TWI634439B - Matching system for to-be-translated documents - Google Patents

Abstract

A mediation system for translating files includes at least one electronic device and a backend server. The electronic device includes a translation file, and the electronic device is connected to the backend server in a network connection manner. The backend server includes a language recognition unit, a category identification unit, an assigned personnel database, a translator database, a media unit, a source database, and a comparison unit. The language recognition unit generates a language analysis result based on the content of the translated file. The category identification unit includes a domain database, and the category identification unit compares the translation file with the data in the domain database to generate a category analysis result. The assigned personnel database includes the dispatch records of a number of different assignees, and the dispatch records are used to count the language and domain categories in which the assignee is good at dividing the case. The translator's database consists of translation records of a number of different translators with the translated language and domain categories that the corresponding translator has translated. The media unit combines the linguistic analysis results and the category analysis results to collect statistical records and translation records of the same language and the same field category.

Description

Translation system

The invention provides a mediation system for translating files, in particular a mediation system with a language recognition unit, a category identification unit and a comparison unit.

Nowadays, the initial job processing process of the translation case usually needs to rely on the manual judgment of the operation program. The "business personnel" is first assigned to the appropriate translation file "assignment personnel", and the "designation personnel" then assigns the translation work to the appropriate "translator". .

In detail: the "business personnel" and "designated personnel" who undertake translation cases must first make a preliminary judgment on the translation documents: determine the language of the documents and the categories of professional fields. The basis for judgment is the information provided by the customer and the knowledge and experience of the judge. If the information provided by the customer is wrong or the knowledge of the judge is insufficient, it is possible to judge the error. Judging errors can make the assigned "translator" inappropriate. In detail, each "translator" has a file type that he is really good at translating. If the "translator" hard-wires a file type that is not good at it, it will take more time and effort to translate. Quality is also prone to problems. In addition, even if the documents are judged correctly, the ability to assign documents to the most suitable "translators" will be affected by the experience of "designating staff" in handling such documents. If it is handled by an "designated person" with insufficient experience, The chances of causing problems will also be higher.

In addition, when there are a large number of or time-stressed translation files that need to be judged, it is not easy for the processing personnel to digest and correctly judge all the documents in a short time. Relying on manual judgment of the operating procedures, except In addition to making it easy to judge mistakes, the efficiency of assigning translation work will not be too good, affecting the progress of the translation case.

Therefore, how to establish an automated translation document matching system to help different types of translation files quickly and correctly find suitable "designers" and "translators" is a common knowledge in the field worth considering.

The object of the present invention is to provide a mediation system for translating documents, which can send translation files to customers, perform "language", "category" identification and "approximate document" comparison, thereby providing suitable "designers" Recommendations for "Translators" and "Approximate Documents" to improve the efficiency and correctness of internal processes, reduce costs and reduce losses caused by human factors.

A mediation system for translating files includes at least one electronic device and a backend server. The electronic device includes a translation file, and the electronic device is connected to the backend server in a network connection manner. In addition, the backend server includes a translator database, a language recognition unit, a category identification unit, and a mediation unit. Among them, the translator database includes translation records of a number of different translators, and the translation records have the translated language and domain categories that the corresponding translator has translated. The language recognition unit generates a language analysis result according to the content of the translation file, and the category identification unit includes a domain database, and the category identification unit compares the translation file with the data in the domain database to generate a category analysis result. Based on the language analysis results and the category analysis results, the media unit combines the translation records of the same language and the same field category to provide translators' recommendations.

The mediation system for translating files as described above, wherein the backend server further includes an assigned personnel database, the dispatcher database includes a plurality of dispatch records of different assignees, and the dispatch records are used to count the assignments. The staff is good at dividing the language and domain categories. In the above-mentioned mediation system for translating documents, the mediation unit combines the statistics of the language analysis result and the analysis result of the category to collect the record records of the same language and the same domain category. Provide a designated person recommendation.

The mediation system for translating the file further includes an original database and a comparison unit. The original database stores a plurality of original materials, each of which has a corresponding translator data and a corresponding Assign personnel information. The comparison unit compares the translated document with each original document and provides similar document recommendation according to the degree of approximation.

The mediation system for translating a file as described above, wherein the electronic device further comprises a display interface, and the language analysis result and the category analysis result are transmitted to the display interface.

