JP2000339313A - Device and method for improving dictionary data and information recording medium - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a dictionary data improving device and a dictionary data improving method which use a great deal of parallel translation documents and can register the KATAKANA (Japanese square syllabary) notation of the reading of English words used as translated words of the English words with a dictionary database for translation and to provide an information recording medium. SOLUTION: This device has a used word collating part 8 that confirms whether or not a word of an object language being a translated word candidate of each morpheme of an original language and the KATAKANA notation word of reading of an original language morpheme exist in a morphological list of the object language, a KATAKANA notation data acquiring part 9 which acquires the KATAKANA notation word of the reading of a designated word and a dictionary data changing part 12 which changes data about the designated word of a dictionary database 13 for translation, performs processing with a designated parallel translation database as an object, extracts the data with which translated word information is changed by a translated word information improvement data extracting part 11 and changes the data of the database 13 with the dictionary data changing part 12.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a dictionary data improving apparatus, a dictionary data improving method, and an information recording medium in machine translation for translating a language (source language) into a different language (target language).

[0002]

2. Description of the Related Art One of the reasons that a user determines that a translation generated by machine translation is inappropriate is that a translation used by a human for translation is not registered in a dictionary for machine translation. Can be In recent years, katakana has tended to be frequently used in Japanese.
In many cases, katakana words that are not registered in the dictionary for machine translation are used, but at present, there is no method of reviewing the translated words other than manually improving individual data.

[0003]

However, when the dictionary data is manually improved, the amount of work increases in proportion to the number of words, so that an enormous amount of work is required. This work also requires expertise in the language.

The present invention takes advantage of the fact that katakana words, which are frequently used in Japanese, are often represented by katakana readings of corresponding English words. Using a large number of bilingual documents that describe the same content in the language, automatically extract katakana transcript data of reading English words that can be registered as translations, and improve the translation of machine translation dictionaries An object of the present invention is to provide a dictionary data improving device, a dictionary data improving method, and an information recording medium.

[0005]

In order to solve this problem, a dictionary data improving apparatus according to the present invention comprises an input means for designating a target document and an execution instruction, and a sentence for dividing the text of a bilingual text file into sentences. A certification processing unit, a target language analysis word dictionary storing information on words required for analysis of the target language, a translation word dictionary storing information on words necessary for translation from the source language to the target language, A target language morphological analysis unit that divides sentences written in the target language into morpheme units by referring to the language analysis word dictionary, and divides sentences written in the source language into morpheme units by referring to the translation word dictionary A morphological analysis processing unit, a morphological list storage unit for storing morphological analysis results of text files described in the source language and the target language, and a target which is a candidate for a translation of each morpheme in the source language. Utilization translation unit that checks whether the katakana notation of the word of the word and the source language morpheme exists in the morpheme list of the target language, and the katakana notation data acquisition that acquires the katakana notation of the reading of the specified word A translated word appearance data storage unit for storing a translated word matching result;
Composes a translated word information improvement data extraction unit that extracts dictionary registered words in the source language for adding and changing translation words, a dictionary data change unit that changes data related to specified words in the translation dictionary database, and a translation word dictionary A translation dictionary database containing katakana notation data of source language word reading, display means for displaying data, etc., a sentence recognition processing unit, a target language morphological analysis processing unit, a source language morphological analysis processing unit, and a used translation word Controls the functions of the collation unit, katakana notation data acquisition unit, translated word information improvement data extraction unit, dictionary data change unit, and the operation between the functions, and controls the flow of data stored in the morpheme list storage unit and translated word appearance data storage unit A translated word information improvement data extraction unit, which reads, from the translated word appearance data storage unit, the form of the source language in which the katakana notation of the reading of the English word is used. In arsenide, and in all data translation appearance data storage unit, it extracts the data about morphemes using the number of data of the katakana words exceeds a predetermined number.

With this configuration, taking advantage of the fact that katakana words frequently used in Japanese are often expressed in katakana readings of corresponding English words, two types of target machine translation source language and target language are used. Using a large number of bilingual documents that describe the same content in a language and can be registered as a translation,
A dictionary data improving apparatus capable of automatically extracting katakana notation word data of reading an English word and improving the translation of a machine translation dictionary can be realized.

[0007]

According to the first aspect of the present invention, there is provided an input means for designating a target document and giving an execution instruction, a sentence recognition processing unit for dividing the text of a bilingual text file for each sentence, and an analysis of a target language. Refer to the target language analysis word dictionary that stores information about the words necessary for the translation, the translation word dictionary that stores information about the words necessary for translation from the source language to the target language, and the target language analysis word dictionary. A target language morphological analysis processing unit that divides a sentence described in the target language into morpheme units, and a source language morphological analysis processing unit that divides a sentence written in the source language into morpheme units with reference to a translation word dictionary, A morpheme list storage unit that stores morphological analysis results of text files described in the source language and the target language, and a target language word that is a candidate for a translation of each morpheme in the source language and a reading format of the source language morpheme. Na notation word, and use translation collating unit for confirming whether or not present in the morpheme list of the target language,
A katakana notation data acquisition unit that acquires katakana notation words for specified word readings, a translated word appearance data storage unit that stores translated word matching results, and a translated word information improvement that extracts the source language dictionary registered words that add or change the translated words A data extraction unit, a dictionary data change unit that changes data related to a specified word in the translation dictionary database, and a translation dictionary database that includes katakana notation data of source language words that constitute a translation word dictionary. Display means for displaying data, etc., sentence recognition processing unit, target language morphological analysis processing unit, source language morphological analysis processing unit, translation word matching unit, katakana notation data acquisition unit, translation word information improvement data extraction unit, dictionary data change And a control unit that controls the flow of data stored in the morpheme list storage unit and the translated word appearance data storage unit, as well as the functions of the units and the operations between the functions. The translated word information improved data extracting unit reads the katakana notation from the translated word appearance data storage unit with the morpheme of the source language in which the katakana notation word of the English word is used, and in all the data in the translated word appearance data storage unit. A dictionary data improving apparatus for extracting data relating to morphemes in which the number of words used exceeds a certain number.

With this configuration, in an actual translation,
When a katakana notation of the source language morpheme is used as a translation of a certain source language morpheme, the katakana notation used in the actual translation is automatically used as a translation of the source language morpheme in the translation word dictionary database. You can register.

According to a second aspect of the present invention, there is provided an input means for designating a target document and giving an execution instruction, a sentence recognition processing unit for dividing the text of a bilingual text file for each sentence, and a word necessary for analyzing a target language. Refer to the target language analysis word dictionary that stores information related to the target language, the translation word dictionary that stores information related to words required for translation from the source language to the target language, and the target language analysis word dictionary, and describe in the target language. A target language morphological analysis processing unit that divides the sentence into morpheme units,
A source language morphological analysis processing unit that divides a sentence described in a source language into morpheme units by referring to a translation word dictionary, and stores a morpheme list that stores morphological analysis results of a text file described in a source language and a target language. A target translation word that checks whether or not katakana notation of the target language word and the source language morpheme reading that is a candidate for a translation of each morpheme of the source language exists in the morpheme list of the target language; A katakana notation data acquisition unit that acquires katakana notation words of word reading, a translated word appearance data storage unit that stores translation word matching results,
Composes a translated word information improvement data extraction unit that extracts dictionary registered words in the source language for adding and changing translation words, a dictionary data change unit that changes data related to specified words in the translation dictionary database, and a translation word dictionary A translation dictionary database containing katakana notation data of source language word reading, display means for displaying data, etc., a sentence recognition processing unit, a target language morphological analysis processing unit, a source language morphological analysis processing unit, and a used translation word Controls the functions of the collation unit, katakana notation data acquisition unit, translated word information improvement data extraction unit, dictionary data change unit, and the operation between the functions, and controls the flow of data stored in the morpheme list storage unit and translated word appearance data storage unit A translated word information improvement data extraction unit, which reads, from the translated word appearance data storage unit, the form of the source language in which the katakana notation of the reading of the English word is used. Acquiring all data related to the element, and extracting all katakana notation of the reading of the English word as a registered translation when all the candidate words are not used as the translation of the source language morpheme in the acquired data. This is a dictionary data improvement device.

