CN105843802A - Corpus intervention module and method in translation - Google Patents

Abstract

The invention discloses a module and method for intervening corpus in translation. The purpose is to realize corpus retrieval and comparison, and the matching corpus can be easily involved in translation, thereby reducing translation time and improving the consistency of expression in translation. The technical solution is: use the corpus reading module to selectively read the historical corpus and the corpus prepared for translation activities; use the translation material reading module to open the materials that need to be translated, and perform sentence processing on the materials that need to be translated; corpus and The translation material retrieval matching module searches for the maximum corpus matching sentence by sentence for the read and processed materials that need to be translated, and finally obtains the position of the matching corpus in the text and the interpretation of the corpus, and displays the matching corpus and corpus through the matching corpus display module. The difference between the translated texts is displayed; finally, the matching corpus translation module is used to copy the matched corpus translation, and paste it in the selected position in the translation, so as to realize the intervention of translation.

Description

Intervention module and method of corpus in translation

technical field

The invention belongs to the technical field of computational linguistics and translation, and in particular relates to a corpus intervention module and method in translation.

Background technique

The corpus comes from the Latin word corpus, which originally means "summary", "collection", etc., and the plural form is corpora or corpuses. A corpus is "a collection of works, and any textual collection on a subject" (OED) is a "collection of written or spoken material that provides a basis for linguistic analysis" (OED). A corpus is "a collection of language-use materials selected and ordered according to explicit linguistic criteria, and intended to be used as a sample of language" (Sinclair, 1986:185-203). A corpus is a large text library assembled for a specific purpose according to clear design criteria (Atkins and Clear, 1992: 1-16). Renouf believes that a corpus is "a text library composed of a large collection of written or spoken language, stored and processed by a computer, and used for linguistic research" (Renouf, 1987: 1). Leech pointed out that a large collection of machine-readable electronic texts is the basis for obtaining "required frequency data" in probabilistic research methods. "To obtain the necessary frequency data, we must backwash a sufficient amount of natural English (or other language) texts, In order to make realistic predictions based on the observed frequency.

Therefore, there is a need for a reliable machine-readable collection of electronic texts, that is, a machine-readable corpus” (leech, 1987: 2). In summary, a corpus has the following basic characteristics:

1) The design and construction of the corpus is carried out under the principles of systematic theoretical linguistics, and the development of the corpus has clear and specific research goals. For example, the main purpose of the BROWN corpus in the early 1960s was to analyze the grammar of American English, and the subsequent LOB corpus basically collected British English of the same era according to the design principles of the BROWN corpus, and the purpose was to compare American English and British English. Analysis and parsing.

2) The composition and sampling of the corpus is to collect the corpus according to clear linguistic principles and adopt a random sampling method, rather than simply accumulating corpus. The collected corpus must be natural-occurred data of language use.

3) As a sample of natural language use, the corpus must be representative. Chomsky once criticized that the corpus is just trying to use a small sample to represent a huge amount of even unlimited actual language materials, and the result is bound to be biased and lacks representativeness. Description will be but a vocabulary" (Chomsky, 1962:159). This critique is valuable to any study based on methods of probability and statistics (McEnery, 1996: 5).

Li Wenzhong believes that the corpus text is a continuous text or discourse fragment (running text or continuous stretches of discourse), rather than encouraging sentences and vocabulary. In corpus research, grammatical relationships, usage, and bulk observations of a search term are made by analyzing the context provided.

At present, the research on corpus is more theoretical, serving the research of corpus translation, and does not involve specific practical applications; the choice of corpus is a research corpus, and most of them are not corpora that can be directly used in specific translation practice; specific translation practice In , there is no specific mention of how the corpus intervenes in translation, or how the corpus forms a help to translation. At present, in the translation industry, there is no relatively mature terminology intervention tool, which is usually manual reference, and the efficiency is low.

Contents of the invention

In order to solve the problems in the prior art, the present invention proposes a translation process that can realize corpus retrieval and comparison during translation, and the matched corpus can be easily involved in translation, thereby reducing translation time and improving the consistency of expression in translation. Corpus intervention modules and methods.

