JP2008146218A - Language analysis system, language analysis method and computer program - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To generate a morpheme-analyzing dictionary to perform accurate morpheme analysis by accurately dividing a technical term difficult in division point setting to extract a morpheme. <P>SOLUTION: A parallel translation pair having registration data wherein the parallel translation pair of Japanese and a foreign language is one word-to-multiple words different from one word-to-one word is extracted from registration data of a bilingual dictionary of Japanese and the foreign language, Japanese of the extracted parallel translation pair is performed with the morpheme analysis, a foreign word corresponding to a generated divided word or divided word string is specified, a divided word or a divided word string corresponding to the foreign word is registered in the morpheme-analyzing dictionary as the morpheme, and the morpheme analysis based on registered morpheme information is performed. <P>COPYRIGHT: (C)2008,JPO&INPIT

Description

  The present invention relates to a language analysis system, a language analysis method, and a computer program. More specifically, for example, a language analysis system that accurately performs a process of dividing a term used in a highly specialized field such as a medical field into a segment word (morpheme) as a language unit for applying to data search and the like, and a language The present invention relates to an analysis method and a computer program.

  For example, a process for extracting a term applied in data processing from a natural language document such as a search key applied in a database search or an index as an index of a term dictionary is a technique required in various data processing fields. is there. A language unit applied to data processing such as a search key in search processing is called a morpheme.

  Research on extracting morphemes from natural language documents has been conducted. For example, if it is a common document such as [car runs on road], it is possible to extract morphemes such as [car], [road], and [run] by applying a general morphological analysis system. . The morpheme analysis system is known as a system that applies a predetermined morpheme analysis rule to segment a sentence into morphemes that are semantic minimum units and perform a part-of-speech recognition process.

However, it is difficult to divide technical terms in highly specialized fields such as the medical field into appropriate morphemes. For example, a specific disease name
"Potential autoimmune thyroiditis"
When a document including this disease name is divided into morphemes by applying a general morphological analysis system, the following morphemes are extracted. The following shows the data extracted from “latent autoimmune thyroiditis” in the form of [morpheme, part of speech].
(1) [Latent, noun]
(2) [Self, noun]
(3) [Immunity, Noun]
(4) [Gender, suffix]
(5) [Thyroid, noun]
(6) [Fire, Suffix]

  When the morphemes of the minimum unit (1) to (6) are obtained, in addition to setting of morphemes using (1) to (6) as individual search keys, (1) + (2), (1) to Appropriate morpheme delimiters are set by setting a plurality of consecutive morpheme sets such as (4) and (2) to (5) as one search key or index, and morpheme to be applied to data processing is set. The above (1) to (6) are the smallest unit morphemes obtained by applying the morpheme analysis system, and data obtained by combining a plurality of these smallest unit morphemes can also be applied as search keys and index data. These data are also called morphemes.

But in fact,
(2)-(6) "Autoimmune thyroiditis" and (5)-(6) "Thyroiditis"
Is a word used alone in medical texts, for example, a morpheme worth using as a search key,
(3)-(6) "Immune thyroiditis"
Cannot exist alone. Also,
(2)-(4) "Autoimmunity"
Is a morpheme that can hardly exist alone in medical text, but this (2)-(4) "autoimmunity"
(2)-(6) "Autoimmune thyroiditis"
Besides
"Autoimmune hepatitis"
"Autoimmune hemolytic anemia"
"Autoimmune chronic thyroiditis"
As part of compound nouns, it is used extensively and frequently, and can be recognized as an independent morpheme.

Therefore, “latent autoimmune thyroiditis”
(1) “Latent”
(2)-(4) "Autoimmunity"
(5)-(6) "Thyroiditis"
It is considered appropriate to divide into these three morphemes. For example, performing database search processing using these three morphemes as search keys makes it possible to efficiently acquire effective information with less noise from the database. It can be said that there is.

  However, it is difficult to obtain such an appropriate morpheme segmentation from the above morphological analysis results without specialized knowledge in the medical field.

  As a method for solving the above problems, a method of adding a technical term dictionary (technical term list) in the medical field to a dictionary for morphological analysis (morpheme list) is generally used. In particular, frequently used disease names and site names such as “thyroiditis” are likely to be included in the technical terminology dictionary in the medical field. Therefore, this method can correctly separate “thyroiditis” into one morpheme. On the other hand, morphemes that rarely appear alone in medical text, such as “autoimmunity”, are often not included in the technical term dictionary and are divided into “self” and “immunity”. become. (“Immunity” and “sex” are processed into one morpheme because “sex” is a suffix.)

  If “autoimmune thyroiditis” is included in the technical term dictionary, it is possible to set “autoimmune thyroiditis” as one morpheme, but it is a disease such as “autoimmune chronic thyroiditis”. Those that are used as names but have a very low frequency of appearance are not included in the technical term dictionary, and as described above, “self” and “immunity” cannot be correctly combined into one morpheme.