With this configuration, in an actual translation,
If the translation candidate of a certain source language morpheme is not used and the katakana notation of the reading of the source language morpheme is used, the katakana notation used in the actual translation is converted into the source language in the translation word dictionary database. It can be automatically registered as a morpheme translation.

According to a third aspect of the present invention, there is provided an input means for designating a target document and giving an execution instruction, a sentence recognition processing unit for dividing the text of a bilingual text file for each sentence, and a word necessary for analyzing a target language. Refer to the target language analysis word dictionary that stores information related to the target language, the translation word dictionary that stores information related to words required for translation from the source language to the target language, and the target language analysis word dictionary, and describe in the target language. A target language morphological analysis processing unit that divides the sentence into morpheme units,
A source language morphological analysis processing unit that divides a sentence described in a source language into morpheme units by referring to a translation word dictionary, and stores a morpheme list that stores morphological analysis results of a text file described in a source language and a target language. A target translation word that checks whether or not katakana notation of the target language word and the source language morpheme reading that is a candidate for a translation of each morpheme of the source language exists in the morpheme list of the target language; A katakana notation data acquisition unit that acquires katakana notation words of word reading, a translated word appearance data storage unit that stores translation word matching results,
Composes a translated word information improvement data extraction unit that extracts dictionary registered words in the source language for adding and changing translation words, a dictionary data change unit that changes data related to specified words in the translation dictionary database, and a translation word dictionary A translation dictionary database containing katakana notation data of source language word reading, display means for displaying data, etc., a sentence recognition processing unit, a target language morphological analysis processing unit, a source language morphological analysis processing unit, and a used translation word Controls the functions of the collation unit, katakana notation data acquisition unit, translated word information improvement data extraction unit, dictionary data change unit, and the operation between the functions, and controls the flow of data stored in the morpheme list storage unit and translated word appearance data storage unit A translated word information improvement data extraction unit, which reads, from the translated word appearance data storage unit, the form of the source language in which the katakana notation of the reading of the English word is used. Acquire all data relating to oxygen, and classified by using translation of the data, a dictionary data improved apparatus and extracting the katakana notation word reading English words by referring to the frequency information as a registered translation.

With this configuration, in an actual translation,
When a katakana notation of the reading of a source language morpheme is frequently used as a translation of a certain source language morpheme, the katakana notation used in the actual translation is replaced with the katakana notation of the source language morpheme in the word dictionary database for translation. It can be automatically registered as a translation.

According to a fourth aspect of the present invention, there is provided an input means for designating a target document and instructing execution, a sentence recognition processing unit for dividing the text of a bilingual text file for each sentence, and a word necessary for analyzing a target language. Refer to the target language analysis word dictionary that stores information related to the target language, the translation word dictionary that stores information related to words required for translation from the source language to the target language, and the target language analysis word dictionary, and describe in the target language. A target language morphological analysis processing unit that divides the sentence into morpheme units,
A source language morphological analysis processing unit that divides a sentence described in a source language into morpheme units by referring to a translation word dictionary, and stores a morpheme list that stores morphological analysis results of a text file described in a source language and a target language. A target translation word that checks whether or not katakana notation of the target language word and the source language morpheme reading that is a candidate for a translation of each morpheme of the source language exists in the morpheme list of the target language; A katakana notation data acquisition unit that acquires katakana notation words of word reading, a translated word appearance data storage unit that stores translation word matching results,
Composes a translated word information improvement data extraction unit that extracts dictionary registered words in the source language for adding and changing translation words, a dictionary data change unit that changes data related to specified words in the translation dictionary database, and a translation word dictionary A translation dictionary database containing katakana notation data of source language word reading, display means for displaying data, etc., a sentence recognition processing unit, a target language morphological analysis processing unit, a source language morphological analysis processing unit, and a used translation word Controls the functions of the collation unit, katakana notation data acquisition unit, translated word information improvement data extraction unit, dictionary data change unit, and the operation between the functions, and controls the flow of data stored in the morpheme list storage unit and translated word appearance data storage unit And a control unit for registering the katakana notation extracted by the translation information improvement data extracting unit as a translation. A dictionary data improved apparatus and registers as a translation.

With this configuration, in an actual translation,
If a katakana spelling of a source language morpheme is frequently used as a translation of a source language morpheme, register the katakana spelling as the first priority translation of the source language morpheme in the translation dictionary database. In, the translation words used for the actual translation can be automatically registered in the dictionary data,
The dictionary data can be changed so that the katakana notation is preferentially used as a frequently used translation.

According to a fifth aspect of the present invention, there is provided an input means for designating a target document and giving an execution instruction, a sentence recognition processing section for dividing the text of a bilingual text file into sentences, and a word necessary for analyzing a target language. Refer to the target language analysis word dictionary that stores information related to the target language, the translation word dictionary that stores information related to words required for translation from the source language to the target language, and the target language analysis word dictionary, and describe in the target language. A target language morphological analysis processing unit that divides the sentence into morpheme units,
A source language morphological analysis processing unit that divides a sentence described in a source language into morpheme units by referring to a translation word dictionary, and stores a morpheme list that stores morphological analysis results of a text file described in a source language and a target language. A target translation word that checks whether or not katakana notation of the target language word and the source language morpheme reading that is a candidate for a translation of each morpheme of the source language exists in the morpheme list of the target language; A katakana notation data acquisition unit that acquires katakana notation words of word reading, a translated word appearance data storage unit that stores translation word matching results,
Composes a translated word information improvement data extraction unit that extracts dictionary registered words in the source language for adding and changing translation words, a dictionary data change unit that changes data related to specified words in the translation dictionary database, and a translation word dictionary A translation dictionary database containing katakana notation data of source language word reading, display means for displaying data, etc., a sentence recognition processing unit, a target language morphological analysis processing unit, a source language morphological analysis processing unit, and a used translation word Controls the functions of the collation unit, katakana notation data acquisition unit, translated word information improvement data extraction unit, dictionary data change unit, and the operation between the functions, and controls the flow of data stored in the morpheme list storage unit and translated word appearance data storage unit A translated word information improvement data extraction unit, which reads, from the translated word appearance data storage unit, the form of the source language in which the katakana notation of the reading of the English word is used. All data about the element is obtained, the data is classified according to the field, and if the katakana notation of reading the English word is used as the translation of the source language morpheme in the data of the specific field, the data is The dictionary data improving apparatus is characterized in that the technical term dictionary translated word information is extracted as data to be changed.

According to this configuration, in a specific field, when a katakana notation for reading an English word is used as a translation of a source language morpheme, the katakana notation used for actual translation is converted into a word dictionary for translation. It can be automatically registered as a translation of the source language morpheme in the technical term dictionary data of the database.

According to a sixth aspect of the present invention, using a large number of bilingual file databases, for each bilingual text file, a source language text file is morphologically analyzed to create a morpheme list listing morphemes and candidate translations, Morphological analysis of the target language text file to create a morpheme list listing the morphemes, and whether the target language morpheme list contains a translation of each morpheme in the source language morpheme list and katakana notation of the reading of the source language morpheme For the morphemes of the source language in which the translation candidate is not used and the katakana notation of the reading is used, refer to the usage translation matching result of the all-translation text file and use the data of the katakana notation. If the number exceeds a certain number, register the katakana notation as a translation of the source language morpheme in the translation dictionary database A dictionary data improvement wherein the Rukoto.

