作者: Yating Yang , Chenggang Mi , Lei Wang , Xiao Li
DOI: 10.2174/1874444301406010561
关键词:
摘要: To alleviate the data sparseness problem during word alignment, we propose a alignment method based on co-occurrence degree. In this paper, new to get statistical information from co- occurrence. We combine counts and fuzzy weights as Fuzzy can be obtained by searching for pairs computing differences of length between current other words in pairs. Experiments show that quality translation performance both improved.