cbrew / chartparse Goto Github PK

The parser should accommodate probabilities, alllowing computation of both inside probability and best-parse probabilities. This is partially done, seems to merit its own issue and branch.

Adjusted documentation to use sphinx instead of epydoc.

The parser could be generalized to cover MCFGs, as defined by Seki. These generalize CFGs (for which the yield of a rule is a string of terminal symbols) to the case where the yield of a rule is a tuple of strings of terminal symbols. Since the grammar describes sets of strings, S (the distinguished sentence symbol) is constrained to produce a 1-tuple of strings.

The most recent work on parsing these formalisms focuses attention on Linear Context-Free Rewriting Systems. There is an IWPT 2013 paper:

Andreas van Cranenburgh, Rens Bod (2013).
Discontinuous Parsing with an Efficient and Accurate DOP Model
Proceedings of the International Conference on Parsing Technologies, Nara, Japan, 27-29 November.

and a CL journal article:

Laura Kallmeyer and Wolfgang Maier (2013): *Data-driven Parsing using Probabilistic Linear Context-Free Rewriting Systems." Computational Linguistics, 39(1).

Both of these have associated software.