Table 4.
Misclassifications that are complemented to be correct by different approaches
Features used + Similarity measure |
Distributional dependencies + α-skew divergence |
Strings without annotations + Naïve Bayesian |
Strings with annotations + Naïve Bayesian |
Distributional dependencies + α-skew divergence |
-- | 10/17 | 12/17 |
Strings without annotations + Naïve Bayesian |
15.5/22.5 | -- | 6/22.5 |
Strings with annotations + Naïve Bayesian |
12.5/17.5 | 1/17.5 | -- |