As a conventional technique for identifying a morpheme that is not included in a technical term dictionary such as “autoimmunity”, Patent Document 1 (Japanese Patent Laid-Open No. 6-19968) discloses a technical term automatic extraction technique. . This method gives each morpheme of the normal morpheme analysis result a score calculated from the part of speech, character type, appearance frequency, etc., and decides whether or not to combine multiple morphemes into technical terms based on the score It is a technique. However, whether or not it is a technical term cannot be determined only by factors such as part of speech, character type, and appearance frequency, and the extraction accuracy of technical term extraction by this method is at a level that is not practically sufficient.
Japanese Patent Laid-Open No. 6-19968

  The present invention has been made in view of the above-mentioned problems, and analyzes a technical term in a highly specialized field such as the medical field and divides the process into appropriate language units to be applied to a search key or the like in a search process. An object of the present invention is to provide a language analysis system, a language analysis method, and a computer program.

The first aspect of the present invention is:
A bilingual dictionary storage means for storing bilingual dictionaries of Japanese and foreign languages other than Japanese,
Based on the registered bilingual dictionary data, one-word-to-multi-word bilingual pairs whose Japanese-foreign language bilingual pairs are different from one-word to one-word are extracted, and based on the Japanese morphological analysis included in the extracted bilingual pairs A one-to-many word extracting unit that generates a plurality of segment words by dividing processing, and that analyzes the foreign language word corresponding to the generated segment word or segment word string based on registration data of the bilingual dictionary;
A morpheme analysis dictionary generation unit that executes processing for registering, in the morpheme analysis dictionary, Japanese corresponding to the segmented word or segmented word string determined to correspond to the foreign language word in the one-to-many word extraction unit; ,
A language analysis system characterized by having

  Furthermore, in one embodiment of the language analysis system of the present invention, the language analysis system further includes a morpheme analysis unit that executes a morpheme analysis to which the morpheme analysis dictionary generated by the morpheme analysis dictionary generation unit is applied. It is characterized by that.

  Furthermore, in one embodiment of the language analysis system of the present invention, the one-to-many word extracting unit is configured such that a Japanese-foreign bilingual translation pair included in the bilingual dictionary registration data is a one-word to one-word bilingual pair. The Japanese word included in the parallel translation pair is output to the morpheme analysis dictionary generation unit, and the morpheme analysis dictionary generation unit converts the Japanese word input from the one-to-many word extraction unit into the morpheme analysis dictionary. The present invention is characterized in that a process for registering is executed.

  Further, in one embodiment of the language analysis system of the present invention, the bilingual dictionary storage means includes a general term bilingual dictionary storing a bilingual dictionary of Japanese and foreign languages for general terms, and Japanese for technical terms. The technical term bilingual dictionary storing a bilingual dictionary with a foreign language is stored, and the one-to-many word extraction unit is configured to search for a bilingual pair of Japanese and foreign language from data registered in the technical term bilingual dictionary. Extract single-word-multi-word bilingual pairs that are different from word-to-one words, generate multiple classifiers by dividing Japanese by morphological analysis of Japanese included in the extracted bilingual pairs, and generate individual classifiers Or it is the structure which performs the process which analyzes the word of the said foreign language corresponding to a division word string based on the registration data of both the said technical term bilingual dictionary and the said general term bilingual dictionary To.

  Furthermore, in one embodiment of the language analysis system of the present invention, the technical term bilingual dictionary is a bilingual dictionary that stores medical terms in the medical field, and the one-to-many word extraction means is a technical term in Japanese. A plurality of categorical words are generated by morphological analysis, and the foreign language words corresponding to the generated categorical words or categorical word strings are analyzed based on registration data of both the technical term bilingual dictionary and the general term bilingual dictionary It is the structure which performs the process to perform.

  Furthermore, in one embodiment of the language analysis system of the present invention, the bilingual dictionary storage means stores a bilingual dictionary of Japanese and English, and the one-to-many word extraction means is registered in the bilingual dictionary. Extract multiple bilingual pairs of Japanese and English bilingual pairs with different one-to-one bilingual pairs of Japanese and English, and split Japanese by morphological analysis of Japanese included in the extracted bilingual pairs It is the structure which performs the process which produces | generates a division word and analyzes the English word corresponding to the produced division word or a division word string based on the registration data of the said bilingual dictionary.

  Furthermore, in one embodiment of the language analysis system of the present invention, the one-to-many word extraction unit generates a plurality of segment words obtained by dividing Japanese by a Japanese morpheme analysis, and the generated segment words are analyzed in the morpheme analysis. Applying the part-of-speech information of the segmented words to generate a segmented word string by executing a grouping process of the segmented words, and executing a process of analyzing a foreign word corresponding to the segmented word string Features.

  Furthermore, in an embodiment of the language analysis system of the present invention, the one-to-many word extraction unit generates a plurality of segment words obtained by dividing Japanese by a Japanese morpheme analysis, and generates a segment word string based on the generated segment words. When generating, the frequency of appearance of the segmented word string in the bilingual dictionary or other database is analyzed, and the processing of analyzing a corresponding foreign language word is performed on the segmented word string having a high frequency of appearance. To do.

  Furthermore, in one embodiment of the language analysis system of the present invention, the language analysis system further includes a foreign language corpus storage means for storing the foreign language text data, and the one-to-many word extraction means is a Japanese language In the foreign language word analysis processing corresponding to the segment word or segment word string generated by morphological analysis, the foreign language text in the corpus storage means is referred to, and the word string that frequently appears in the foreign language text is specified, An analysis process is performed in which the word string is regarded as one foreign language word.

