Abstract
In this paper, we propose a semi-automatic tree annotating workbench for building a Korean treebank. Generally, building a treebank requires an enormous effort by the annotator. In order to improve annotating efficiency, decrease the number of intervention required by the annotator, and help maintain consistent annotation in building a treebank, we have developed a semi-automatic tree annotating workbench consisting of following three stages: syntactic pattern extraction, syntactic pattern selection, and syntactic pattern application. The experiment was carried out with 27,966 tree tagged sentences as a training set and 3,108 sentences as a test set. As a result, the burden of manual annotation can be reduced by about 47% with the best selection of the feature set by using the proposed tree annotating workbench.
| Original language | English |
|---|---|
| Pages (from-to) | 253-257 |
| Number of pages | 5 |
| Journal | Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) |
| Volume | 2945 |
| Publication status | Published - 2004 Dec 1 |
Fingerprint
ASJC Scopus subject areas
- Computer Science(all)
- Biochemistry, Genetics and Molecular Biology(all)
- Theoretical Computer Science