Incremental Parser for Czech
| Authors | |
|---|---|
| Year of publication | 2005 |
| Type | Article in Proceedings |
| Conference | Proceedings of the 4th International Symposium on Information and Communication Technologies (WISICT05) |
| MU Faculty or unit | |
| Citation | |
| Web | http://nlp.fi.muni.cz/publications/wisict2005_smrz_xkadlec/ |
| Field | Informatics |
| Keywords | parsing |
| Description | The presented paper deals with an efficient implementation of incremental bottom-up chart parser for Czech. We shortly introduce the system based on a metagrammar that generates a context-free backbone supplemented by contextual constraints. The main part of the paper focuses on the way the pruning constraints are evaluated in the incremental parsing mode. The algorithm takes advantage of a special form of feature structures employed in our grammar, namely the limited number of values that can be produced for each type of the constraints. Instead of pruning the original packed share forest, the parser builds a new forest of values resulting from the evaluation. The described method enables efficient parsing even for extremely ambiguous grammars typical for free word order languages. |
| Related projects: |