JP6138065B2 - Program, apparatus and method for outputting search keywords suitable for different language systems - Google Patents

Description

  The present invention relates to a technique for selecting parallel translation words between different language systems.

  Everyone often learns through media data such as images and sounds in order to learn a foreign language for their native language. For example, even in an educational setting, if it is difficult to express the meaning of a foreign language word in the native language, the learner may be made to understand the visual and auditory image by media data.

  Conventionally, there is a technique of a foreign language learning apparatus that gives a learner a problem in consideration of confusion between word pairs similar to each of the correct text / meaning / phoneme (see, for example, Patent Document 1). According to this technique, based on the rules of the correct text / meaning / phoneme, candidate words of the correct text / meaning / phonological interference word for the designated learning word are extracted. And according to these interference degrees, the integrated difficulty level is estimated and a problem series is created. The generated problem is presented to the user in combination with an image and pronunciation sound.

  In addition, there is a technique for searching for metadata associated with an image in advance using an input text query and presenting the image to a searcher (see, for example, Non-Patent Document 1). According to this technique, an image search keyword for acquiring an appropriate image is generated from a given foreign language word and a native language obtained by translating the foreign language into the native language.

Japanese Patent No. 4432079

Google Image Search, [online], [Search January 22, 2014], Internet <URL: http: //images.google.com/> Pre-ranking processing using a vector space model, [online], [Search January 22, 2014], Internet <URL: http://www.jpo.go.jp/shiryou/s_sonota/hyoujun_gijutsu/search_engine/ b / b51.htm>

  However, according to the technique described in Patent Document 1, it is necessary to store in advance an image corresponding to the meaning of a foreign language word. Further, according to the technique described in Non-Patent Document 1, when an adverb, adjective, or a word having ambiguity is used as a search keyword, an image that does not match the meaning of a foreign language word or a plurality of ambiguity images are presented. There is also.

  Here, even if the inventors of the present application search for media data such as an image using a keyword obtained by translating a foreign language into its native language, the media data does not match the original meaning of the foreign language. Is there a case? I thought. At that time, considering the relationship between words that co-occur with a foreign language and words that co-occur with a translated native language, is it possible to extract a search keyword for an appropriate native language? I thought.

  Accordingly, an object of the present invention is to provide a program, an apparatus, and a method for outputting an appropriate search keyword for different language systems. This search keyword can be used to search for media data that should be understood.

According to the present invention, the first original language of the first language system and the second original language of the second language system, which is a parallel translation of the first original language, are input, and suitable as a usage for the first original language A program that causes a computer to function to output a second co-occurrence word of a second language system,
First co-occurrence word extracting means for extracting a plurality of first co-occurrence words co-occurring with the first original word using a corpus dictionary of the first language system;
First bilingual co-occurrence word extraction means for extracting a second bilingual co-occurrence word that becomes a bilingual translation of the first co-occurrence word using a bilingual dictionary of the first language system;
Using a corpus dictionary of a second language system, first appearance frequency search means for searching for an appearance frequency for each word string combining the second original word and the second parallel co-occurrence word;
A computer is caused to function as a co-occurrence word determining means for determining a second parallel co-occurrence word in a word string having the highest appearance frequency as a second co-occurrence word suitable as a usage for the first original word. .

According to another embodiment of the program of the present invention,
The first language system is a foreign language, the second language system is a native language,
As a usage of the first original language of the foreign language, it is also preferable to cause the computer to function so that a word string composed of the second original language of the native language and the second co-occurrence word is output.

According to another embodiment of the program of the present invention,
The first language system is a native language, the second language system is a foreign language,
As a usage of the first original language of the native language, it is also preferable to cause the computer to function so that a word string composed of the second original language of the foreign language and the second co-occurrence word is output.

According to another embodiment of the program of the present invention,
Using the word sequence composed of the second original word and the second co-occurrence word determined by the co-occurrence word determination means as a key, further causing the computer to function as search means for transmitting the search result. Is also preferable.

According to another embodiment of the program of the present invention,
It is also preferred to have the computer function so that the media server is a text, image or audio media server.

According to another embodiment of the program of the present invention,
Among the plurality of first co-occurrence words extracted from the first co-occurrence word extracting means, the co-occurrence word having the same and similar case as the first original word is deleted, and the other first co-occurrence words It is also preferable to further cause the computer to function as first co-occurrence word filter means for outputting to the first parallel co-occurrence word extraction means.