With this configuration, in an actual translation,
When a katakana notation of the source language morpheme is used as a translation of a certain source language morpheme, the katakana notation used in the actual translation is automatically used as a translation of the source language morpheme in the translation word dictionary database. You can register.

The invention according to claim 7 uses a large number of bilingual file databases, morphologically analyzes the source language text file in each bilingual text file, and creates a morpheme list listing morphemes and candidate translations. Morphological analysis of the target language text file to create a morpheme list listing the morphemes, and whether the target language morpheme list contains a translation of each morpheme in the source language morpheme list and katakana notation of the reading of the source language morpheme For the morphemes in the source language where the candidate translation is not used and the katakana spelling of the reading is used, the data of the corresponding source language morpheme is extracted from the results of the usage translation matching in all the translated text files If all the candidate words are not used in all the extracted data, A dictionary data improved method and registers the katakana words as translation word morphemes.

With this configuration, in an actual translation,
If the translation candidate of a certain source language morpheme is not used and the katakana notation of the reading of the source language morpheme is used, the katakana notation used in the actual translation is converted into the source language in the translation word dictionary database. It can be automatically registered as a morpheme translation.

The invention according to claim 8 uses a large number of bilingual file databases, morphologically analyzes a source language text file in each bilingual text file, and creates a morpheme list listing morphemes and candidate translations. Morphological analysis of the target language text file to create a morpheme list listing the morphemes, and whether the target language morpheme list contains a translation of each morpheme in the source language morpheme list and katakana notation of the reading of the source language morpheme For the morphemes in the source language where the candidate translation is not used and the katakana spelling of the reading is used, the data of the corresponding source language morpheme is extracted from the results of the usage translation matching in all the translated text files Then, the data is classified for each used translation, and the katakana notation of the reading of the English word is used as the translation by referring to the frequency information. Extracts data to be recorded, a dictionary data improved method and registers the katakana words as translation of raw language morpheme translation dictionary database.

With this configuration, in an actual translation,
When a katakana notation of the reading of a source language morpheme is frequently used as a translation of a certain source language morpheme, the katakana notation used in the actual translation is replaced with the katakana notation of the source language morpheme in the word dictionary database for translation. It can be automatically registered as a translation.

According to a ninth aspect of the present invention, a large number of bilingual file databases are used, and in each bilingual text file, a source language text file is morphologically analyzed to form a morpheme list listing morphemes and translation candidates. Morphological analysis of the target language text file to create a morpheme list listing the morphemes, and whether the target language morpheme list contains a translation of each morpheme in the source language morpheme list and katakana notation of the reading of the source language morpheme For the morphemes in the source language where the candidate translation is not used and the katakana spelling of the reading is used, the data of the corresponding source language morpheme is extracted from the results of the usage translation matching in all the translated text files Then, based on the usage of katakana notation words and candidate translations for reading English words, the katakana table is used as a translation for the source language. When registering a word, a dictionary data improved method and registers as a first priority translation data on raw language translation dictionary database.

With this configuration, in an actual translation,
If a katakana spelling of a source language morpheme is frequently used as a translation of a source language morpheme, register the katakana spelling as the first priority translation of the source language morpheme in the translation dictionary database. In, the translation words used for the actual translation can be automatically registered in the dictionary data,
The dictionary data can be changed so that the katakana notation is preferentially used as a frequently used translation.

According to a tenth aspect of the present invention, a large number of bilingual file databases are used, and in each bilingual text file, a source language text file is morphologically analyzed.
Create a morpheme list listing morphemes and candidate translations, morphologically analyze the target language text file, create a morpheme list listing the morphemes, translate the morphemes in the source language morpheme list, and read katakana for the source language morphemes Use translated words in the target language morpheme list to check whether or not the target word is present in the target language morpheme list. When the data of the corresponding source language morpheme is extracted from the collation result, the data is classified according to the field, and the katakana notation of reading an English word is used as the translation of the source language morpheme in the data of the specific field. Register the katakana notation as a translation of the source language morpheme in the data of the technical term dictionary of the field in the translation dictionary database A dictionary data improvement wherein the Rukoto.

According to this configuration, in a specific field, when a katakana notation for reading an English word is used as a translation of a source language morpheme, the katakana notation used for actual translation is converted into a word dictionary for translation. It can be automatically registered as a translation of the source language morpheme in the technical term dictionary data of the database.

[0027] According to an eleventh aspect of the present invention, there is provided a step of extracting one source language text file, a step of performing sentence recognition in accordance with the written language, and a step of referring to a translation word dictionary to store one source language text file. A morpheme analysis process for each sentence, storing a morpheme list in which morpheme candidates are listed in order of priority in each morpheme in a morpheme list storage unit, and a step of extracting a corresponding target language text file; Performing a sentence recognition, referring to the target language analysis word dictionary, performing a morphological analysis of the target language text one sentence at a time, and storing a morpheme list in a morpheme list storage unit; The translation and collation of the source language morpheme are checked for katakana notation in the target language's morpheme list. Storing the data in the translated word appearance data storage unit, obtaining the data in which the candidate word is not used and the katakana notation is used from the data in the translated word appearance data storage unit, and storing the data in the translated word appearance data storage unit Extracting data related to the source language morpheme and katakana notation word from the data of all the translated translation files, confirming the frequency of the extracted data, and converting the katakana word into the translation dictionary database as a translation of the source language morpheme. And a program for executing a step of registering a notation word.

With this configuration, in an actual translation,
When a katakana notation of the source language morpheme is used as a translation of a certain source language morpheme, the katakana notation used in the actual translation is automatically used as a translation of the source language morpheme in the translation word dictionary database. You can register.

According to a twelfth aspect of the present invention, there is provided a step of taking out one source language text file, a step of performing sentence recognition in accordance with a written language, and referring to a translation word dictionary to store one source language text file. A morpheme analysis process for each sentence, storing a morpheme list in which morpheme candidates are listed in order of priority in each morpheme in a morpheme list storage unit, and a step of extracting a corresponding target language text file; Performing a sentence recognition, referring to the target language analysis word dictionary, performing a morphological analysis of the target language text one sentence at a time, and storing a morpheme list in a morpheme list storage unit; The translation and collation of the source language morpheme are checked for katakana notation in the target language's morpheme list. Storing the translated word appearance data in the translated word appearance data storage unit, obtaining the morpheme of the source language in which the translation candidate is not used and the katakana notation is used, from the data in the translated word appearance data storage unit, Extracting the data relating to the source language morpheme from the data of all the bilingual files stored in the translation unit; confirming the translation used in the extracted data; and translating the katakana as the translation of the source language morpheme in the translation dictionary database. And a program for executing a step of registering a notation word.

With this configuration, in an actual translation,
If the translation candidate of a certain source language morpheme is not used and the katakana notation of the reading of the source language morpheme is used, the katakana notation used in the actual translation is converted into the source language in the translation word dictionary database. It can be automatically registered as a morpheme translation.

According to a thirteenth aspect of the present invention, there is provided a step of taking out one source language text file, a step of performing sentence recognition in accordance with the written language, and referring to a translation word dictionary to store one source language text. A morpheme analysis process for each sentence, storing a morpheme list in which morpheme candidates are listed in order of priority in each morpheme in a morpheme list storage unit, and a step of extracting a corresponding target language text file; Performing a sentence recognition, referring to the target language analysis word dictionary, performing a morphological analysis of the target language text one sentence at a time, and storing a morpheme list in a morpheme list storage unit; The translation and collation of the source language morpheme are checked for katakana notation in the target language's morpheme list. Storing the translated word appearance data in the translated word appearance data storage unit, obtaining the morpheme of the source language in which the translation candidate is not used and the katakana notation is used, from the data in the translated word appearance data storage unit, Extracting data related to the source language morpheme from the data of all the bilingual files stored in the translation unit, confirming the translated words used in the extracted data, classifying the extracted data according to the translated words, and a translation dictionary. Registering the katakana notation as a translation of the source language morpheme in a database.

