07
NA	2	100 ± 0.01	99.98 ± 0.02
	3	99.98 ± 0.11	99.96 ± 0.13
	4	100	99.99 ± 0.02
	5	100	99.99 ± 0.02
	6	100	99.99 ± 0.02
	7	100	99.99 ± 0.03
	8	100	99.99 ± 0.03
	9	100	99.99 ± 0.02
	10	100	100 ± 0.01
MP	2	56.72 ± 29.09	56.41 ± 28.95
	3	84.59 ± 5.75	83.94 ± 5.48
	5	38.46 ± 2.44	38.39 ± 2.54
	8	71.15 ± 40.04	71.24 ± 40.10
	9	81.01	80.96
	12	81.51	80.72
	19	91.66	91.14

ISSN: 1756-0381