According to another embodiment of the program of the present invention,
Second co-occurrence word extraction means for extracting a plurality of second co-occurrence words co-occurring with the second original word using a corpus dictionary of the second language system;
Second bilingual co-occurrence word extraction means for extracting a first bilingual co-occurrence word that is a bilingual translation of the second co-occurrence word using a bilingual dictionary of the second language system;
Using the corpus dictionary of the first language system, the computer further functions as second appearance frequency search means for searching for the appearance frequency for each word string combining the first original word and the first parallel co-occurrence word. ,
The co-occurrence word determining means determines the second parallel co-occurrence word in the word string having the highest appearance frequency as a second co-occurrence word suitable as a usage for the first original word, and the word string having the highest appearance frequency. The first parallel co-occurrence word in is determined as the first co-occurrence word suitable as a usage for the second original word, and the one with the higher occurrence frequency in the first co-occurrence word and the second co-occurrence word is determined. It is also preferable to cause the computer to function so as to determine one of a word string composed of the first original word and the first co-occurrence word, or a word string composed of the second original word and the second co-occurrence word.

According to another embodiment of the program of the present invention,
Among the plurality of second co-occurrence words extracted from the second co-occurrence word extracting means, the co-occurrence words having the same and similar case as the second original word are deleted, and the other second co-occurrence words It is also preferable to further cause the computer to function as second co-occurrence word filtering means for outputting to the second bilingual co-occurrence word extracting means.

According to another embodiment of the program of the present invention,
The first co-occurrence word output from the first co-occurrence word filter means and the first parallel translation co-occurrence word output from the second appearance frequency search means are compared, and the matched first parallel translation co-words are compared. Second parallel co-occurrence word comparison means for outputting only the word to the co-occurrence word determination means;
The second co-occurrence word output from the second co-occurrence word filter means is compared with the second bi-translation co-occurrence word output from the first appearance frequency search means, and the matched second parallel co-occurrence words are compared. It is also preferable that the computer further function as first parallel co-occurrence word comparison means for outputting only the word to the co-occurrence word determination means.

According to the present invention, the first original language of the first language system and the second original language of the second language system, which is a parallel translation of the first original language, are input, and suitable as a usage for the first original language A device for outputting a second co-occurrence word of a second language system,
First co-occurrence word extracting means for extracting a plurality of first co-occurrence words co-occurring with the first original word using a corpus dictionary of the first language system;
First bilingual co-occurrence word extraction means for extracting a second bilingual co-occurrence word that becomes a bilingual translation of the first co-occurrence word using a bilingual dictionary of the first language system;
Using a corpus dictionary of a second language system, first appearance frequency search means for searching for an appearance frequency for each word string combining the second original word and the second parallel co-occurrence word;
Co-occurrence word determining means for determining the second parallel co-occurrence word in the word string having the highest appearance frequency as a suitable second co-occurrence word as a usage for the first original word.

According to the present invention, the first original language of the first language system and the second original language of the second language system that is a parallel translation of the first original language are input using the device, A method of outputting a second co-occurrence word of a second language system suitable as a usage for an original language,
A first step of extracting a plurality of first co-occurrence words co-occurring with a first original word using a corpus dictionary of a first language system;
A second step of extracting a second parallel co-occurrence word that is a parallel translation of the first co-occurrence word using the parallel translation dictionary of the first language system;
A third step of searching for the appearance frequency for each word string combining the second original word and the second parallel co-occurrence word using a corpus dictionary of the second language system;
And a fourth step of determining a second co-occurrence word in the word string having the highest appearance frequency as a second co-occurrence word suitable as a usage for the first original word.

  According to the program, apparatus and method of the present invention, it is possible to output search keywords suitable for different language systems. This search keyword can be used to search for media data that should be understood.

It is a system block diagram containing the bilingual keyword extraction apparatus in this invention. It is a functional block diagram of the program of this invention which extracts Japanese appropriate for the usage with respect to a foreign language. It is a functional block diagram of the program of this invention which extracts the foreign language suitable as usage to Japanese. It is a functional block diagram which included the co-occurrence word filter part in the structure of FIG. FIG. 4 is a functional configuration diagram in which the configuration of FIG. 2 and the configuration of FIG. 3 are configured in parallel. FIG. 6 is a functional configuration diagram comparing co-occurrence words between different language systems in the configuration of FIG. 5.

  Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings.