With this configuration, in an actual translation,
When a katakana notation of the reading of a source language morpheme is frequently used as a translation of a certain source language morpheme, the katakana notation used in the actual translation is replaced with the katakana notation of the source language morpheme in the word dictionary database for translation. It can be automatically registered as a translation.

According to a fourteenth aspect of the present invention, one source language text file is taken out, a sentence is recognized according to the written language, and one source language text is referenced by referring to a translation word dictionary. A morpheme analysis process for each sentence, storing a morpheme list in which morpheme candidates are listed in order of priority in each morpheme in a morpheme list storage unit, and a step of extracting a corresponding target language text file; Performing a sentence recognition, referring to the target language analysis word dictionary, performing a morphological analysis of the target language text one sentence at a time, and storing a morpheme list in a morpheme list storage unit; The translation and collation of the source language morpheme are checked for katakana notation in the target language's morpheme list. Storing the translated word appearance data in the translated word appearance data storage unit, obtaining the morpheme of the source language in which the translation candidate is not used and the katakana notation is used, from the data in the translated word appearance data storage unit, Extracting data related to the source language morpheme from the data of all the bilingual files stored in the translation unit, confirming the translated words used in the extracted data, classifying the extracted data according to the translated words, and a translation dictionary. Registering the katakana notation as the first priority translation of the source language morpheme in a database.

With this configuration, in the actual translation,
If a katakana spelling of a source language morpheme is frequently used as a translation of a source language morpheme, register the katakana spelling as the first priority translation of the source language morpheme in the translation dictionary database. In, the translation words used for the actual translation can be automatically registered in the dictionary data,
The dictionary data can be changed so that the katakana notation is preferentially used as a frequently used translation.

According to a fifteenth aspect of the present invention, there is provided a step of taking out one source language text file, a step of performing sentence recognition in accordance with the written language, and referring to a translation word dictionary to store one source language text file. A morpheme analysis process for each sentence, storing a morpheme list in which morpheme candidates are listed in order of priority in each morpheme in a morpheme list storage unit, and a step of extracting a corresponding target language text file; Performing a sentence recognition, referring to the target language analysis word dictionary, performing a morphological analysis of the target language text one sentence at a time, and storing a morpheme list in a morpheme list storage unit; The translation and collation of the source language morpheme are checked for katakana notation in the target language's morpheme list. Storing the translated word appearance data in the translated word appearance data storage unit, obtaining the morpheme of the source language in which the translation candidate is not used and the katakana notation is used, from the data in the translated word appearance data storage unit, Extracting data related to the source language morpheme from the data of all the bilingual files stored in the section, checking the translation words used in the extracted data, classifying the extracted data by field, and data for each field. An information recording medium storing a program for executing a step of confirming a translation word used in the translation dictionary database and a step of registering the katakana notation as a translation of the source language morpheme in data of a technical term dictionary in a specific field of the translation dictionary database It is.

According to this configuration, when a katakana notation for reading an English word is used as a translation of a source language morpheme in a specific field, the katakana notation used for actual translation is converted into a word dictionary for translation. It can be automatically registered as a translation of the source language morpheme in the technical term dictionary data of the database.

Hereinafter, embodiments of the present invention will be described with reference to the drawings. FIG. 1 is a functional block diagram of a dictionary data improving device according to an embodiment of the present invention, FIG. 2 is a circuit block diagram of the dictionary data improving device, FIG. 3 is a flowchart showing the operation of the data analysis process, and FIG. FIG. 5 is a diagram of the source language morpheme list, FIG. 5 is a diagram of the target language morpheme list, FIG. 6 is a flowchart showing the operation of the used translation word matching process in the data analysis process, FIG. 7 is an example diagram of the translated word appearance data, FIG. 9, 10, and 11 are flowcharts showing the operation of the dictionary data change processing.

In FIG. 1, 1 is an input means for designating a target document and giving an execution instruction. 2 is a sentence recognition processing unit which divides the text of a bilingual text file for each sentence. 3 is necessary for analyzing a target language. A target language analysis word dictionary storing information on various words, a translation word dictionary storing information on words necessary for translation from the source language to the target language, and a target language analysis word dictionary 3 A target language morphological analysis processing unit that refers to and divides a sentence described in the target language into morpheme units; and a source language that divides a sentence described in the source language into morpheme units by referring to the translation word dictionary 4 A morphological analysis processing unit 7 is a morphological list storage unit that stores morphological analysis results of a text file described in the source language and the target language, and a morphological list storage unit 8 is a target language word and an original Word Utilization translation unit 9 for checking whether katakana notation of morpheme reading is present in the morpheme list of the target language, and katakana notation data for acquiring katakana notation of reading of specified word from translation dictionary database 13 An acquisition unit, 10 is a translated word appearance data storage unit for storing translated word matching results, 11 is a translated word information improvement data extracting unit, which extracts dictionary registered words of the source language for adding or changing translated words,
Reference numeral 12 denotes a dictionary data changing unit that changes data relating to a specified word in the translation dictionary database 13. Reference numeral 13 denotes a translation dictionary database that forms the translation word dictionary 4.
4 is a display means for displaying data and the like, 15 is a sentence recognition processing unit 2, a target language morphological analysis processing unit 5, a source language morphological analysis processing unit 6, a used translation word collation unit 8, a katakana notation data acquisition unit 9, a translated word The control unit controls the functions of the information improvement data extraction unit 11 and the dictionary data change unit 12 and the operation between the functions, and the flow of data stored in the morpheme list storage unit 7 and the translated word appearance data storage unit 10.

FIG. 2 is a circuit block diagram of the dictionary data improving apparatus according to the present embodiment, wherein 16 is a keyboard, 17 is a central processing unit (hereinafter abbreviated as CPU), and 18 is a cathode ray tube display (hereinafter abbreviated as CRT). , 19 is a read only memory (hereinafter abbreviated as ROM), 20 is a random access memory (hereinafter abbreviated as RAM), 21
Is a disk drive for reading data from a recording medium 22 such as a CD-ROM.

The input means 1 shown in FIG.
Thus, the morpheme list storage unit 7, the translated word appearance data storage unit 10, and the translation dictionary database 13 are stored in the RAM 20, and the target language analysis word dictionary 3 and the translation word dictionary 4 are stored in R.
By the OM 19, the sentence recognition processing unit 2, the target language morphological analysis processing unit 5, the source language morphological analysis processing unit 6, the used translation word matching unit 8, the translated word information improvement data extraction unit 11, the dictionary data change unit 12, and the control unit 15 Is ROM 19 and R
The display means 14 is realized by the CRT 18 by executing a program stored in the program storage area of the ROM 19 while exchanging data with the AM 20.

Note that, in the present embodiment, the CPU 17
Although a form in which control is performed by executing a program stored in the OM 19 is shown, a control program recorded on a computer-readable recording medium 22 is read from a disk drive 21 and expanded on a RAM 20. The form executed by the CPU 17 may be used. By taking such a form, the present invention can be easily realized by a general-purpose computer.

(Embodiment 1) In the dictionary data improving apparatus configured as described above, when a bilingual file database to be used is specified by the input means 1 and the execution is instructed, data analysis performed first is performed. The operation of the process will be described below with reference to the flowchart of FIG.

At step A 1, the control unit 15 extracts one source language text file from the bilingual file database and passes it to the sentence recognition processing unit 2. In step A2, the sentence recognition processing unit 2 targets the specified text file
Sentence recognition is performed according to the written language. In step A3, the source language morphological analysis processing unit 6 performs a morphological analysis on the text divided in units of sentences in step A2, referring to the translation word dictionary 4, one sentence at a time. A morpheme list in which candidates are listed in order of priority is stored in the morpheme list storage unit 7. FIG. 4 shows an example of a morpheme list stored in the morpheme list storage unit 7.

