Informace o publikaci
Three Approaches to Word Sense Disambiguation for Czech
| Autoři | |
|---|---|
| Rok publikování | 2001 |
| Druh | Článek ve sborníku |
| Konference | Text, Speech and Dialogue, 4th International Conference, TSD 2001 |
| Fakulta / Pracoviště MU | |
| Citace | |
| Obor | Počítačový hardware a software |
| Popis | Before building a full wsd system it is necessary to have a balanced and representative corpus annotated with sense tags. This requirement is not certainly fulfilled for the Czech language. Thus, we decided to develop some particular methods for annotating texts and we have started with the most common nouns. In our approach, the disambiguation algorithm based on sets of words (called bags) was used. The advantage of this approach is the possibility of filling bags in various ways. Our ultimate goal is to reduce manual work as much as possible. Here we present three basic ways of filling bags. The first one is based on the machine readable version of SSJČ, the second takes the advantage of learning from manually annotated text and the strategy of pseudoclustering is the third one. |
| Související projekty: |