Furthermore, the second aspect of the present invention provides
A language analysis method for executing language analysis processing in a language analysis system,
The one-to-many word extracting means refers to a bilingual dictionary storing a bilingual dictionary of Japanese and a foreign language other than Japanese, and the bilingual translation pair of Japanese and foreign language is one word to one word from the registered data of the bilingual dictionary. The bilingual pairs of different one-word vs. multi-words are extracted, a plurality of segment words are generated by the division processing based on the Japanese morphological analysis included in the extracted bi-translation pairs, and the corresponding segment words or segment word strings are A one-to-many word extraction step of analyzing a foreign language word based on registration data of the bilingual dictionary;
The morpheme analysis dictionary generating means executes a process of registering, in the morpheme analysis dictionary, Japanese corresponding to the segmented word or segmented word sequence determined to correspond to the foreign language word in the one-to-many word extraction step. A morphological analysis dictionary generation step to perform,
A language analysis method characterized by comprising:

  Furthermore, in one embodiment of the language analysis method of the present invention, the language analysis method further includes a morpheme analysis unit that performs a morpheme analysis to which the morpheme analysis dictionary generated in the morpheme analysis dictionary generation step is applied. It has an analysis step.

  Furthermore, in one embodiment of the language analysis method of the present invention, the one-to-many word extraction step includes a Japanese-foreign bilingual translation pair included in the bilingual dictionary registration data for a one-to-one word bilingual pair. Outputting a Japanese word included in the bilingual pair to the morphological analysis dictionary generating means, wherein the morphological analysis dictionary generating step morphologically analyzes the Japanese word input in the one-to-many word extraction step A process of registering in a dictionary for use is executed.

  Furthermore, in one embodiment of the language analysis method of the present invention, the bilingual dictionary includes a general term bilingual dictionary storing a bilingual dictionary of Japanese and foreign languages for general terms, and Japanese and foreign languages for technical terms. And the bilingual word extraction dictionary stores a bilingual dictionary of Japanese and a foreign language from the data registered in the bilingual bilingual dictionary. Extract single-word-to-multi-word parallel translation pairs, generate multiple segment words by dividing Japanese by morphological analysis of Japanese included in the extracted parallel pairs, and correspond to each segment word or segment word string generated This is a step of executing a process of analyzing the foreign language word based on registration data of both the technical term bilingual dictionary and the general term bilingual dictionary.

  Further, in one embodiment of the language analysis method of the present invention, the technical term bilingual dictionary is a bilingual dictionary storing medical technical terms, and the one-to-many word extraction step is Japanese medical terms. A plurality of categorical words are generated by morphological analysis, and the foreign language words corresponding to the generated categorical words or categorical word strings are analyzed based on registration data of both the technical term bilingual dictionary and the general term bilingual dictionary It is the step which performs the process to perform.

  Further, in one embodiment of the language analysis method of the present invention, the bilingual dictionary is a bilingual dictionary of Japanese and English, and the one-to-many word extraction step is performed by using Japanese and English from the data registered in the bilingual dictionary. A bilingual pair of one-word-to-multiple words whose bilingual pairs are different from one-to-one word is extracted, and a plurality of segmented words are generated by dividing Japanese by morphological analysis of Japanese included in the extracted bilingual pairs. It is a step of executing a process of analyzing English words corresponding to a divided word or a divided word string based on registration data of the bilingual dictionary.

  Furthermore, in an embodiment of the language analysis method of the present invention, the one-to-many word extraction step generates a plurality of segment words obtained by dividing Japanese by a Japanese morpheme analysis, and the generated segment words are analyzed in the morpheme analysis. Applying the part-of-speech information of the segmented words to execute a grouping process of the segment words to generate a segment word string, and to execute a process of analyzing a foreign word corresponding to the segment word string Features.

  Furthermore, in an embodiment of the language analysis method of the present invention, the one-to-many word extraction step generates a plurality of segment words obtained by dividing Japanese by a Japanese morphological analysis, and a segment word string based on the segment words generated is generated. When generating, it is a step of analyzing the appearance frequency of the segment word string in the bilingual dictionary or other database, and executing a process of analyzing a corresponding foreign language word for the segment word string having a high appearance frequency. To do.

  Furthermore, in one embodiment of the language analysis method of the present invention, the one-to-many word extraction step includes the step of analyzing the foreign language in the foreign word analysis corresponding to the segment word or segment word string generated by the morphological analysis of Japanese. Analysis processing that refers to a foreign language text stored in a foreign language corpus storage means storing text data, identifies a word string that frequently appears in the foreign language text, and regards the word string as one foreign language word It is characterized by performing.