Next, the corresponding target language text file is analyzed. In step A4, the control unit 15 extracts a target language text file corresponding to the source language text file extracted in step A1 from the bilingual file database, and passes it to the sentence recognition processing unit 2. Step A5
Then, the sentence recognition processing unit 2 performs sentence recognition on the specified text file according to the written language. In step A6, the target language morphological analysis processing unit 5 targets the text divided in sentence units in step A5.
With reference to the target language analysis word dictionary 3, morphological analysis is performed for each sentence, and a morpheme list listing each morpheme is stored in the morpheme list storage unit 7. FIG. 5 shows an example of the morpheme list stored in the morpheme list storage unit 7.

In step A7, the translation word of each morpheme of the source language morpheme list (FIG. 4) and the katakana notation of the reading stored in the morpheme list storage unit 7 by the use translation word collation unit 8 are converted into the morpheme list of the target language. The use translation word collation processing of whether or not it exists in FIG. 5) is performed, and the collation result is stored in the translation word appearance data storage unit 10. In step A8, if there is an unprocessed source language file in the bilingual file database, the process returns to step A1 to continue the process. If there is no unprocessed file, the data analysis process ends, and then the data change process is performed.

The more detailed operation of the used word matching process performed in step A7 of the data analysis process will be described below with reference to FIG.
A description will be given based on the flowchart of FIG.

At step A7-1, morpheme data for one word is obtained from the source language morpheme list stored in the morpheme list storage 7. In the case of the source language morpheme list of FIG. 4, "device noun" device tool trademark strategy
Measure Data of "design" is acquired.

In step A7-2, step A7-1
A high-priority translation is extracted from the morphological data extracted in step (1). In step A7-3, it is confirmed whether or not the translated word extracted in step A7-2 exists in the morpheme list of the target language. If it exists in the list, the subsequent same morpheme does not need to be processed.
Skips the same morphological data in step A7-5.
Then, the source language being processed is stored in the translated word appearance data storage unit 10 in association with the translated word present in the list in step A7-3, and the process returns to step A7-1 until the next morpheme data cannot be obtained. Continue.

If the acquired translation does not exist in the morpheme list, the process returns to step A7-2. If another translation candidate exists, the translation having the next highest priority is taken out and the processing is continued in the same manner. If no translation word candidate can be obtained in step A7-2, the translation word candidate of the source language morpheme does not exist in the corresponding target language text file, so that step A7-6.
Proceed to.

The "device" acquired in step A7-1
e Noun Device Tool Trademark Strategic Strategy Design
In the example of (1), first, in step A7-2, the translated word "device" is extracted, and in step A7-3, it is confirmed whether or not "device" exists in the target language morpheme list shown in FIG. Since "device" does not exist in the list of FIG. 5, step A7-
Return to Step 2 and take out the next translated word "tool". This process is repeated until the last translated word "pattern". Since all translated words are not present in the target language morpheme list, the process proceeds to step A7-6.

In step A7-6, the data is stored in the translated word appearance data storage unit 10 assuming that the translated language has not appeared in the source language being processed. In step A7-7, the next morpheme in the source language morpheme list is stored. Find out. If the morpheme currently being processed is different from the next morpheme, the process proceeds to step A7-8. If the morpheme is the same, the process returns to step A7-1 to process the next morpheme data.

At step A7-8, katakana word data of the reading of the source language word is obtained. Step A7-
In step 9, the katakana notation acquired in step A7-8 is
Check whether it exists in the morpheme list of the target language.
If the word exists in the list, the process proceeds to step A7-5, and the source language being processed and the katakana notation word are stored in the translated word appearance data storage unit 10 in association with each other.

To explain using the above example, in step A7-7, the morpheme “device” currently being processed and the next morpheme “enabl” in the source language morpheme list (FIG. 4)
Since e "is different, the process proceeds to step A7-8. In step A7-8, katakana notation words" devise "and" device "for reading" device "are acquired. In step A7-9, it is confirmed whether or not "divide" and "device" are present in the morpheme list of the target language. Since "device" is present in the list, the process proceeds to step A7-5, and the source being processed is processed. Data "device" in which the language "device" is associated with the katakana notation "device"
KANA KANA device "is stored in the translated word appearance data storage unit 10.

FIG. 7 shows translated word appearance data as a result of performing the above processing using the morphological analysis lists shown in FIGS.

In the specified bilingual file database,
After the analysis processing shown in FIG. 3 is completed for all text files described in the source language and the target language, the translated word information of the translation dictionary database 13 is translated using the translated word matching data in the translated word appearance data storage unit 10. Make changes.

The operation of the dictionary data improving apparatus of the present embodiment for changing the translated word information will be described below with reference to the flowchart of FIG. In step B1, the translated word information improved data extraction unit 11 acquires, from the data stored in the translated word appearance data storage unit 10, the morpheme of the source language in which the katakana notation is used and the katakana notation. In the data of FIG. 7, “device” and “device” are acquired.

At step B2, the translated word appearance data storage unit 1
0, the source language morpheme (“dev” in the above example) acquired in step B1 from the data of all the bilingual files stored in
(i.e., "device") in the same katakana word ("device" in the above example).

In step B3, if the number of data extracted in step B2 has not reached the predetermined number, the process returns to step B1 to acquire the next target data. Step B2
If the number of data extracted in step (b) has reached a certain number, in step B4, the dictionary data changing unit 12 adds the extracted katakana notation words to the morpheme information of the source language in the translation dictionary database 13 to register the translated word information. Make the change and return to step B1.

In the above example, the translation dictionary database 13
Is registered as a translation of "device".

In step B1, the next target data is obtained and the process is continued. When the target data no longer exists, the process ends.

(Embodiment 2) The operation of the dictionary data improving apparatus configured as described above will be described below with reference to the drawings. When the translation file database to be used is specified by the input unit 1 and the execution is instructed, the data analysis process performed first is the same as that of the first embodiment.
Is the same as

After the analysis process shown in FIG. 3 is completed, the process of changing the translated word information of the dictionary data improving apparatus of the present embodiment, which is performed using the translated word collation data in the translated word appearance data storage unit 10, will be described below. This will be described with reference to the flowchart of FIG.

In step C 1, the translated word information improvement data extraction unit 11 obtains a morpheme of the source language in which the katakana notation is used from the data stored in the translated word appearance data storage unit 10. In the data of FIG. 7, "device" is obtained.

In step C2, the translated word appearance data storage unit 1
0, the source language morpheme acquired in step C1 (“dev
ice ”) is extracted.

In step C3, if the number of data extracted in step C2 does not reach the fixed number set in accordance with the data amount, the data is insufficient, so that the dictionary data changing process for this morpheme is not performed, and the process proceeds to step C1. Return and obtain the next target morpheme. If the number of data extracted in step C2 has reached a certain number, it is checked in step C4 whether or not a candidate for a translation word is used in the data extracted in step C2. In this case, the process proceeds to Step C5. If any of the candidate words is used, the process returns to step C1 to acquire the next target morpheme.

In step C5, the dictionary data changing unit 12 changes the translated word information by registering the extracted katakana notation words in the morpheme information of the source language in the translation dictionary database 13, and returns to step C1. At this time, a change such as registering the katakana word as the first priority translation is also possible.

In the above example, “d” extracted in step C2
If all of the translations “device”, “tool”,..., “design” are not used in all the data related to “device”, “device” in the translation dictionary database 13
"device" is registered as a translation of e ". At this time, a change such as registering “device” as the first priority translation of “device” in the translation dictionary database 13 is also possible.

In step C1, the next target morpheme is obtained and the process is continued. When the target morpheme no longer exists, the process ends.