  FIG. 1 is a system configuration diagram including a bilingual keyword extracting apparatus according to the present invention.

  According to FIG. 1, the bilingual keyword extracting apparatus is connected to a network such as the Internet. In addition, a media server capable of searching for images and sounds is connected to the Internet. The media server stores a large amount of media data (image, video, audio) in advance, receives a query keyword, searches for media data appropriate for the keyword, and returns it. The media server associates multilingual text as metadata with media data.

  According to FIG. 1, it is assumed that the user inputs the foreign language “much” to be learned to the user terminal. The user terminal transmits the foreign language “much” to the bilingual keyword extraction device. The bilingual keyword extraction device translates the foreign language “much” into the native language “many”. The user terminal may translate the foreign language “much” into the native language “many” and transmit both “much” and “many” to the bilingual keyword extracting apparatus.

  On the other hand, it is assumed that the bilingual keyword extraction apparatus selects a search keyword “many moneys” in an appropriate native language in order for the learner to understand the foreign language “much”. Then, the bilingual keyword extracting device searches the media server using the native language “a lot of money” as a query. The media server returns media data (image, video, or sound) associated with metadata close to “a lot of money” to the bilingual keyword extraction device.

  Then, the bilingual keyword extraction device returns the media data received from the media server to the user terminal. By playing the media data on the user terminal, the user can visually understand the meaning of the input foreign language “much” from the media data related to the mother tongue “much money”. .

  Here, the bilingual keyword extraction device may, conversely, receive the native language to be learned from the user terminal and return the foreign language media data related to the native language to the user terminal. . That is, the bilingual keyword extracting apparatus according to the present invention applies the first original language from the first original language of the first language system and the second original language of the second language system that is the parallel translation of the first original language. The second co-occurrence word of the second language system suitable for usage is output.

  FIG. 2 is a functional configuration diagram of the program of the present invention for extracting Japanese appropriate for usage in a foreign language.

  FIG. 2 shows the configuration of a program that causes a computer installed in a bilingual keyword extraction apparatus (server) to function. Here, the first language system is a foreign language, and the second language system is Japanese. In addition, as a usage for the first original language of the foreign language, a word string composed of the second Japanese original language and the second co-occurrence language is output.

  Referring to FIG. 2, the bilingual keyword extraction device includes a foreign language corpus dictionary 111, a foreign language bilingual dictionary 112, and a Japanese corpus dictionary 121. A corpus is a database in which sentences in natural language are structured and accumulated on a large scale, and a part of speech or a co-occurrence word is associated with each word. The “foreign language corpus dictionary 111” is one in which a co-occurrence word having a high frequency of use is associated with each foreign language word. Similarly, the “Japanese corpus dictionary 121” is one in which a co-occurrence word having a high use frequency is associated with each Japanese word. The “foreign language parallel translation dictionary 112” is a dictionary in which Japanese words that are parallel translations of foreign language words are associated with each other.

  Further, according to FIG. 2, the bilingual keyword extraction device includes a bilingual word input unit 2, a first co-occurrence word extraction unit 31, a first bilingual co-occurrence word extraction unit 32, and a first appearance frequency search. The unit 33, the co-occurrence word determination unit 5, and the search unit 6 are included. These functional components are realized by executing a program that causes a computer installed in the apparatus to function.

[Parallel translation word input part 2]
The bilingual keyword extraction device inputs a foreign language original language and a Japanese original language to be translated. The bilingual Japanese source language may be translated by a bilingual keyword extraction device or received from a user terminal. According to FIG. 2, for example, the following words are input.
(Foreign language original) much <-> (Japanese original language) Many foreign language original words “much” are output to the first co-occurrence word extraction unit 31. Also, the Japanese original word “many” is output to the first appearance frequency search unit 33.

[First co-occurrence word extraction unit 31]
The first co-occurrence word extraction unit 31 uses a foreign language corpus dictionary 111 (a corpus dictionary of a first language system) to generate a plurality of first co-occurrence words that co-occur with a foreign language original word “much” (first original language). Extract one co-occurrence word. For example, the following words are extracted from the original language “much” of the foreign language.
(Foreign language) much-> (first co-occurrence)
better
pretty
money
success
music
The extracted co-occurrence words are output to the first parallel co-occurrence word extraction unit 32.