Furthermore, the third aspect of the present invention provides
A computer program that executes language analysis processing in a language analysis system,
The one-to-many word extraction means refers to a bilingual dictionary storing a bilingual dictionary of Japanese and a foreign language other than Japanese, and the bilingual translation pair of Japanese and foreign language is converted into a one-to-one word from the registered data of the bilingual dictionary. A pair of different one-word-to-multi-word translations is extracted, and a plurality of segment words are generated by a segmentation process based on Japanese morphological analysis included in the extracted pair of translation pairs, and corresponding segment words or segment word strings are generated. A one-to-many word extraction step for analyzing the foreign language word based on registration data of the bilingual dictionary;
The morpheme analysis dictionary generating means executes a process of registering, in the morpheme analysis dictionary, Japanese corresponding to the segmented word or segmented word sequence determined to correspond to the foreign language word in the one-to-many word extraction step. A morphological analysis dictionary generation step to be performed;
There is a computer program characterized by comprising:

  Note that the computer program of the present invention is a recording medium provided in a computer-readable format for a computer system capable of executing various program codes, for example, a recording medium such as a CD, FD, or MO. A computer program that can be provided by a medium or a communication medium such as a network. By providing such a program in a computer-readable format, processing corresponding to the program is realized on the computer system.

  Other objects, features, and advantages of the present invention will become apparent from a more detailed description based on embodiments of the present invention described later and the accompanying drawings. In this specification, the system is a logical set configuration of a plurality of devices, and is not limited to one in which the devices of each configuration are in the same casing.

  According to the configuration of the present invention, a bilingual pair having one-word-to-multi-word registered data in which a bilingual pair of Japanese and foreign language is different from one-word to one-word from registered data in a bilingual dictionary of Japanese and foreign languages. Extracts and analyzes Japanese morphemes contained in the extracted bilingual pairs to generate a plurality of segment words. Based on the registered data in the bilingual dictionary, foreign language words corresponding to the generated segment words or segment word strings are generated. The acquired vocabulary word or word sequence corresponding to the acquired foreign language word is registered as a morpheme in the morpheme analysis dictionary, and morpheme analysis based on the registered morpheme information can be performed. According to this configuration, for example, even in terms of a specialized field such as a medical field, morpheme registration information based on the correct classification can be generated and used, and an accurate morpheme analysis process is realized.

  Hereinafter, a language analysis system, a language analysis method, and a computer program according to embodiments of the present invention will be described in detail with reference to the drawings.

[1. Basic example]
With reference to FIG. 1, the configuration and processing of a language analysis system according to an embodiment of the present invention will be described. As shown in FIG. 1, a language analysis system 100 according to an embodiment of the present invention includes a general term bilingual dictionary storage unit 101, a technical term bilingual dictionary storage unit 102, a one-to-many word extraction unit 103, a morphological analysis dictionary generation unit 104, A morphological analysis unit 105 is included. The one-to-many word extracting unit 103 includes a morpheme analyzing unit 111. The morphological analysis dictionary generation means 104 generates and updates the morphological analysis dictionary 112.

  Hereinafter, the configuration and processing of each unit will be described with specific examples. The language analysis system of the present invention is a system that analyzes documents including technical terms in specialized fields. In the following embodiments, the medical field will be described as an example of specialized fields. The present invention can be applied to analysis of technical terms in various specialized fields such as economics, architecture, technology, etc.

Each means which comprises the language analysis system 100 shown in FIG. 1 is demonstrated.
[General term bilingual dictionary storage means]
The general term parallel translation dictionary storage means 101 is a means for storing a plurality of general Japanese-English parallel translation pairs, and is a dictionary database storing [Japanese-English] parallel translation pairs. As an example of specific stored data, for example,
"Latent-latent"
"Experimental-experimental"
It is configured as a dictionary database storing a plurality of Japanese-English word pairs of general words such as

[Technical term bilingual dictionary storage means]
The technical term parallel dictionary storage means 102 is a dictionary database that stores a plurality of Japanese-English parallel translation pairs of technical terms in the medical field. As an example of specific stored data, for example,
"Thyroiditis"
"Encephalomyelitis-encephalomyelitis"
"Autoimmune thyroiditis-autoimmune thyroiditis"
"Experimental autoimmune encephalomyelitis-experimental autoimmune encephalomyelitis"
It is configured as a dictionary database that stores multiple bilingual pairs of technical terms in Japanese and English.

[One-to-many word extraction means]
The one-to-many word extraction unit 103 includes a morpheme analysis unit 111, which performs morpheme analysis processing on Japanese technical terms extracted from the Japanese-English bilingual translation pair in the technical term bilingual dictionary storage unit 102, and based on morphological analysis A plurality of segment words (morphemes) are generated by the division process of the above, and the foreign language words corresponding to the generated segment words (morphemes) or segment word strings (morpheme strings) are stored in the general term bilingual dictionary storage unit 101 and the technical term bilingual dictionary A morpheme is retrieved from the registered data of the bilingual dictionary stored in the storage means 102, and Japanese corresponding to a segmented word (morpheme) or a segmented word string (morpheme string) determined to correspond to a foreign language word is regarded as one morpheme. Processing to pass to the morphological analysis dictionary generation means 104 as data to be registered in the analysis dictionary is executed.

