HoxPred: automated classification of Hox proteins using combinations of generalised profiles

BMC Bioinformatics

Table 4 Comparison of the 4 classification methods performance for OG predictions among each PG.

PG	2-groups	PG-groups	anterior/posterior	all-groups
PG1	2.00	2.00	2.00	2.00
PG2	0.73	0.60	0.62	0.62
PG3	1.70	1.78	1.78	1.40
PG4	1.70	1.65	1.60	1.82
PG5	0.43	0.43	0.43	0.38
PG6	0.43	0.43	0.43	0.43
PG7	1.00	0.85	0.85	0.85
PG8	0.40	0.43	0.43	0.36
PG9	1.89	1.85	1.63	1.63
PG10	0.40	0.43	0.43	0.39
PG11	1.00	1.00	1.00	1.00
PG12	1.00	1.00	1.00	1.00
PG13	1.89	1.85	1.85	1.85

Each value is the surface of the polygon that represents one classification method on a radar plot. The polygon is quadrilateral for PG containing 4 OG and triangular for PG containing 3 OG. For PG containing 2 OG, the value is the product of the 2 OG accuracies, for a given classification method.

ISSN: 1471-2105