In order to achieve the above object, the technical solution adopted in the present invention is:

A corpus intervention module in translation, comprising:

Corpus reading module: used to selectively read historical corpora and corpora prepared for translation activities;

Translation material reading module: used to open the material that needs to be translated, read the material that needs to be translated, and process the material that needs to be translated by sentence;

Corpus and translation material retrieval matching module: used to search for the largest corpus match sentence by sentence starting from the first word for the read and sentence-processed materials that need to be translated, and finally obtain the position and corpus of the matching corpus in the text Interpretation;

Matching corpus display module: used to display the matching corpus and the translation of the corpus;

Matching corpus intervention translation module: used to copy the matched corpus translation and paste it in the selected position in the translation, so as to realize the intervention of translation.

A method for corpus intervention in translation, comprising the following steps:

1) The translation material reading module opens the materials that need to be translated, reads the materials that need to be translated, and performs sentence processing on the materials that need to be translated. At the same time, the corpus reading module selectively reads the historical corpus and the corpus prepared for translation activities ;

2) The corpus and translation material retrieval matching module searches for the largest corpus matching sentence by sentence from the first word to the read and sentence-processed materials that need to be translated, and finally obtains the position of the matching corpus in the text and the interpretation of the corpus; and Display the matching corpus and the translation of the corpus through the matching corpus display module;

3) The matching corpus intervention translation module copies the matched corpus translation, and pastes it at a selected location in the translation, so as to realize the corpus intervention in translation.

In the described step 1), the translation material reading module calls the Com interface of Word to the word board and the Word document to obtain the text in the word; the Com interface of the excel document to call excel obtains the text in the excel form.

In the step 1), the translation material reading module defines a sentence terminator according to the punctuation rules, divides the material to be translated into sentences, and judges it as the end of a sentence when the terminator is encountered.

Described translation material reading module needs to judge whether English full stop is abbreviation punctuation, contains acronym in the thesaurus, searches full stop and the word before the full stop in the thesaurus, if can search then be acronym punctuation, is ignored as a sentence terminator.

The corpus reading module in the step 1) saves the corpus read in the historical corpus and the corpus prepared for translation activities in list form, and sorts the corpus alphabetically.

Described step 2) in the corpus and the translation material retrieval matching module, the specific steps of the matching of the material that needs translation include:

2.1) Take a word to a word group, and search the word group in the corpus list;

2.2) If a fully matched corpus is found, save the information of the corpus; continue to step 2.1) to search for a larger match;

2.3) If a sub-match is found in the search, that is, the word group is a part of the corpus, then go to step 2.1) to continue searching;

2.4) If no match is found, clear the word group, and go to step 2.1) from the last matched word group until all translation materials are searched.

In the step 2), the matching corpus display module displays the translation of the calibrated matching corpus through a floating window or a symbolic label, and the translation can be edited.

Compared with the prior art, the present invention utilizes the corpus reading module to selectively read the historical corpus and the corpus prepared for translation activities; utilizes the translation material reading module to open the material that needs to be translated, reads the material that needs to be translated, and The materials that need to be translated are processed in sentences; the corpus and translation material retrieval matching module searches for the largest corpus match sentence by sentence from the first word of the read and translated materials that need to be translated, and finally obtains the matching corpus in the text position and corpus interpretation, and display the matching corpus and the translation of the corpus through the matching corpus display module; finally, copy the matching corpus translation through the matching corpus intervention translation module, and paste it at the selected position in the translation, so as to realize the translation Intervention of translation. It can realize corpus retrieval and comparison during translation, and the matched corpus can be easily involved in translation, thus reducing translation time and improving the consistency of expression in translation.

Furthermore, the translation material reading module defines sentence terminators according to the rules of punctuation marks, divides the material to be translated into sentences, and judges it as the end of a sentence when a terminator is encountered. For English periods, it needs to judge whether it is an abbreviation punctuation, thesaurus contains acronyms, search the period and the word before the period in the thesaurus, if found, it is an acronym punctuation, and it is ignored and not used as a sentence terminator, which further improves the accuracy of the sentence processing of the translation material reading module and improve the translation efficiency.