However, an English technical term extracted from a Japanese-English bilingual translation pair in the technical term bilingual dictionary storage means 102 is a bilingual pair of [Japanese-English] words, for example,
"Thyroiditis"
“Encephalomyelitis-encephalomyelitis” etc. For Japanese technical terms set as a bilingual pair of [Japanese-English] words, the technical term bilingual dictionary is not stored. The Japanese technical terms extracted from the Japanese-English bilingual pair in the means 102 are output to the morphological analysis dictionary generation means 104 as they are. The morpheme analysis dictionary generation unit 104 registers such a Japanese word input from the one-to-multiword extraction unit 103 in the morpheme analysis dictionary 112.

  On the other hand, the one-to-many word extraction unit 103, for example, has a Japanese-English bilingual pair extracted from a Japanese-English bilingual pair in the technical term bilingual dictionary storage unit 102. Extract one-to-many multi-word bilingual pairs different from one word, generate multiple segment words by dividing processing based on Japanese morphological analysis included in the extracted bi-directional pairs, and generate the segment words or segment word strings The corresponding foreign language word is analyzed based on the registered data in the bilingual dictionary. For example, a morphological analysis unit 111 performs morpheme analysis on a Japanese technical term of a bilingual pair in which English is composed of a plurality of words with respect to one Japanese word. This morpheme analysis is performed as a morpheme analysis process to which an existing dictionary for morpheme analysis is applied.

A specific example will be described. For example, in a [Japanese-English] bilingual pair extracted from a Japanese-English bilingual translation pair in the technical term bilingual dictionary storage means 102, a bilingual translation in which English is composed of a plurality of words for one Japanese word. There are the following translation pairs as pairs.
[Autoimmune thyroiditis-autoimmune thyroiditis]
The Japanese terminology of this translation pair,
"Autoimmune thyroiditis"
An example of the above Japanese processing will be described.

The one-to-many word extracting means 103 is the morpheme analyzing means 111,
"Autoimmune thyroiditis"
A morphological analysis process is applied to. As described above, the morpheme analysis is a process of applying a predetermined morpheme analysis rule and segmenting a sentence into segment words (morpheme) that are semantic minimum units to perform a part of speech recognition process. The following analysis results are obtained by morphological analysis. The following data is shown in the form of [morpheme (category), part of speech] (1) [self, noun]
(2) [Immunity, Noun]
(3) [Gender, suffix]
(4) [Thyroid, noun]
(5) [Fire, Suffix]

The morpheme analyzing unit 111 further sets a classification word (morpheme) as an effective data unit in the data processing from the above-described minimum unit morpheme extraction result. For example, when the suffix is suffixed to a noun, the summarization process of (2) + (3) and (4) + (5) is executed by the process of collectively considering it as one morpheme, and “autoimmune thyroid gland” "Fire"
(A) "Self"
(B) “Immunity”
(C) “Thyroiditis”
Are divided into three segment words (morphemes).

  The above-described morpheme summarization process is an example of the summarization process based on part-of-speech analysis. For example, a plurality of minimum morphemes (for example, morphemes (1) to (5) above) are obtained by analyzing stored documents in a document database in a specialized field. It is also possible to perform a process of counting the number of occurrences of the sequence and determining a morpheme sequence having the number of occurrences equal to or greater than a preset threshold value as one morpheme.

Next, the one-to-many word extraction unit 103 uses the bilingual dictionaries in the general term bilingual dictionary storage unit 101 and the technical term bilingual dictionary storage unit 102 to use the three divided words (morpheme) (a) to (c) above. ) Is identified.
(A) "Self"
(B) “Immunity”
(C) “Thyroiditis”

As described above, the technical term bilingual dictionary storage means 102 has [Japanese-English] bilingual pairs as follows.
[Autoimmune thyroiditis-autoimmune thyroiditis]
The above translation pairs are registered.
The one-to-many word extraction unit 103 first converts each English word contained in the above-described bilingual pair into a Japanese-English bilingual translation pair recorded in the general term bilingual dictionary storage unit 101 and the technical term bilingual dictionary storage unit 102. It is searched whether it is registered as. That is, in the above example, it is determined whether or not a single English word is registered as a bilingual pair for each of “autoimmune” and “thyridotis”.

For example, among “autoimmune” and “thyridotis”, only “thyroidis” is stored in the technical term bilingual dictionary storage means 102.
"Thyroiditis"
Is detected as one-to-many word extraction means 103,
[Autoimmune thyroiditis-autoimmune thyroiditis]
It can be specified that “thyroiditis” in the two English words included in the parallel translation pair corresponds to “thyroiditis”.

Furthermore, the one-to-many word extracting means 103 is adapted to the above-mentioned correspondence, that is,
"Thyroiditis"
The determination result of this correspondence and the bilingual pair registered in the technical term bilingual dictionary storage means 102,
[Autoimmune thyroiditis-autoimmune thyroiditis]
Based on these two registered bilingual pair data, the two morphemes in Japanese, “self” and “immunity”, are one English word, ie
"Autoimmune"
It is determined that it corresponds to.

As a result, the one-to-many word extraction unit 103 associates a plurality of English words with respect to one Japanese word, and registers the bilingual pairs registered in the technical term bilingual dictionary storage unit 102.
[Autoimmune thyroiditis-autoimmune thyroiditis]
about,
"Thyroiditis"
"Autoimmune -autoimmune"
The Japanese corresponding to each of these English words is determined. As described above, the one-to-many word extraction unit 103 outputs a segment word (morpheme) or a segment word string (a set of morphemes) corresponding to one English word to the morpheme analysis dictionary generation unit 104 as one morpheme. The morphological analysis dictionary generation means 104 registers “thyroiditis” and “autoimmunity” in the morphological analysis dictionary 112 as morphemes.

