Skip to main content

Table 3 F dep score, all sentences, loose matching criterion

From: Benchmarking natural-language parsers for biological applications using dependency graphs

Parser

F dep

Charniak-Lease

81.0

Bikel (0.9.8)

81.0

Bikel (0.9.9c)

81.0

Charniak (Aug 05)

78.0

Stanford (unlexicalised)

74.5

Collins (model 2)

72.5

Collins (model 1)

72.5

Stanford (lexicalised)

72.5

Collins (model 3)

71.5

  1. Parser effectiveness based on comparison of dependency graphs to the graphs generated from the GENIA treebank, summed over the entire corpus, disregarding dependency types.