(Embodiment 3) The operation of the dictionary data improving apparatus configured as described above will be described below with reference to the drawings.

When the bilingual file database to be used is specified by the input means 1 and the execution is instructed,
The first data analysis process is the same as in the first embodiment.

After the analysis processing shown in FIG. 3 is completed, the operation of the translation word information change processing of the dictionary data improving apparatus of the present embodiment, which is performed using the translated word collation data in the translated word appearance data storage unit 10, will be described below. This will be described with reference to the flowchart of FIG.

Steps up to step D5 in the flowchart in FIG. 10 are the same as those up to step C5 in the flowchart in FIG. However, in step D4, step D2
If the morpheme of the candidate word is used in the data extracted in step, the process proceeds to step D6 without returning to step D1.

In step D6, the data extracted in step D2 is classified by use translation. In step D7, in the classification of step D6, if the ratio of the katakana notation word use data to all the data exceeds a certain value, the process proceeds to step D5.

At Step D5, the dictionary data changing unit 12
With this, the translated word information is changed by registering the extracted katakana notation word in the morpheme information of the source language in the translation dictionary database 13, and the process returns to step D1. At this time, a change such as registering the katakana word as the first priority translation is also possible.

Using the above example again, in the data relating to “device” extracted in step D 2, the katakana notation “device” is used for a certain amount or more (for example, 90%) of the data, and the remaining (for example, 10%) When the translation word “apparatus” is used in the data, “device” is registered as a translation of “device” in the translation dictionary database 13. At this time, “d” in the translation dictionary database 13
A change such as registering "device" as the first priority translation of "device" is also possible.

In step D1, the next morpheme of interest is obtained and the process is continued. When the morpheme of interest no longer exists, the process is terminated.

(Embodiment 4) The operation of the dictionary data improving apparatus configured as described above will be described below with reference to the drawings. When the translation file database to be used is specified by the input unit 1 and the execution is instructed, the data analysis process performed first is the same as that of the first embodiment.
Is the same as However, it is assumed that each bilingual file used in the present embodiment is given a field code based on the description.

After the analysis process shown in FIG. 3 is completed, the translation word priority changing process of the dictionary data improving apparatus of this embodiment, which is performed using the translated word collation data in the translated word appearance data storage unit 10, will be described below. This will be described with reference to the flowchart of FIG.

Steps up to step E5 in the flowchart in FIG. 11 are the same as those up to step C5 in the flowchart in FIG.

However, in step E4, when the morpheme of the candidate word is used in the data extracted in step E2, the process proceeds to step E6 without returning to step E1.

At step E6, the data extracted at step E2 is classified by field. In step E7, in the data for each field classified in step E6, a field in which the number of data reaches a certain number set according to the data amount is searched. If there is no field in which the number of data has reached the certain number set according to the data amount, the process returns to step E1 without performing dictionary data change processing due to lack of data.

At step E1, the next target morpheme is obtained, and the process is continued. When the target morpheme no longer exists, the process ends. If there is a field in which a certain number or more of data exists, in step E8, the data in the field is classified by use translation.

At step E9, in the classification of step E8, if the ratio of the katakana word use data to all the data exceeds a certain level, the process proceeds to step E10.
If not, the dictionary data change processing for this field is not performed, and the process returns to step E7 to continue data processing for another field.

In step E10, the dictionary data changing unit 1
2, the data of the technical term dictionary of the field of the translation dictionary database 13 is changed. If the translation dictionary database 13 contains data related to the morpheme of the source language as the data of the technical term dictionary of the field, the katakana notation is registered as a translation, and the data related to the morpheme of the source language does not exist. Then, data relating to the source language morpheme is added to the data of the technical term dictionary in the field, and the katakana notation is registered as the first priority translation. After the end of the data change process for the field, the process returns to step E7 to continue the data process for another field.

In the present embodiment, a plurality of data change methods have been described separately for different embodiments, but a plurality of data change methods can be used simultaneously.

[0086]

As described above, according to the present invention, a large amount of bilingual text analysis data is used to determine the actual appearance of katakana notation words in the source language morpheme in the corresponding target language morpheme list. Katakana notation words used in translation can be automatically registered as translations in the translation word dictionary database.

[Brief description of the drawings]

FIG. 1 is a functional block diagram of a dictionary data improving device according to an embodiment of the present invention.

FIG. 2 is a circuit block diagram of a dictionary data improving device according to an embodiment of the present invention.

FIG. 3 is a flowchart showing an operation of a data analysis process in the embodiment of the present invention.

FIG. 4 is an example diagram of a source language morpheme list according to the embodiment of the present invention.

FIG. 5 is an example diagram of a target language morpheme list according to the embodiment of the present invention.

FIG. 6 is a flowchart showing an operation of a used word matching process in the data analysis process according to the embodiment of the present invention

FIG. 7 is an example diagram of translated word appearance data according to the embodiment of the present invention.

FIG. 8 is a flowchart showing the operation of dictionary data change processing according to the embodiment of the present invention.

FIG. 9 is a flowchart illustrating an operation of a dictionary data change process according to the embodiment of the present invention.

FIG. 10 is a flowchart illustrating an operation of a dictionary data change process according to the embodiment of the present invention.

FIG. 11 is a flowchart illustrating an operation of a dictionary data change process according to the embodiment of the present invention.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 Input means 2 Sentence recognition processing part 3 Target language analysis word dictionary 4 Translation word dictionary 5 Target language morphological analysis processing part 6 Source language morphological analysis processing part 7 Morphological list storage part 8 Use translation word collation part 9 Katakana notation data acquisition part 10 translated word appearance data storage unit 11 translated word information improved data extraction unit 12 dictionary data changing unit 13 translation dictionary database 14 display means 15 control unit 16 keyboard 17 CPU 18 CRT 19 ROM 20 RAM 21 disk drive 22 storage medium

Claims (15)