  In the extraction of co-occurrence words, one or more co-occurrence words for the original word are extracted. The corpus may be associated with the appearance frequency (co-occurrence frequency) of the original word and the co-occurrence word. Examples of co-occurrence words include synonyms, antonyms, idiomatic phrases, idioms, usages, and the like having the same meaning.

[First parallel co-occurrence word extraction unit 32]
The first bilingual co-occurrence word extraction unit 32 uses the foreign language bilingual dictionary 112 (the bilingual dictionary of the first language system) to extract a second bilingual co-occurrence word that is a bilingual translation of the first co-occurring word. To do. For example, assume that the following words are extracted.
(Foreign language first co-occurrence) (Japanese second parallel co-occurrence)
better-> better
pretty-> cute
money-> money
success-> success
music-> music The extracted second parallel co-occurrence word is output to the first appearance frequency search unit 33.

[First appearance frequency search unit 33]
The first appearance frequency search unit 33 uses the Japanese corpus dictionary 121 (the corpus dictionary of the second language system) to combine the second Japanese original word “many” and the second parallel co-occurrence word. Search for appearance frequency for each word string. For example, appearance frequencies for the following word strings are extracted.
(Japanese original words and parallel co-occurrence words) (Appearance frequency)
Lots of money-> 1500
Lots of music-> 780
Lots of success-> 550
Many good-> 30
Many cute-> 20
The frequency of appearance is, for example, the frequency of appearance from the direction of “many” to “money” and the frequency of occurrence from “money” to the direction of “many”, and combining them to generate the frequency of occurrence between words. It may be.

[Co-occurrence word determination unit 5]
The co-occurrence word determining unit 5 determines the second parallel co-occurrence word in the word string “many moneys” having the highest appearance frequency as a second co-occurrence word suitable as a usage for the first original word. The determined “a lot of money” is output to the search unit 6.

  As another embodiment, the first appearance frequency search unit 33 extracts a characteristic word string based on TF-IDF (Term Frequency-Inverse Document Frequency) instead of a simple appearance frequency. (For example, refer nonpatent literature 2). TF-IDF is a technology that weights each word, expresses sentences from a query in a vector space, and ranks them based on the similarity between the sentences and the query.

[Search unit 6]
The search unit 6 transmits, to the media server, the word string “much money” made up of the second original word and the second co-occurrence word determined by the co-occurrence word determination unit 5 as a key. On the other hand, the media server returns media data associated with metadata close to “a lot of money” to the search unit 6 of the bilingual keyword extraction device. Then, the media data is returned as a search result to the user terminal.

  The user terminal reproduces the media data returned from the bilingual keyword extraction device. Users can understand the meaning of the foreign language “much” by viewing metadata related to Japanese “a lot of money”.

  FIG. 3 is a functional configuration diagram of a program according to the present invention for extracting a foreign language suitable as a usage for Japanese.

  According to FIG. 3, the first language system is Japanese and the second language system is a foreign language. As a usage for the first Japanese original language, a word string composed of the second foreign language foreign word and the second co-occurrence word is output. That is, as compared with FIG. 2, the language system to be searched is reversed.

[Parallel translation word input part 2]
The bilingual keyword extracting device inputs the following words, for example.
(Original Japanese) <-> (Foreign language) much
The original Japanese word “many” is output to the second co-occurrence word extraction unit 41. Further, the foreign language original word “much” is output to the second appearance frequency search unit 43.

[Second co-occurrence word extraction unit 41]
The second co-occurrence word extraction unit 41 uses a Japanese corpus dictionary 121 (a corpus dictionary of the second language system) to generate a plurality of second words that co-occur with the Japanese original word “many” (second original word). Two co-occurrence words are extracted. For example, the following words are extracted from the Japanese original word “many”.
(Original Japanese) Lots-> (Second co-occurrence)
Many
money
life
musics
The extracted co-occurrence words are output to the second bilingual co-occurrence word extraction unit 42.

[Second parallel co-occurrence word extraction unit 42]
The second bilingual co-occurrence word extraction unit 42 uses the Japanese bilingual dictionary 122 (the bilingual dictionary of the second language system) to extract the first bilingual co-occurrence word that is the bilingual translation of the second co-occurrence word. To do. For example, assume that the following words are extracted.
(Second co-occurrence word in Japanese) (First parallel co-occurrence word in foreign language)
Many-> many
Money-> money
Life-> live
Music-> music
Smile-> smile
The extracted first parallel translation co-occurrence word is output to the second appearance frequency search unit 43.