Another processing example in the one-to-many word extraction unit 103 will be described. In a Japanese-English bilingual pair extracted from a Japanese-English bilingual translation pair in the technical term bilingual dictionary storage means 102, a bilingual pair in which English is composed of a plurality of words with respect to one Japanese word. As an example,
[Experimental autoimmune encephalomyelitis-experimental autoimmune encephalomyelitis]
A processing example for the parallel translation pair will be described.

The one-to-many word extracting means 103 is the morpheme analyzing means 111,
"Experimental autoimmune encephalomyelitis"
Execute morphological analysis processing for. The following analysis results are obtained as morphological analysis results. The following data is shown in the form of [morpheme, part of speech] (1) [experiment, noun]
(2) [Target, suffix]
(3) [Self, noun]
(4) [Immunity, Noun]
(5) [Gender, suffix]
(6) [Brain, noun]
(7) [Spine, noun]
(8) [Fire, Suffix]

The morpheme analyzing unit 111 further sets a morpheme as an effective data unit in the data processing from the morpheme extraction result of the minimum unit. For example, when a suffix is suffixed to a noun, rounding processing of (1) + (2), (4), + (5), (7) + (8) is executed by processing that is regarded as one morpheme. "Experimental autoimmune encephalomyelitis"
(A) "Experimental"
(B) "Self"
(C) “Immunity”
(D) “Brain”
(E) "Myelitis"
Divided into these five morphemes.

  Next, the one-to-many word extraction unit 103 uses the bilingual dictionaries in the general term bilingual dictionary storage unit 101 and the technical term bilingual dictionary storage unit 102 to correspond to the above five morphemes (a) to (e). Identify English words.

As described above, the technical term bilingual dictionary storage means 102 has [Japanese-English] bilingual pairs as follows.
[Experimental autoimmune encephalomyelitis-experimental autoimmune encephalomyelitis]
The above translation pairs are registered.
The one-to-many word extraction unit 103 first converts each English word contained in the above-described bilingual pair into a Japanese-English bilingual translation pair recorded in the general term bilingual dictionary storage unit 101 and the technical term bilingual dictionary storage unit 102. It is searched whether it is registered as. That is, in the above example,
"Experimental"
"Autoimmune"
"Encephalomyelitis"
It is determined whether or not these words are registered.

As a judgment result,
It is assumed that “encephalomyelitis” is detected as a registered bilingual pair in the technical term bilingual dictionary storage unit 102. That is,
"Encephalomyelitis-encephalomyelitis"
It is assumed that the bilingual pair is extracted as registration data in the technical term bilingual dictionary storage unit 102.

The one-to-many word extraction means 103, based on the detection of the parallel translation pair, from among the five morphological analysis results (a) to (e) above,
(D) “Brain”
(E) "Myelitis"
Is determined to be one morpheme corresponding to one English word.

Furthermore, it is assumed that “experimental” is detected as a registered translation pair in the general term parallel dictionary storage unit 101. That is,
"Experimental-experimental"
It is assumed that the bilingual pair is extracted as registration data in the general term bilingual dictionary storage unit 101.

The one-to-many word extraction means 103, based on the detection of the parallel translation pair, from among the five morphological analysis results (a) to (e) above,
(A) "Experimental"
Is determined as one morpheme corresponding to one English word.

Furthermore, the one-to-many word extraction means 103
"Encephalomyelitis-encephalomyelitis"
"Experimental-experimental"
Based on the correspondence between these English words, the remaining one English word “autoimmune” is the remaining two morphemes in the five morphological analysis results of (a) to (e) above,
(B) "Self"
(C) “Immunity”
It is determined that it corresponds to. That is,
"Autoimmune -autoimmune"
It is determined that there is a correspondence relationship.

As a result, the one-to-many word extraction unit 103 associates a plurality of English words with respect to one Japanese word, and registers the bilingual pairs registered in the technical term bilingual dictionary storage unit 102.
[Experimental autoimmune encephalomyelitis-experimental autoimmune encephalomyelitis]
about,
"Experimental-experimental"
"Autoimmune -autoimmune"
"Encephalomyelitis-encephalomyelitis"
The Japanese corresponding to each of these English words is determined. In this way, the one-to-many word extraction unit 103 outputs a morpheme or a set of morphemes corresponding to one English word to the morpheme analysis dictionary generation unit 104 as one morpheme. The morphological analysis dictionary generation means 104 registers “experimental”, “autoimmunity”, and “encephalomyelitis” in the morphological analysis dictionary 112 as morphemes.

  The one-to-many word extracting unit 103 is thus two or more Japanese morpheme sequences corresponding to one English word, and is included in the general term bilingual dictionary storage unit 101 and the technical term bilingual dictionary storage unit 102. A Japanese morpheme string not extracted is extracted and output to the morpheme analysis dictionary generation means 104. The morpheme analysis dictionary generation unit 104 registers the data input from the one-to-multiword extraction unit 103 in the morpheme analysis dictionary 112 as one morpheme data.

