Super Tagger

The CCG supertagger uses a set of 425 lexical categories from CCGbank. This set contains all those categories which occur at least 10 times in Sections 2-21, and has good coverage on unseen CCGbank data. The per-word accuracy of the supertagger is around 92% on unseen WSJ text. Using the multi-supertagger increases the accuracy significantly -- to over 98% -- with only a small cost in increased ambiguity.