[Second appearance frequency search unit 43]
The second appearance frequency search unit 43 uses the foreign language corpus dictionary 111 (corpus dictionary of the first language system) to combine the first original word “much” of the foreign language and the first parallel co-occurrence word. Search for appearance frequency for each word string. For example, appearance frequencies for the following word strings are extracted.
(Foreign language and parallel co-occurrence words) (Appearance frequency)
much smile-> 2500
much money-> 1200
much music-> 800
much live-> 340
much many-> 1

[Co-occurrence word determination unit 5]
The co-occurrence word determination unit 5 determines the second parallel co-occurrence word in the word string “much smile” having the highest appearance frequency as a second co-occurrence word suitable as a usage for the first original word. The determined “much smile” is output to the search unit 6.

[Search unit 6]
The search unit 6 transmits the word string “much smile” composed of the second original word and the second co-occurrence word determined by the co-occurrence word determination unit 5 to the media server as a key. In response to this, the media server returns the media data associated with metadata close to “much smile” to the search unit 6 of the bilingual keyword extraction device. Then, the media data is returned as a search result to the user terminal.

  The user terminal reproduces the media data returned from the bilingual keyword extraction device. Users can understand the meaning of Japanese “many” by viewing metadata related to the foreign language “much smile”.

  FIG. 4 is a functional configuration diagram including a co-occurrence word filter unit in the configuration of FIG.

  According to FIG. 4, a first co-occurrence word filter unit 34 is provided between the first co-occurrence word extraction unit 31 and the first parallel translation co-occurrence word extraction unit 32.

[First Co-occurrence Word Filter Unit 34]
The first co-occurrence word filter unit 34 has the same case as the first original word “much” among the plurality of first co-occurrence words extracted from the first co-occurrence word extraction unit 31. The co-occurrence words are deleted, and the other first co-occurrence words are output to the first parallel co-occurrence word extraction unit 32. For example, co-occurrence words are deleted as follows.
(Foreign language) much-> (first co-occurrence)
× better (same adverb case as much)
× pretty (adverbic case similar to much)
money
success
music
The first co-occurrence word filter unit 34 preferably excludes definite articles such as a and then and prepositions such as at and from.

  Although not shown in FIG. 4, it is also preferable to provide a second co-occurrence word filter unit 44 between the second co-occurrence word extraction unit 41 and the second parallel translation co-occurrence word extraction unit 42. .

[Second Co-occurrence Word Filter Unit 44]
The second co-occurrence word filter unit 44 is a co-occurrence word having the same similarity as the second original word among the plurality of second co-occurrence words extracted from the second co-occurrence word extraction unit 41. And the other second co-occurrence words are output to the second parallel co-occurrence word extraction unit 42. For example, co-occurrence words are deleted as follows.
(Original Japanese) Lots-> (Second co-occurrence)
× Many (adverbial case similar to “many”)
money
life
musics
Smile

  FIG. 5 is a functional configuration diagram in which the configuration of FIG. 2 and the configuration of FIG. 3 are configured in parallel.

According to FIG. 5, the configuration of FIG. 2 and the configuration of FIG. 3 are integrated by the co-occurrence word determination unit 5. The co-occurrence word determination unit 5 determines the co-occurrence word in the following steps.
(S1) The second bilingual co-occurrence word “money” in the word sequence having the highest appearance frequency is determined as a second co-occurrence word suitable as a usage for the first original word “many”. That is, “a lot of money” is determined.
(S2) The first parallel co-occurrence word “smile” in the word string having the highest appearance frequency is determined as the first co-occurrence word suitable as a usage for the second original word. That is, “much smile” is determined.
(S3) A word string composed of the first original word and the first co-occurrence word, or the second original word and the second one having the higher appearance frequency in the first co-occurrence word and the second co-occurrence word. One of the word strings consisting of the co-occurrence words is determined. Here, “much smile” having a high appearance frequency is determined.

  FIG. 6 is a functional configuration diagram in which co-occurrence words are compared between different language systems in the configuration of FIG.

[First parallel co-occurrence word comparison unit 35]
The first bilingual co-occurrence word comparison unit 35 outputs the second co-occurrence word output from the second co-occurrence word filter unit 44 and the second bilingual co-occurrence word output from the first appearance frequency search unit 33. The word is compared, and only the matched second parallel co-occurrence word is output to the co-occurrence word determination unit 5. According to FIG. 6, the comparison can be made as follows.
(Output of first appearance frequency search unit 33) (Output of second co-occurrence word filter unit 44)
○ A lot of money Money
× Many successes
○ Lots of music Music
Smile By this, the first parallel co-occurrence word comparison unit 35 outputs “much money” and “much music” to the co-occurrence word determination unit 5.

