The Automatic Extraction of Translation Patterns and Matching Algorithm in an English-Chinese Machine Translation System
Jian Li, Bo Wang · 2006
In this paper, an automatic extraction algorithm of translation patterns between English and Chinese is presented. The patterns are extracted from a bilingual corpus aligned at the sentence level. Before the extraction of translation patterns, the sentences in the corpus will be pre-processed by the rule-based method. In the pattern-based method, a new algorithm of calculating sentence similarity is presented. Based on DP matching algorithm, we design a structure matching algorithm and the pattern matching rule. Then, a semantic similarity algorithm is presented. The experiment shows that the score of translation quality is 68.4 and the accurate rate can reach 88%.