Table 10.
McNemar's tests on labeling disagreements
Null hypothesis | P-value |
---|---|
Exp1 vs. AIMed | 2.04e-09 |
Exp2 vs. AIMed | 3.57e-11 |
Exp3 vs. GENETAG | 0.0254 |
Exp4 vs. GENETAG | 0.00013 |
The experiments are as described in Table 6. AIMed and GENETAG represent the experiments with the pure AIMed and GENETAG corpora, respectively.