[Second parallel co-occurrence word comparison unit 45]
The second parallel co-occurrence word comparison unit 45 outputs the first co-occurrence word output from the first co-occurrence word filter unit 34 and the first parallel translation co-occurrence word output from the second appearance frequency search unit 43. The words are compared with each other, and only the matched first parallel co-occurrence word is output to the co-occurrence word determining unit 5.
(Output of Second Appearance Frequency Search Unit 43) (Output of Second Co-occurrence Word Filter Unit 44)
× much smile success
○ much money money
○ much music music
× much live
As a result, the second parallel co-occurrence word comparison unit 45 outputs “much money” and “much music” to the co-occurrence word determination unit 5.

According to FIG. 6, the co-occurrence word determination unit 5 determines a word string having the highest appearance frequency.
Lots of money-> 1500
Lots of music-> 780
much money-> 1200
much music-> 800
Finally, the co-occurrence word determination unit 5 determines “a lot of money” as a search keyword and outputs it to the search unit 6.

  The search unit 6 transmits the word string “a lot of money” as a key to the media server. On the other hand, the media server returns media data associated with metadata close to “a lot of money” to the search unit 6 of the bilingual keyword extraction device. Then, the media data is returned as a search result to the user terminal.

  The user terminal reproduces the media data returned from the bilingual keyword extraction device. Users can understand the meaning of the foreign language “much” by viewing metadata related to Japanese “a lot of money”.

  As described above in detail, according to the program, apparatus and method of the present invention, it is possible to output search keywords suitable for different language systems. This search keyword can be used to search for media data that should be understood.

  In particular, according to the present invention, for example, for a foreign language word learning apparatus, it is not necessary to previously store learning materials of media data (images, videos, or sounds) to be used. In each case, appropriate media data can be acquired by searching a vast number of external media servers. In addition, according to the present invention, media data can be searched between a foreign language and a native language using keywords that are frequently used in the native language as much as possible. Therefore, even foreign languages that cannot be understood by simply translating the foreign language into the native language are automatically supplemented with keywords that are frequently used as the native language, and media data appropriate for the user can be viewed. Become.

  Various changes, modifications, and omissions of the above-described various embodiments of the present invention can be easily made by those skilled in the art. The above description is merely an example, and is not intended to be restrictive. The invention is limited only as defined in the following claims and the equivalents thereto.

DESCRIPTION OF SYMBOLS 1 Parallel keyword extraction apparatus 111 Foreign language corpus dictionary 112 Foreign language parallel dictionary 121 Japanese corpus dictionary 122 Japanese parallel dictionary 2 Parallel word input part 31 1st co-occurrence word extraction part 32 1st parallel translation word extraction part 33 1st appearance frequency search part 34 1st co-occurrence word filter part 35 1st parallel translation co-occurrence word comparison part 41 2nd co-occurrence word extraction part 42 2nd parallel translation co-occurrence word extraction part 43 2nd appearance Frequency search unit 44 Second co-occurrence word filter unit 45 Second parallel co-occurrence word comparison unit 5 Co-occurrence word determination unit 6 Search unit

Claims (11)