  The concept of the medical field expressed by one word in English is naturally a fixed concept as long as it is expressed by one word. Many medical terms in Japanese are translations of English medical terms. Therefore, a plurality of Japanese morpheme strings corresponding to one English word are very likely to represent one concept, and it is considered appropriate to regard these morpheme strings as one morpheme.

Various Japanese technical terms that are divided into multiple Japanese morphemes in the conventional morphological analysis system, such as “occipital temporal”, “dog caterpillar disease”, “colonitis”, “neoplastic protein”, etc. Also, it is analyzed that these technical terms are associated with one English word by performing the above-described processing of the one-to-many word extracting unit 103. That is,
“Occipitotemporal”
"Toxocariasis"
"Ileocolitis"
“Oncoprotein”, etc.
As described above, “occipital temporal”, “canine roundworm”, “colonitis”, “oncoprotein” were analyzed that these terms correspond to one English word, That is, “occipital temporal”, “canine roundworm”, “colonitis”, and “oncoprotein” can be registered in the morphological analysis dictionary as morphemes.

[Dictionary generation means for morphological analysis]
The morpheme analysis dictionary generation unit 104 is based on a general morpheme analysis dictionary applied by a conventional morpheme analysis system, and additionally stores morpheme data input from the one-to-multiword extraction unit 103 described above. Generate and update the file.

  That is, Japanese words in the general term bilingual dictionary storage unit 101, “Japanese technical terms in a pair whose English technical term is a single English word” in the technical term bilingual dictionary storage unit 102, and one-to-many word extraction A morpheme analysis dictionary 112 in which “Japanese morpheme sequences corresponding to a single word in English technical terms” (for example, “autoimmunity”) extracted by the means 103 is added as morpheme data. Generate and update.

[Morphological analysis means]
The morpheme analysis unit 105 is a unit that executes morpheme analysis processing using the morpheme analysis dictionary 112 generated or updated by the morpheme analysis dictionary generation unit 104. For example, medical text is received as input, morpheme analysis is executed by using the morpheme analysis dictionary 112 generated by the morpheme analysis dictionary generation unit 104, and the analysis result is output.

  By using the morphological analysis dictionary 112 generated by the morphological analysis dictionary generating means 104, for example, "latent autoimmune thyroiditis" and "autoimmune chronic thyroiditis" are respectively expressed as "latent", "autoimmunity", " It becomes possible to correctly divide into “thyroiditis”, “autoimmunity”, “chronic”, and “thyroiditis”.

  With the above configuration, even a technical term including a morpheme such as “autoimmunity” that is not included in the medical terminology dictionary in the medical field can be divided into correct morphemes with high accuracy.

  Next, a processing sequence executed in the language analysis system of the present invention will be described with reference to the flowchart shown in FIG. The processing flow shown in FIG. 2 is composed of five steps S101 to S105. Among these steps, the processing of steps S101 to S103 is the processing of the one-to-many word extraction means 103 shown in FIG. 1, the processing of step S104 is the processing of the morpheme analysis dictionary generation means 104, and the processing of step S104 is the morpheme analysis means. This corresponds to the process 105.

  First, in step S101, the one-to-many word extraction unit 103 performs a morphological analysis process on the Japanese technical terms extracted from the Japanese-English parallel translation pairs in the technical term bilingual dictionary storage unit 102. Specifically, as described above, from the data registered as a Japanese-English bilingual translation pair in the technical term bilingual dictionary storage unit 102, a translation in which a plurality of Japanese morphemes correspond to one word in the English technical term. A pair is extracted, Japanese included in the extracted parallel translation pair is extracted, and morpheme analysis is performed in the morpheme analysis unit 111 of the one-to-multiword extraction unit 103.

The processing of this step S101 is, for example, in the above-described processing example, a Japanese-English bilingual translation pair in the technical term bilingual dictionary storage means 102,
[Autoimmune thyroiditis-autoimmune thyroiditis]
This bilingual paired Japanese technical term “autoimmune thyroiditis” corresponds to a process of performing a morphological analysis process. First, the following analysis results are obtained.
(1) [Self, noun]
(2) [Immunity, Noun]
(3) [Gender, suffix]
(4) [Thyroid, noun]
(5) [Fire, Suffix]
Furthermore, the morpheme analyzing unit 111 determines a morpheme as an effective data unit in the data processing from the morpheme extraction result of the minimum unit. For example, when the suffix is suffixed to a noun, the summarization process of (2) + (3) and (4) + (5) is executed by the process of collectively considering it as one morpheme, and “autoimmune thyroid gland” "Fire"
(A) "Self"
(B) “Immunity”
(C) “Thyroiditis”
Are divided into three morphemes.

In step S <b> 102, the one-to-many word extraction unit 103 specifies the corresponding English word using each bilingual dictionary in the general term bilingual dictionary storage unit 101 and the technical term bilingual dictionary storage unit 102. In the above processing example, English words corresponding to the three morphemes (a) to (c) are specified. That is, [Japanese-English] parallel translation pairs registered in the technical term bilingual dictionary storage means 102,
[Autoimmune thyroiditis-autoimmune thyroiditis]
, For each of “autoimmune” and “thyridotis”, it is determined whether or not a single English word is registered as a bilingual pair.
"Thyroiditis"
Detect that the above translation pair is registered.

