Device and method for aligning word and example sentence paginal translation dictionary

Abstract

The invention provides a device and a method for aligning words and an example sentence paginal translation dictionary, which improve the invoking rate and the precision. A device (200) for aligning the words comprises a word paginal translation dictionary (208) for storing the words and paginal translation thereof and the words and explaining translating words thereof, an input part (202) for inputting example sentences and example sentence paginal translation pairs serving as the paginal translation of the example sentences, a morpheme analyzing part (204) for carrying out morpheme analysis on the inputted example sentence paginal translation pairs, a dictionary-based aligning part (210) for aligning the words of the example sentence paginal translation pairs extracted by the morpheme analysis according to the word paginal translation dictionary, a word list token-based aligning part (214) for aligning words to be aligned by the dictionary-based aligning part (210), a more-to-more aligning part (218), a bipartite graph-based aligning part (220) and an output part (222) for outputting aligning results among the words.

Description

technical field [0001] The invention relates to an automatic word alignment (Alignment) device for translated example sentences, in particular to a learning device for word alignment example sentence translation dictionary and translation word extraction in a machine translation device based on example sentences. Background technique [0002] Machine translation is the use of computers to convert from a certain language to other languages. Such research and development have been carried out for half a century in the world. Machine translation methods can be broadly classified into 1) analysis-based machine translation methods, 2) statistics-based machine translation methods, and 3) example sentence-based machine translation methods. [0003] Analysis-based machine translation is a technology that performs analysis (morphological analysis, grammar / meaning analysis, etc.) in a first language, converts the analysis results into a second language, and generates a translation in ...

AI Technical Summary

Problems solved by technology

[0026] 3) The matching alignment method based on the maximum and minimum weighting of the bipartite graph is very effective in 1-to-1 word alignment, but it may not be an effective method in many-to-many word alignment

Images