Deep processing method for characters of document

Abstract

The invention discloses a deep processing method for characters of a document. The deep processing method comprises the following steps: (1) extracting character information of the document to be translated; (2) splitting the document into a data set of characters to be translated with minimum unit and eliminating data of non-translation characters and repeated characters to be translated; (3) establishing the document processed before translation, firstly replicating the data of the characters to be translated into an original text, then writing the original text and a translated text relating to special terms into corresponding special term lists, thus obtaining the document processed before translation and with special terms; (4) replacing character data of the original text in the special term list included in an original text list in the document processed before translation with character data of the translated text in the special term list, and secondarily processing the character data to obtain a final document processed before translation; (5) translating the original text list by a translator; and (6) replacing the original text with the translated text to obtain the translated text. The deep processing method disclosed by the invention has the advantages that repeated single contents in the document can be deeply eliminated in advance before the document is processed, so that the purpose of improving the translation efficiency can be achieved.

Description

technical field [0001] The invention relates to the technical field of translation, in particular to a method for in-depth processing of document text. Background technique [0002] Since the mid-1980s, based on the extensive use of corpus and multi-engine machine translation methods, the performance and efficiency of translation software have been significantly improved, and various translation software have sprung up. Translating with pre-written software programs greatly improves the translation speed of texts. However, due to the particularity of language expression, the translation quality of translation software has been repeatedly criticized. The principle of translation software is to store the semantics of the two languages ​​in one-to-one correspondence, and replace them mechanically during translation. Due to the diversity of language expressions, each word , Words, phrases or single sentences often correspond to more than one meaning, and the translation obtaine...

AI Technical Summary

Problems solved by technology

[0003] In the existing technology, for a project or a long document, it is often divided into multiple parts in a team for translation, but due to the different translation habits of translators, different translators often translate a sentence with the same meaning, resulting in inconsistent translations Happening

In addition, this method of dividing the team into multiple translations causes the translators to repeatedly translate words, phrases or single sentences with the same meaning, which not only greatly increases the translation intensity of the translators, but also greatly reduces the translation efficiency

Images