The first language of the first language system and the second language of the second language system, which is a parallel translation of the first language, are input, and the second language system suitable as a usage for the first language A program that causes a computer to function to output the second co-occurrence word of
First co-occurrence word extracting means for extracting a plurality of first co-occurrence words co-occurring with the first original word using a corpus dictionary of the first language system;
First bilingual co-occurrence word extraction means for extracting a second bilingual co-occurrence word that becomes a bilingual translation of the first co-occurrence word using a bilingual dictionary of the first language system;
Using a corpus dictionary of a second language system, first appearance frequency search means for searching for an appearance frequency for each word string combining the second original word and the second parallel co-occurrence word;
A computer is caused to function as a co-occurrence word determining means for determining a second parallel co-occurrence word in a word string having the highest appearance frequency as a second co-occurrence word suitable as a usage for the first original word. program. The first language system is a foreign language, the second language system is a native language,
The computer is caused to function as a usage of the first original language of the foreign language so that a word string composed of the second original language and the second co-occurrence language of the native language is output. The program described in. The first language system is a native language, the second language system is a foreign language,
2. The computer according to claim 1, wherein the computer functions so as to output a word string composed of the second original word and the second co-occurrence word of the foreign language as a usage for the first original word of the native language. The program described in. The computer is further caused to function as search means for transmitting to the media server using the word string consisting of the second original word and the second co-occurrence word determined by the co-occurrence word determination means as a key and outputting the search result. The program according to any one of claims 1 to 3, wherein: The program according to any one of claims 1 to 4, wherein the media server causes a computer to function as a text, image, or audio media server. Among the plurality of first co-occurrence words extracted from the first co-occurrence word extracting means, the co-occurrence word having the same and similar case as the first original word is deleted, and the other first co-occurrence words 6. The program according to claim 1, wherein the computer is further caused to function as first co-occurrence word filtering means for outputting to the first parallel co-occurrence word extracting means. Second co-occurrence word extraction means for extracting a plurality of second co-occurrence words co-occurring with the second original word using a corpus dictionary of the second language system;
Second bilingual co-occurrence word extraction means for extracting a first bilingual co-occurrence word that is a bilingual translation of the second co-occurrence word using a bilingual dictionary of the second language system;
Using the corpus dictionary of the first language system, the computer further functions as second appearance frequency search means for searching for the appearance frequency for each word string combining the first original word and the first parallel co-occurrence word. ,
The co-occurrence word determining means determines the second parallel co-occurrence word in the word string having the highest appearance frequency as a second co-occurrence word suitable as a usage for the first original word, and the word having the highest appearance frequency. The first parallel co-occurrence word in the column is determined as the first co-occurrence word suitable as a usage for the second original word, and the one with the highest appearance frequency in the first co-occurrence word and the second co-occurrence word is determined. The computer is made to function so as to determine one of a word string composed of the first original word and the first co-occurrence word, or a word string composed of the second original word and the second co-occurrence word. The program according to any one of claims 1 to 6. Among the plurality of second co-occurrence words extracted from the second co-occurrence word extracting means, the co-occurrence words having the same and similar case as the second original word are deleted, and the other second co-occurrence words The program according to claim 7, further causing the computer to function as second co-occurrence word filtering means for outputting to the second parallel co-occurrence word extraction means. The first co-occurrence word output from the first co-occurrence word filter means and the first parallel translation co-occurrence word output from the second appearance frequency search means are compared, and the matched first parallel translation co-words are compared. Second parallel co-occurrence word comparison means for outputting only the word to the co-occurrence word determination means;
The second co-occurrence word output from the second co-occurrence word filter means is compared with the second bi-translation co-occurrence word output from the first appearance frequency search means, and the matched second parallel co-occurrence words are compared. 9. The program according to claim 8, further causing a computer to function as first parallel co-occurrence word comparing means for outputting only the word to the co-occurrence word determining means. The first language of the first language system and the second language of the second language system, which is a parallel translation of the first language, are input, and the second language system suitable as a usage for the first language A device for outputting the second co-occurrence word of
First co-occurrence word extracting means for extracting a plurality of first co-occurrence words co-occurring with the first original word using a corpus dictionary of the first language system;
First bilingual co-occurrence word extraction means for extracting a second bilingual co-occurrence word that becomes a bilingual translation of the first co-occurrence word using a bilingual dictionary of the first language system;
Using a corpus dictionary of a second language system, first appearance frequency search means for searching for an appearance frequency for each word string combining the second original word and the second parallel co-occurrence word;
And a co-occurrence word determining means for determining a second parallel co-occurrence word in a word string having the highest appearance frequency as a second co-occurrence word suitable as a usage for the first original word. Using the apparatus, the first original language of the first language system and the second original language of the second language system, which is a parallel translation of the first original language, are input and suitable as a usage for the first original language A method for outputting a second co-occurrence word of a second language system,
A first step of extracting a plurality of first co-occurrence words co-occurring with a first original word using a corpus dictionary of a first language system;
A second step of extracting a second parallel co-occurrence word that is a parallel translation of the first co-occurrence word using the parallel translation dictionary of the first language system;
A third step of searching for the appearance frequency for each word string combining the second original word and the second parallel co-occurrence word using a corpus dictionary of the second language system;
And a fourth step of determining the second parallel co-occurrence word in the word sequence having the highest appearance frequency as a second co-occurrence word suitable as a usage for the first original word.

Images