In the mediation system for translating the file, the display interface includes a learning feedback mechanism. The learning feedback mechanism transmits a feedback message to the category identification unit according to a user, and then the learning information is imported into the domain database.

In order to make the above features and advantages more apparent, the preferred embodiments are described below, and in conjunction with the accompanying drawings.

1, 2‧‧‧Mediating system for translation documents

10‧‧‧Backend server

101, 201‧‧‧Translator Database

101D‧‧‧ translation record

102‧‧‧Language Identification Unit

102A‧‧‧Language analysis results

103‧‧‧Category Identification Unit

103B‧‧‧ Domain Database

103A‧‧‧ category analysis results

104‧‧‧Media Unit

201D‧‧‧ translation record

305‧‧‧ Designated personnel database

305D‧‧‧ record

403‧‧‧ comparison unit

405‧‧‧ original database

405A‧‧‧ original information

405B‧‧‧Translator Information

405C‧‧‧Designation of personnel information

12‧‧‧Electronic devices

120‧‧‧Translated documents

121‧‧‧Display interface

121R‧‧‧ learning feedback mechanism

1212‧‧‧Feedback

FIG. 1 is a schematic diagram showing a table of the mediation system 1 for translating files according to the embodiment.

FIG. 2 is a schematic diagram of the display interface 121.

FIG. 3 is a table diagram showing a mediation system 2 for translating files according to another embodiment.

Please refer to FIG. 1. FIG. 1 is a schematic diagram of a table of the mediation system 1 for translating files according to the embodiment. The mediation system 1 for translating files includes at least one electronic device 12 and a backend server 10. The electronic device 12 includes a translation file 120 and a display interface 121. The electronic device 12 is, for example, a desktop computer, a notebook computer, a tablet computer, and a smart phone. The translated file 120 is a translation case of a new customer. It is a document that has not been divided into "designated personnel" and "translators". In addition, the electronic device 12 is connected to the backend server 10 by means of a network connection. The backend server 10 includes a translator database 101, a language recognition unit 102, a category identification unit 103, and a mediation unit 104. . The translator database 101 includes a plurality of different translation records 101D, and the translation record 101D records that the corresponding translator has turned over. Translated language and domain categories. In detail, each translation record 101D corresponds to a translator (different translation records 101D correspond to translator A, translator B and translator C respectively), and each translator has a "language" that is good at translation. And "domain categories." Among them, common "language", such as: Chinese, English, Japanese, Korean, etc., "domain category" refers to the professional field to which the translation file 120 belongs, such as: electronics, machinery, computers, chemistry, motors, communications, Physics...etc. Further, the language recognizing unit 102 analyzes the content of the translated file 120 using the features of each language to determine the "language" to which the translated file 120 belongs. For example, if the character encoding, the punctuation, the sentence structure, and the like of the translation file 120 belong to the traditional Chinese, the language recognition unit 102 determines that the "language" to which the translation file 120 belongs is "traditional Chinese". The language analysis result 102A is transmitted to the display interface 121. For example, the proportion of traditional Chinese shown in Figure 2 accounts for 95.9% of the entire file, and the proportion of English accounts for 4.1% of the entire file. In this way, the user can also understand the language composition of the translated file 120 through the display interface 121.

In addition, the category identification unit 103 includes a domain database 103B. The domain database 103B stores template data sets belonging to various professional fields. The source of the template data sets is mainly Wikipedia. After downloading the original data, the classification is performed. After the analysis and collation of the text exploration technology, the model data sets of different professional fields are completed. The category identifying unit 103 compares the translated file 120 with the data in the domain database 103B to generate a category analysis result 103A. The category analysis result 103A will be displayed in a similar proportion to the translated domain 120, for example, the proportion of the communication/network field shown in Figure 2 is 39.8%, the proportion of the transportation field is 20.9%, and the computer/information technology The proportion of the field is 20.2%, and the proportion of the motor field is 19.1%. The category analysis result 103A is also transmitted to the display interface 121, so that the user can also understand the main area of interest of the translated file 120 through the display interface 121.