[Claims] An input means for designating a target document and giving an execution instruction, a sentence recognition processing unit for dividing the text of a bilingual text file for each sentence, and a target language storing information on words necessary for analyzing the target language By referring to the analysis word dictionary, the translation word dictionary that stores information on words necessary for translation from the source language to the target language, and the target language analysis word dictionary, the sentence described in the target language is converted into morpheme units. Referring to the target language morphological analysis processing unit to be divided and the word dictionary for translation,
A source language morphological analysis unit that divides a sentence described in the source language into morpheme units, a morpheme list storage unit that stores morphological analysis results of text files written in the source language and the target language, and morphemes in the source language A target translation word matching unit that checks whether the katakana notation of the target language word and the source language morpheme that is a candidate for the translation exists in the morpheme list of the target language, and a katakana notation word of the specified word reading A katakana notation data acquisition unit to be acquired, a translated word appearance data storage unit to store the translated word matching result, a translated word information improvement data extraction unit to extract the source language dictionary registered words to be added or changed, and a translation dictionary database A dictionary data change unit that changes data related to a specified word, and a translation dictionary data that includes katakana notation data of the source language word reading that constitutes a translation word dictionary. A database, display means for displaying data, etc., statements certified processor, target language morphological analysis unit,
Each function of the source language morphological analysis processing unit, the translated word matching unit, the katakana notation data acquisition unit, the translated word information improvement data extraction unit, the dictionary data change unit, and the operation between the functions, the morphological list storage unit, the translated word appearance data storage unit And a control unit for controlling the flow of data stored in the translation word information improvement data extraction unit, the translation word appearance data storage unit, from the source word morpheme in which the katakana notation of the reading of English words is used ,
A dictionary data improving apparatus for extracting, from all data in a translated word appearance data storage unit, data relating to morphemes in which the number of used data of the katakana notation exceeds a certain number. 2. An input means for designating a target document and giving an execution instruction; a sentence recognition processing unit for dividing text of a bilingual text file for each sentence; and a target language storing information on words necessary for analysis of the target language. By referring to the analysis word dictionary, the translation word dictionary that stores information on words necessary for translation from the source language to the target language, and the target language analysis word dictionary, the sentence described in the target language is converted into morpheme units. Referring to the target language morphological analysis processing unit to be divided and the word dictionary for translation,
A source language morphological analysis unit that divides a sentence described in the source language into morpheme units, a morpheme list storage unit that stores morphological analysis results of text files written in the source language and the target language, and morphemes in the source language A target translation word matching unit that checks whether the katakana notation of the target language word and the source language morpheme that is a candidate for the translation exists in the morpheme list of the target language, and a katakana notation word of the specified word reading A katakana notation data acquisition unit to be acquired, a translated word appearance data storage unit to store the translated word matching result, a translated word information improvement data extraction unit to extract the source language dictionary registered words to be added or changed, and a translation dictionary database A dictionary data change unit that changes data related to a specified word, and a translation dictionary data that includes katakana notation data for reading a source language word that constitutes a translation word dictionary. A database, display means for displaying data, etc., statements certified processor, target language morphological analysis unit,
Each function of the source language morphological analysis processing unit, the translated word matching unit, the katakana notation data acquisition unit, the translated word information improvement data extraction unit, the dictionary data change unit, and the operation between the functions, the morphological list storage unit, the translated word appearance data storage unit And a control unit for controlling the flow of data stored in the translation word information improvement data extraction unit. The translation word appearance improvement data extraction unit relates, from the translated word appearance data storage unit, to the morpheme of the source language in which the katakana notation of the reading of the English word is used. A dictionary for acquiring all data and extracting katakana notation of reading an English word as a registered translation if all translation candidates are not used as translations of the source language morpheme in the acquired data. Data improvement device. 3. An input means for designating a target document and instructing execution, a sentence recognition processing section for dividing the text of a bilingual text file for each sentence, and a target language storing information on words necessary for analyzing the target language. By referring to the analysis word dictionary, the translation word dictionary that stores information on words necessary for translation from the source language to the target language, and the target language analysis word dictionary, the sentence described in the target language is converted into morpheme units. Referring to the target language morphological analysis processing unit to be divided and the word dictionary for translation,
A source language morphological analysis unit that divides a sentence described in the source language into morpheme units, a morpheme list storage unit that stores morphological analysis results of text files written in the source language and the target language, and morphemes in the source language A target translation word matching unit that checks whether the katakana notation of the target language word and the source language morpheme that is a candidate for the translation exists in the morpheme list of the target language, and a katakana notation word of the specified word reading A katakana notation data acquisition unit to be acquired, a translated word appearance data storage unit to store the translated word matching result, a translated word information improvement data extraction unit to extract the source language dictionary registered words to be added or changed, and a translation dictionary database A dictionary data change unit that changes data related to a specified word, and a translation dictionary data that includes katakana notation data of the source language word reading that constitutes a translation word dictionary. A database, display means for displaying data, etc., statements certified processor, target language morphological analysis unit,
Each function of the source language morphological analysis processing unit, the translated word matching unit, the katakana notation data acquisition unit, the translated word information improvement data extraction unit, the dictionary data change unit, and the operation between the functions, the morphological list storage unit, the translated word appearance data storage unit And a control unit for controlling the flow of data stored in the translation word information improvement data extraction unit. The translation word appearance improvement data extraction unit relates, from the translated word appearance data storage unit, to the morpheme of the source language in which the katakana notation of the reading of the English word is used. A dictionary data improving apparatus for acquiring all data, classifying the data for each translation, and extracting a katakana spelling of an English word reading as a registered translation with reference to frequency information. 4. An input means for designating a target document and instructing execution, a sentence recognition processing unit for dividing a text of a bilingual text file for each sentence, and a target language storing information on words necessary for analyzing the target language. By referring to the analysis word dictionary, the translation word dictionary that stores information on words necessary for translation from the source language to the target language, and the target language analysis word dictionary, the sentence described in the target language is converted into morpheme units. Referring to the target language morphological analysis processing unit to be divided and the word dictionary for translation,
A source language morphological analysis unit that divides a sentence described in the source language into morpheme units, a morpheme list storage unit that stores morphological analysis results of text files written in the source language and the target language, and morphemes in the source language A target translation word matching unit that checks whether the katakana notation of the target language word and the source language morpheme that is a candidate for the translation exists in the morpheme list of the target language, and a katakana notation word of the specified word reading A katakana notation data acquisition unit to be acquired, a translated word appearance data storage unit to store the translated word matching result, a translated word information improvement data extraction unit to extract the source language dictionary registered words to be added or changed, and a translation dictionary database A dictionary data change unit that changes data related to a specified word, and a translation dictionary data that includes katakana notation data of the source language word reading that constitutes a translation word dictionary. A database, display means for displaying data, etc., statements certified processor, target language morphological analysis unit,
Each function of the source language morphological analysis processing unit, the translated word matching unit, the katakana notation data acquisition unit, the translated word information improvement data extraction unit, the dictionary data change unit, and the operation between the functions, the morphological list storage unit, the translated word appearance data storage unit And a control unit for controlling the flow of data stored in the storage unit. In the data in which the katakana notation extracted by the translation information improvement data extraction unit is registered as a translation, the katakana notation is registered as a first priority translation. A dictionary data improvement device characterized by the following. 5. An input means for designating a target document and instructing execution, a sentence recognition processing section for dividing the text of a bilingual text file for each sentence, and a target language storing information on words necessary for analyzing the target language. By referring to the analysis word dictionary, the translation word dictionary that stores information on words necessary for translation from the source language to the target language, and the target language analysis word dictionary, the sentence described in the target language is converted into morpheme units. Referring to the target language morphological analysis processing unit to be divided and the word dictionary for translation,
A source language morphological analysis unit that divides a sentence described in the source language into morpheme units, a morpheme list storage unit that stores morphological analysis results of text files written in the source language and the target language, and morphemes in the source language A target translation word matching unit that checks whether the katakana notation of the target language word and the source language morpheme that is a candidate for the translation exists in the morpheme list of the target language, and a katakana notation word of the specified word reading A katakana notation data acquisition unit to be acquired, a translated word appearance data storage unit to store the translated word matching result, a translated word information improvement data extraction unit to extract the source language dictionary registered words to be added or changed, and a translation dictionary database A dictionary data change unit that changes data related to a specified word, and a translation dictionary data that includes katakana notation data of the source language word reading that constitutes a translation word dictionary. A database, display means for displaying data, etc., statements certified processor, target language morphological analysis unit,
