Fig 2. The effect of training set size on classification performance.

For the sake of visual clarity, the results of OPF using the distance metrics Bray-Curtis and Chi-Square, and SVM using linear and polynomial kernels are excluded.