CONVERSATIONAL ENGLISH-JAPANESE MACHINE TRANSLATION SYSTEM

摘要

PURPOSE:To solve such problem that multiple part of speech property exists in a word in an conversational machine translation system and to reduce burden on analysis. CONSTITUTION:This system is provided with an input means 1 which inputs a translation target English text as a character string, a word segmenting means 2 which segments the word in an inputted English text, and a storage means 3 which stores a be-verb, an auxiliary verb, a preposition, and an article as headers. Also, a system side can estimate a clause by providing rule groups 3, 4 which estimate the range of clause from the position of the be-verb, the auxiliary verb, the preposition, or the article in the text, a clause estimation means 9, 10 which estimate the clause by using the rule groups 3, 4, a display means 11 which displays an estimation result, and an estimation result change means 12 which changes the estimation result based on estimation information. The estimation result is represented to a user, and the user can correct the estimation result, and the analysis of the English text can be performed by using clause information obtained as a result.