KR970066941A - Multilingual translation system using token separator - Google Patents

Abstract

The present invention relates to a multilingual translation system using a token separator. In order to make one concept correspond to one word, words having a significant influence on the definition of a concept are defined as a reference word, and a token separator In order to provide a multilingual translation system installed at the front end of an interpreter, a first step (31,32) of separating input sentences using a token separator by a token separator when a source language sentence is inputted; A second step 33 of interpreting the primitive language using the conceptual structure; And a third step (34,35) of translating the analysis result into the target language and simultaneously generating and outputting the target language. The concept is clarified, the size of the analysis and generation grammar can be remarkably reduced, The processing speed can be improved, and the grammar writer who writes the grammar can write the grammar much more easily and quickly.

Description

Multilingual translation system using token separator

Since this is a trivial issue, I did not include the contents of the text.

FIG. 3 is a flow chart of a multilingual translation system according to the present invention; FIG. 4 is a flowchart of a token separation method according to the present invention; FIG.

Claims (2)

In a multilingual translation system using a conceptual structure, when a source language sentence is input, a first step (31,32) of separating a sentence input by a token separator using a reference word into tokens; A second step 33 of interpreting the primitive language using the conceptual structure; And a third step (34,35) of translating the analysis result into the target language and generating and outputting the target language. The method according to claim 1, wherein the token separation process of the first step (31, 32) comprises the steps of: searching a reference word by using a reference word dictionary (13) for each word A fourth step (42, 43) for judging whether or not there is an abnormality; If it is determined in the fourth step 42 or 43 that the reference word exists, a fifth step of separating the token using the reference word and cutting out the reference word and repeating the fourth step 42 and 43 44); A sixth step (45) of outputting a separate token list if there is no more reference word as a result of the fourth step (42, 43); And a seventh step (46) of repeating the fourth step (42,43) to the sixth step (45) for all sentences inputted. ※ Note: It is disclosed by the contents of the first application.