Methods and Practices of Dynamic Programming Algorithms to Promote the Fluency of Japanese Literary Metaphor Translation

Abstract

This paper uses a dynamic programming algorithm to achieve automatic sentence alignment for Japanese-Chinese metaphor translation, based on a bilingual dictionary. The dynamic programming algorithm is utilized to find out the sentence pair path from the beginning to the end of the text paragraph with the largest total evaluation value, and the sentence pairs on the path are output in reverse, which are the aligned sentence pairs of the paragraph. Using the translation dictionary and dynamic programming algorithm to align the algorithm for obtaining Japanese-Chinese translations, the correctness rate of the final extracted translation units is 85.19%, and the recall rate is about 62.2% by sampling test, and compared with the traditional alignment algorithms, the fluency of the dynamic programming alignment algorithm is about 15%–23% higher. The algorithm not only inherits the advantages of the dictionary word-based alignment algorithm but also enhances its shortcomings, thereby improving both the correctness and fluency of the alignment. According to the statistics on sentence length relationships between Chinese and Japanese metaphorical translation, the more stable the relationship is, the longer the text length. A significant difference can be seen in the Sig values of fluency between manual translation and the dynamic planning alignment algorithm, which are less than 0.01.

Keywords

• dynamic programming algorithm
• metaphorical translation
• aligned sentence pairs
• fluency
• 68m11