Table 1.
ROC analysis of the classification accuracy achieved with different kernel functions
Organism | Accession no. | Linear | Polynomial | Gaussian | Δbest-linear |
---|---|---|---|---|---|
E.coli O157:H7 | BA000007 | 0.960 | 0.960 | 0.968 | 0.008 |
T.pallidum subsp. Pallidum str. Nichols | AE000520 | 0.920 | 0.930 | 0.929 | 0.010 |
C.trachomatis D/UW-3/CX | AE001273 | 0.976 | 0.982 | 0.987 | 0.009 |
B.aphidicola str. APS | BA000003 | 0.986 | 0.991 | 0.989 | 0.005 |
The ROC0.1 measures the discriminatory power of the SVM in gene identification based on sequence composition with a linear, polynomial or Gaussian kernel function.