In addition, the backend server 10 of the mediation system 1 that translates files further includes an assigned personnel database 305. The assigned personnel database 305 includes a plurality of dispatching records 305D of different assigning personnel, and the dispatching record 305D is used to count the attributes of the assigning personnel to be good at dividing the case. In detail, each case record 305D corresponds to an assigned person (different dispatch records 305D correspond to assigning personnel D, assigning personnel E, and assigning personnel F respectively), and the attribute is also referred to as "language" and "category" of the translated file 120. Therefore, the merging unit combines the linguistic analysis result 102A and the category analysis result 103A to collect the genre record 305D of the same language and the same domain category. In this way, different translation files 120 can quickly find a suitable "designator" to perform the divisional work.

Thereafter, the matching unit 104 combines the translation records 101D of the same language and the same domain category according to the "language" and the field "category" of the translated document. Therefore, the translated translator 101D can find the recommended translator (translator corresponding to the translation record 101D). In this way, various types of translation files 120 can quickly find a suitable "translator" to translate, to improve the efficiency and correctness of the internal processing process, and reduce the loss of human factors.

Referring to FIG. 2, the display interface 121 further has a learning feedback mechanism 121R. The learning feedback mechanism 121R transmits a learning message 1212 to the category identifying unit 103 according to a user, and then imports a learning information into the domain database 103B. The feedback message 1212 is, for example, "correct", "error" or "unsure". In this way, the learning feedback mechanism 121R can help the category recognizing unit 103 perform self-learning to improve the correct rate of the professional domain category to which the translation file 120 belongs later.

Please refer to FIG. 3. FIG. 3 is a schematic diagram of a table of the mediation system 2 for translating files according to still another embodiment. The mediation system 2 for translating files is evolved through the mediation system 1 for translating files. The mediation system 2 for translating files further includes an original document database 405 and a comparison unit 403. The original database 405 stores a plurality of original materials 405A, and each of the original materials 405A has a corresponding translator information 405B and a corresponding assigned personnel data 405C. Moreover, the comparison unit 403 is configured to compare the degree of approximation of each original document 405A with the translation file 120, find one or more original text materials 405A of the approximate translation file 120, and simultaneously obtain the corresponding translator data. 405B and assign personnel data 405C. One or more similar original texts found 405A, which can be used by "business people" and "designated personnel": direct use plus the draft to complete the translation or provide to the translator assigned to the translation document for reference or near The original translator 405B, like the original material 405A, completes the translation. Finding the approximate source material 405A can help reduce the number of words to be translated, speed up the time for translators, and save translation costs, in addition to helping to find suitable translator candidates.

Although the present invention has been disclosed in the above preferred embodiments, it is not intended to limit the invention, and any one of ordinary skill in the art can make some modifications and refinements without departing from the spirit and scope of the invention. Therefore, the scope of the invention is defined by the scope of the appended claims.

Claims (5)

A mediation system for translating files, comprising: at least one electronic device, including a translation file; and a backend server connected to the backend server by a network connection, the backend server , comprising: a database of translators, including translation records of a plurality of different translators, the translation records having corresponding language and domain categories translated by the translator; a language recognition unit, the language recognition unit is based on the translation The content of the file generates a language analysis result; a category identification unit includes a domain database, and the category identification unit compares the translation file with the data in the domain database to generate a category analysis result; The media unit combines the translation result of the language and the analysis result of the same language and the same domain category according to the language analysis result; wherein the backend server further includes an assigned personnel database, and the designated personnel database includes multiple The record of the assignment of different assignees, which is used to count the language and field in which the assignee is good at dividing the case. do not. For example, the mediation system for translating documents according to item 1 of the patent application, wherein the media unit combines the results of the language analysis and the analysis results of the category to count the records of the same language and the same field category. The mediation system for translating documents as described in claim 1 further includes: a source database storing a plurality of source materials, each of the source materials having a corresponding translator profile and a Corresponding assigned personnel data; a comparison unit that compares the approximate degree of each original material with the translated file. The mediation system of the translation file of claim 1, wherein the electronic device further comprises a display interface, and the language analysis result and the category analysis result are transmitted to the display interface. The mediation system of the translation document described in claim 4, wherein the display interface comprises a learning feedback mechanism, and the learning feedback mechanism transmits a feedback message to the category identification unit according to a user, and then learns a learning information. Into the field database.