French standard parsing experiments (test set, sentences ≤ 40 words). FactLex uses basic POS tags predicted by the parser and morphological analyses from Morfette. FactLex* uses gold morphological analyses.
French . | Leaf Ancestor . | Evalb . | ||||
---|---|---|---|---|---|---|
. | Sent. . | Corpus . | LP . | LR . | F1 . | EX% . |
PAPCFG | 0.857 | 0.840 | 73.5 | 72.8 | 73.1 | 14.5 |
SplitPCFG | 0.870 | 0.853 | 77.9 | 77.1 | 77.5 | 16.0 |
Berkeley | 0.905 | 0.894 | 83.9 | 83.4 | 83.6 | 24.0 |
DP-TSG | 0.858 | 0.841 | 77.1 | 76.8 | 76.9 | 16.0 |
Stanford | 0.869 | 0.853 | 78.5 | 79.6 | 79.0 | 17.6 |
Stanford+FactLex | 0.877 | 0.860 | 79.0 | 79.6 | 79.3 | 19.6 |
Stanford+FactLex* | 0.890 | 0.874 | 82.8 | 84.0 | 83.4 | 27.4 |
French . | Leaf Ancestor . | Evalb . | ||||
---|---|---|---|---|---|---|
. | Sent. . | Corpus . | LP . | LR . | F1 . | EX% . |
PAPCFG | 0.857 | 0.840 | 73.5 | 72.8 | 73.1 | 14.5 |
SplitPCFG | 0.870 | 0.853 | 77.9 | 77.1 | 77.5 | 16.0 |
Berkeley | 0.905 | 0.894 | 83.9 | 83.4 | 83.6 | 24.0 |
DP-TSG | 0.858 | 0.841 | 77.1 | 76.8 | 76.9 | 16.0 |
Stanford | 0.869 | 0.853 | 78.5 | 79.6 | 79.0 | 17.6 |
Stanford+FactLex | 0.877 | 0.860 | 79.0 | 79.6 | 79.3 | 19.6 |
Stanford+FactLex* | 0.890 | 0.874 | 82.8 | 84.0 | 83.4 | 27.4 |