Further, the corpus reading module can selectively read the historical corpus and the corpus specially prepared for this translation activity, or mainly read the corpus prepared for this translation activity, and use the historical corpus as an auxiliary reference to read, The read corpus is saved in a list, and the corpus is sorted alphabetically, which can improve the efficiency of corpus matching and search, thereby reducing translation time.

Further, the corpus and translation material retrieval matching module adopts the principle of maximum corpus matching for the matching of the materials to be translated, which can better match the corpus of the materials to be translated, and further improve the efficiency of the present invention.

detailed description

The present invention will be further explained below in conjunction with specific examples.

The present invention is made of five modules:

Module 1: Corpus reading module: You can selectively read the historical corpus and the corpus specially prepared for this translation activity, or you can read the corpus prepared for this translation activity mainly, and use the historical corpus as an auxiliary reference to read Pick. The read corpus is saved in a list, and the corpus is sorted alphabetically to improve the efficiency of corpus matching search;

Module 2: Translation material reading module: Open the material that needs to be translated, and process the material into sentences while opening the material. According to punctuation marks and rules, the English text is divided into sentences one by one, and sentence terminators are defined, such as English full stops, exclamation marks, question marks, etc. When encountering a terminator, it is judged as the end of a sentence, and the English full stop needs to be judged whether it is an abbreviation , the thesaurus contains acronyms, search for the period and the word before the period in the thesaurus, if it can be found, it will be an acronym punctuation, and it will be ignored and not used as a sentence terminator;

Module 3: Retrieval and matching of corpus and translation materials Module: For the translation materials that have been read and processed sentence by sentence, search for the largest corpus match sentence by sentence starting from the first word, and finally get the position of the matching corpus in the text and the corpus (corpus + Interpretation); specifically include: (1) get a word to word group, and search the word group in the corpus list; (2) if a fully matched corpus is found, save the information of the corpus (position+corpus+paraphrase), and continue to go to Step (1) Search for a larger match; (3) If a sub-match is found (the phrase is part of the corpus), go to step (1); (4) If no match is found, the phrase is cleared, starting from the last Start to go to step (1) after a matching phrase group, until all translation materials have been searched;

Module 4: Matching corpus display module: All tagged corpus are already matched corpus. When translating the sentence, there are many ways to display it:

1) Display method 1: The color of the matching corpus is displayed (the color can be set, and two colors can be set to distinguish between the corpus prepared for this translation activity and the corpus in the historical corpus), when the mouse is placed on the corpus , a text box of the translation of the corpus appears next to the mouse. When the mouse moves over the text box, you can choose to copy the translation. If the mouse leaves the text box, the text box will exit;

2) Display method 2: Display the color of the matching corpus (the color can be set, and two colors can be set to distinguish between the corpus prepared for this translation activity and the corpus in the historical corpus), and the translation of the corpus can be directly used in the device The specified symbols are displayed directly behind the corpus;

3) Display method 3: The color of the matching corpus is displayed (the color can be set, and two colors can be set to distinguish between the corpus prepared for this translation activity and the corpus in the historical corpus), and the translation of the corpus is displayed in a floating position Above the corpus, when the data is moved to the translation, the corpus can be edited, for example, the content of the translation can be copied and changed;

Module 5: Matching corpus to intervene in translation module: The translation of corpus with different display methods can be copied, and then pasted at a selected location in the translation, so as to achieve intervention in translation.

The complete steps of the inventive method:

Open the text to be translated in the tool interface (the format can be Word, Excel, Notepad, WordPad, etc.), directly use the general file reading module to obtain the text of the text file, and the WordPad and Word documents call the Com interface of Word to obtain the text in Word , excel calls the Com interface of excel to obtain the text in the excel table; then click "corpus intervention" (the corpus is a historical corpus or a corpus specially made for this project), and select the corpus according to the prompts (the list of corpus forms is displayed in two columns, and the left column is The corpus, the right column is the definition) file, call the corpus and translation material retrieval matching module to obtain the matching corpus information;

