Korean-Chinese Machine Translation Using Three-Stage Verb Pattern Matching
In this paper, we describe three-stage pattern matching approach and an effective pattern construction process in the Korean-Chinese Machine Translation System for technical documents. We automatically extracted about 100,000 default verb patterns and about 10,000 default ordering patterns from the existing patterns. With the new three-stage approach, additionally using default verb and ordering patterns, the matching hit rate increases 3.8% , comparing with one-stage approach.
KeywordsKorean-Chinese machine translation verb pattern pattern construction
Unable to display preview. Download preview PDF.
- 1.Takeda, K.: Pattern-Based Context-Free Grammars for Machine TranslationGoogle Scholar
- 2.Watanabe, H., Takeda, K.: A Pattern-based Machine Translation System Extended by Example-based Processing. In: Proceedings of the 36th annual meeting on Association for Computational Linguistics, vol. 2, pp. 1369–1373Google Scholar