Files in this item
|(no description provided)|
|Title:||Table-driven parsing of natural language with constraint-based grammars|
|Doctoral Committee Chair(s):||Morgan, Jerry L.|
|Department / Program:||Linguistics|
|Degree Granting Institution:||University of Illinois at Urbana-Champaign|
|Abstract:||In this dissertation, it is shown how efficient natural language parsing with constraint-based grammars can be achieved. In so doing, a variant of the LR-parsing algorithm is developed that deals with grammars using feature-based categories. It is shown that constraint-based grammars bring about the problems of potential nontermination and unnecessary nondeterminism with the LR parsing method, which do not arise from grammars using monadic categories. To achieve a terminating and deterministic algorithm, an extended method for constructing parsing tables is developed. The extended LR method is unable to handle schematic rules as in HPSG and CUG, however. To remedy this problem, a rule inference algorithm is developed that instantiates underspecified rules into more specified ones containing enough information to construct sets of items and parsing tables. The implementation and evaluation of the extended LR parsing method is also discussed.
The LR-based parsing method is further explored to find out whether it can be used to model human sentence processing. Traditional approaches to local ambiguity resolution are reviewed. It is shown how the reduce-first LR parser and combinatory categorial grammar can be used with a semantics-based local ambiguity resolution technique.
Finally, prominent problems with the two-level model for morphological parsing are discussed. A two-level morphological processor augmented with feature-based CF word grammars is proposed which allows for more enriched morphosyntactic descriptions. It is shown how LR predictions manifested in the parsing table can used to optimize the dictionary lookup process.
|Rights Information:||Copyright 1993 Lee, Kang-Hyuk|
|Date Available in IDEALS:||2011-05-07|
|Identifier in Online Catalog:||AAI9314899|