There are two options for matching display. 1) Use special symbols directly for the matched corpus, such as [] to display the definition, search the matching module according to the corpus and translation materials to get the position of the matching corpus in the translated text, and insert the corpus in order to simplify Influenced by the position of the text, the translation text is inserted into the paraphrase of the matching corpus from the back to the front;

2) The hovering display on the corpus, when the mouse moves over the corpus, the stay time exceeds the set value (default is 3 seconds), the position of the mouse is obtained, and the sentence is obtained according to the position, and the sentence is retrieved through the corpus and translation materials The matching module obtains the matching corpus of the modified sentence and displays it in a floating window where the mouse is located;

Directly copy the paraphrase displayed by the two methods, paste it to the translation position, and complete the inclusion of the corpus in the translation.

The invention can realize corpus retrieval and comparison during translation, and the matched corpus can be easily involved in translation, thereby reducing translation time and improving expression consistency in translation.

Claims (8)

1. A corpus intervention module in translation, characterized in that, comprising: Corpus reading module: used to selectively read historical corpora and corpora prepared for translation activities; Translation material reading module: used to open the material that needs to be translated, read the material that needs to be translated, and process the material that needs to be translated by sentence; Corpus and translation material retrieval matching module: used to search for the largest corpus match sentence by sentence starting from the first word for the read and sentence-processed materials that need to be translated, and finally obtain the position and corpus of the matching corpus in the text Interpretation; Matching corpus display module: used to display the matching corpus and the translation of the corpus; Matching corpus intervention translation module: used to copy the matched corpus translation and paste it in the selected position in the translation, so as to realize the intervention of translation. 2. A method for intervening corpus in translation, comprising the following steps: 1) The translation material reading module opens the materials that need to be translated, reads the materials that need to be translated, and performs sentence processing on the materials that need to be translated. At the same time, the corpus reading module selectively reads the historical corpus and the corpus prepared for translation activities ; 2) The corpus and translation material retrieval matching module searches for the largest corpus matching sentence by sentence from the first word to the read and sentence-processed materials that need to be translated, and finally obtains the position of the matching corpus in the text and the interpretation of the corpus; and Display the matching corpus and the translation of the corpus through the matching corpus display module; 3) The matching corpus intervention translation module copies the matched corpus translation, and pastes it at a selected location in the translation, so as to realize the corpus intervention in translation. 3. the intervening method of corpus in a kind of translation according to claim 2, is characterized in that, in described step 1), translation material reading module transfers the Com interface of Word to writing board, Word document and obtains the text in word; Call the excel Com interface on the excel document to get the text in the excel table. 4. a kind of corpus intervening method in translation according to claim 3, is characterized in that, in described step 1), translation material reading module defines sentence terminator according to punctuation mark rule, and the material that needs translation is segmented is a sentence, and it is judged as the end of a sentence when a terminator is encountered. 5. a kind of corpus intervening method in the translation according to claim 4, it is characterized in that, described translation material reading module needs to judge whether it is abbreviation punctuation to English period, comprises acronym in the thesaurus, in Search the thesaurus for the period and the word before the period. If it can be found, it will be an acronym punctuation, and it will be ignored and not used as a sentence terminator. 6. a kind of corpus intervention method in translation according to claim 2, is characterized in that, in described step 1), corpus reading module reads in the corpus of historical corpus and for the corpus that translation activity prepares with list Forms are saved and the corpus is sorted alphabetically. 7. a kind of corpus intervening method in translation according to claim 2, is characterized in that, described step 2) in the corpus and translation material retrieval matching module, the concrete steps of the matching of the material that needs translation comprise: 2.1) Take a word to a word group, and search the word group in the corpus list; 2.2) If a fully matched corpus is found, save the information of the corpus; continue to step 2.1) to search for a larger match; 2.3) If a sub-match is found in the search, that is, the word group is a part of the corpus, then go to step 2.1) to continue searching; 2.4) If no match is found, clear the word group, and go to step 2.1) from the last matched word group until all translation materials are searched. 8. A method for intervening corpus in translation according to claim 2, characterized in that the matching corpus display module in the step 2) displays the translation of the marked matching corpus through a floating window or a symbolic label form, and the translation able to edit.