NDT 2.0 with Constituent Structure

In this version of the Norwegian Dependency Treebank 2.0 constituent structure (c-structure) similar to the one found in NorGramBank has been added.

This can be used to train one syntactic parser for both grammatical frameworks (dependency- and constituentanalyses). This has been shown to be beneficial to the performance of parsers.

The c-structure has been assigned automatically. Hence, caution should be exercised with respect to the results, since the analyses have not been checked manually.

Feel free to contact us at sprakbanken@nb.no if you have questions or comments about this resource.