Hybrid Algorithm for Word-Level Alignment of Parallel Texts
Given a text in two languages, word alignment task consists of identifying in the two variants of the text specific word occurrences that are mutual translations. The majority of existing text alignment systems follow either a linguistic or a statistical approach. We argue for that both approaches are insufficient when used separately, and suggest a flexible algorithm that combines statistical and linguistic techniques.
KeywordsLinguistic Processing Linguistic Resource Word Alignment Linguistic Approach Pointwise Mutual Information
- 1.Borin, L.: You’ll take the high road and i’ll take the low road: Using a third language to improve bilingual word alignment. In: ACL 2000, vol. 1, pp. 97–103 (2000)Google Scholar
- 2.Mihalca, R., Pedersen, T.: An evaluation exercise for word alignment. In: HLT-NAACL 2003 Workshop on Building and using parallel texts, vol. 3, pp. 1–10 (2003)Google Scholar