Each function of the source language morphological analysis processing unit, the translated word matching unit, the katakana notation data acquisition unit, the translated word information improvement data extraction unit, the dictionary data change unit, and the operation between the functions, the morphological list storage unit, the translated word appearance data storage unit And a control unit for controlling the flow of data stored in the translation word information improvement data extraction unit. When all data is obtained, the data is classified by field, and katakana notation of reading an English word is used as a translation of the source language morpheme in data of a specific field,
A dictionary data improving apparatus for extracting the data as data for changing terminology terminology translation information in the field. 6. Using a large number of bilingual file databases, for each bilingual text file, morphologically analyze the source language text file, create a morpheme list listing morphemes and candidate translations, and morphologically analyze the target language text file. Then, a morpheme list listing the morphemes is created, and a translation of each morpheme in the source language morpheme list and a katakana notation of the reading of the source language morpheme are subjected to use translation matching to determine whether or not the target language morpheme list exists, When the translation candidate is not used and the katakana spelling of the source language is used, refer to the results of using the translated words in all bilingual text files, and the number of data of the katakana spelling exceeds a certain number. Registering the katakana notation as a translation of the source language morpheme in a translation dictionary database. Over data improved method. 7. Using a large number of bilingual file databases, for each bilingual text file, morphologically analyze the source language text file, create a morphological list listing morphemes and candidate translations, and morphologically analyze the target language text file. Then, a morpheme list listing the morphemes is created, and a translation of each morpheme in the source language morpheme list and a katakana notation of the reading of the source language morpheme are subjected to use translation matching to determine whether or not the target language morpheme list exists, For the source language morphemes in which the translation candidate is not used and the pronunciation katakana notation is used, the data of the corresponding source language morpheme is extracted from the translated word matching results of all bilingual text files. If not all translation candidate words are used, the translation word of the source language morpheme in the translation dictionary database is Dictionary data improved method and registers the Tacana notation language. 8. Using a large number of bilingual file databases, for each bilingual text file, morphologically analyze the source language text file, create a morpheme list listing morphemes and candidate translations, and morphologically analyze the target language text file. Then, a morpheme list listing the morphemes is created, and a translation of each morpheme in the source language morpheme list and a katakana notation of the reading of the source language morpheme are subjected to use translation matching to determine whether or not the target language morpheme list exists, For the source language morphemes in which the translation candidate is not used and the pronunciation katakana notation is used, the data of the corresponding source language morpheme is extracted from the result of the use translation matching of the all bilingual text files, and the data is used as the translation target. , And refer to the frequency information to extract the data that registers the katakana notation of the reading of English words as a translation, Dictionary data improved method and registers the katakana words as translation of raw language morphemes translation dictionary database. 9. Using a large number of bilingual file databases, for each bilingual text file, morphologically analyze the source language text file, create a morpheme list listing morphemes and candidate translations, and morphologically analyze the target language text file. Then, a morpheme list listing the morphemes is created, and a translation of each morpheme in the source language morpheme list and a katakana notation of the reading of the source language morpheme are subjected to use translation matching to determine whether or not the target language morpheme list exists, For the source language morphemes where the translation candidate is not used and the pronunciation katakana notation is used, extract the corresponding source language morpheme data from the translated word matching results of all bilingual text files, and When registering the katakana notation as a translation of the source language based on the usage status of katakana notation and candidate translations, Dictionary data improved method and registers as a first priority translation data on raw language book database. 10. Using a large number of bilingual file databases, for each bilingual text file, morphologically analyze the source language text file, create a morpheme list listing morphemes and candidate translations, and morphologically analyze the target language text file. Then, a morpheme list listing the morphemes is created, and a translation of each morpheme in the source language morpheme list and a katakana notation of the reading of the source language morpheme are subjected to use translation matching to determine whether or not the target language morpheme list exists, For the source language morphemes in which the translation candidate is not used and the pronunciation katakana notation is used, the data of the corresponding source language morphemes is extracted from the results of matching translations of all bilingual text files, and the data is classified by field. Classify and use katakana notation of reading English words as a translation of the source language morpheme in data of a specific field If so, the data of the terminology dictionary of the field of translation dictionary database, dictionary data improved method and registers the katakana words as translation of raw language morpheme. 11. A step of extracting one source language text file, a step of performing sentence recognition in accordance with the written language, and referring to a translation word dictionary to store one source language text.
A morpheme analysis process for each sentence, storing a morpheme list in which morpheme candidates are listed in order of priority in each morpheme in a morpheme list storage unit, and a step of extracting a corresponding target language text file; Performing a sentence recognition, referring to the target language analysis word dictionary, performing a morphological analysis of the target language text one sentence at a time, and storing a morpheme list in a morpheme list storage unit; A step of collating the translated word and the use of katakana notation of the source language morpheme in the morpheme list of the target language, and storing the collation result in a translated word appearance data storage unit; Obtaining data in which katakana notation is used without using candidate translations from the data; Extracting data related to the source language morpheme and katakana words from the data of all the bilingual files stored in the storage unit; checking the frequency of the extracted data; Registering the katakana notation word as a translation word. 12. A step of extracting one source language text file, a step of performing sentence recognition in accordance with the written language, and referring to a translation word dictionary to store one source language text.
A morpheme analysis process for each sentence, storing a morpheme list in which morpheme candidates are listed in order of priority in each morpheme in a morpheme list storage unit, and a step of extracting a corresponding target language text file; Performing a sentence recognition, referring to the target language analysis word dictionary, performing a morphological analysis of the target language text one sentence at a time, and storing a morpheme list in a morpheme list storage unit; A step of collating the translated word and the use of katakana notation of the source language morpheme in the morpheme list of the target language, and storing the collation result in a translated word appearance data storage unit; Acquiring from the data the source language morphemes in which katakana notation is used and no translation candidate is used; Extracting the data relating to the source language morpheme from the data of all the bilingual files stored in the data storage unit, confirming the translation used in the extracted data, and converting the source language morpheme into a translation dictionary database. And a step of registering the katakana word. 13. A step of extracting one source language text file, a step of performing sentence recognition in accordance with the written language, and referring to a translation word dictionary to store one source language text file.
A morpheme analysis process for each sentence, storing a morpheme list in which morpheme candidates are listed in order of priority in each morpheme in a morpheme list storage unit, and a step of extracting a corresponding target language text file; Performing a sentence recognition, referring to the target language analysis word dictionary, performing a morphological analysis of the target language text one sentence at a time, and storing a morpheme list in a morpheme list storage unit; A step of collating the translated word and the use of katakana notation of the source language morpheme in the morpheme list of the target language, and storing the collation result in a translated word appearance data storage unit; Acquiring from the data the source language morphemes in which katakana notation is used and no translation candidate is used; Extracting data on the source language morpheme from the data of all the bilingual files stored in the data storage unit, confirming the translation of the extracted data, classifying the extracted data by translation, and translating Registering the katakana notation as a translation of the source language morpheme in a dictionary database for use. 14. A step of taking out one source language text file, a step of performing sentence recognition in accordance with the written language, and referring to a translation word dictionary to store one source language text file.
A morpheme analysis process for each sentence, storing a morpheme list in which morpheme candidates are listed in order of priority in each morpheme in a morpheme list storage unit, and a step of extracting a corresponding target language text file; Performing a sentence recognition, referring to the target language analysis word dictionary, performing a morphological analysis of the target language text one sentence at a time, and storing a morpheme list in a morpheme list storage unit; A step of collating the translated word and the use of katakana notation of the source language morpheme in the morpheme list of the target language, and storing the collation result in a translated word appearance data storage unit; Acquiring from the data the source language morphemes in which katakana notation is used and no translation candidate is used; Extracting data on the source language morpheme from the data of all the bilingual files stored in the data storage unit, confirming the translation of the extracted data, classifying the extracted data by translation, and translating Registering the katakana notation as the first priority translation of the source language morpheme in the dictionary database for business use. 15. A step of extracting one source language text file, a step of performing sentence recognition in accordance with the written language, and referring to a translation word dictionary to store one source language text file.
A morpheme analysis process for each sentence, storing a morpheme list in which morpheme candidates are listed in order of priority in each morpheme in a morpheme list storage unit, and a step of extracting a corresponding target language text file; Performing a sentence recognition, referring to the target language analysis word dictionary, performing a morphological analysis of the target language text one sentence at a time, and storing a morpheme list in a morpheme list storage unit; A step of collating the translated word and the use of katakana notation of the source language morpheme in the morpheme list of the target language, and storing the collation result in a translated word appearance data storage unit; Acquiring from the data the source language morphemes in which katakana notation is used and no translation candidate is used; Extracting data relating to the source language morpheme from the data of all the bilingual files stored in the data storage unit, confirming the translation words used in the extracted data, classifying the extracted data by field, Storing a program for executing a step of confirming a translation word of the data of the above and a step of registering the katakana notation as a translation of the source language morpheme in the data of a technical term dictionary of a specific field of the translation dictionary database. An information recording medium characterized by the above-mentioned.