  In step S103, based on the word identification result in step S102, the remaining English words and Japanese words that cannot be directly matched from the bilingual pairs registered in the general term bilingual dictionary storage unit 101 and the technical term bilingual dictionary storage unit 102 are displayed. The correspondence with the morpheme string is determined.

In the above processing example, the one-to-many word extraction unit 103 is a bilingual pair registered in the technical term bilingual dictionary storage unit 102 or the general term bilingual dictionary storage unit 101.
[Autoimmune thyroiditis-autoimmune thyroiditis]
"Thyroiditis"
Based on these parallel translation pairs, it is specified that “thyroiditis” corresponds to “thyroiditis”.

  The one-to-many word extraction means 103 outputs the processing results of steps S101 to S103 to the morphological analysis dictionary generation means 104.

  Step S104 is processing of the morpheme analysis dictionary generation unit 104, and adds a morpheme string input from the one-to-many word extraction unit 103 and a Japanese expression corresponding to a single English word in the bilingual dictionary to the morpheme analysis dictionary. Execute the process. Specifically, a Japanese word in the general term bilingual dictionary storage unit 101, a “Japanese technical term in a pair whose English technical term is a single English word” in the technical term bilingual dictionary storage unit 102, and For morpheme analysis in which a “Japanese morpheme sequence in which a plurality of Japanese morphemes correspond to one word in English technical terms” (for example, “autoimmunity”) extracted by the one-to-many word extraction unit 103 is added as morpheme data A dictionary 112 is generated and updated.

  The process of step S105 is a process of the morpheme analysis unit 105, and is a unit that executes the morpheme analysis process using the morpheme analysis dictionary 112 generated or updated by the morpheme analysis dictionary generation unit 104 in step s104. For example, medical text is received as input, morpheme analysis is executed by using the morpheme analysis dictionary 112 generated by the morpheme analysis dictionary generation unit 104, and the analysis result is output.

  With the above configuration, even a technical term including a morpheme such as “autoimmunity” that is not included in the medical terminology dictionary in the medical field can be divided into correct morphemes with high accuracy.

  In the above-described embodiment, the example in which the specialized field is the medical field has been described. However, as described above, the present invention is not limited to the medical field, and various other specialized fields such as economy, architecture, technology, and the like. Applicable to analysis of technical terms in specialized fields. When processing corresponding to each specialized field is performed, processing applying a technical term bilingual dictionary corresponding to each specialized field may be executed.

  The configuration example of the language analysis system shown in FIG. 1 shows an example in which the general term bilingual dictionary storage unit 101 and the technical term bilingual dictionary storage unit 102 are configured as independent units. The storage unit 101 and the technical term bilingual dictionary storage unit 102 may be combined into a single dictionary storage unit.

  In the above-described embodiment, a processing example in which a bilingual dictionary of Japanese and English is applied has been described. However, a configuration in which a bilingual dictionary of languages other than English and Japanese is applied may be adopted. For example, even when a bilingual dictionary such as Japanese and German or Japanese and French is applied, the same processing as in the above embodiment can be realized. That is, by using a bilingual dictionary of a language and Japanese in which words are clearly divided by space symbols, the same processing as the processing applying the above Japanese-English bilingual dictionary can be performed.

[2. Other Examples]
Hereinafter, a plurality of modified embodiments different from the basic embodiment described above will be described.

(2-1) Modified Example 1
In the processing example of the language analysis system described with reference to FIGS. 1 and 2, the one-to-many word extraction unit 103 has Japanese and English pairs extracted from the bilingual pairs registered in the technical term bilingual dictionary storage unit 102. Data configured as a pair of one Japanese word and a plurality of English words is extracted, and the morpheme analysis unit 111 executes a Japanese morpheme analysis. This morpheme analysis is an existing morpheme analysis dictionary. It has been described as an example of execution as a morphological analysis process to which is applied.

  As shown in FIG. 3, the language analysis system 200 according to the first modified example includes a morpheme analysis dictionary generated and updated by a morpheme analysis generation unit in the morpheme analysis performed by the morpheme analysis unit 111 of the one-to-multiword extraction unit 103. 112 is applied. That is, in the previous analysis process, data registered as morphemes based on the Japanese-English term pairs whose correspondence relationship has been found is applied, and the morpheme analysis unit 111 of the one-to-many word extraction unit 103 executes a new morpheme analysis.

  For example, in the analysis process executed earlier, it is possible to specify that “autoimmune” corresponds to “self” and “immunity”, and when “autoimmunity” is additionally registered in the morphological analysis dictionary 112, In the morpheme analysis performed by the morpheme analysis unit 111 of the one-to-many word extraction unit 103, it is possible to immediately determine “autoimmunity” as one morpheme, thereby realizing efficient processing. In addition, by recursively collecting morpheme strings in this way, it is possible to realize extraction of technical term analysis morphemes with less leakage.

(2-2) Modification Example 2
In the basic embodiment described above, when the morpheme analysis process is performed by the morpheme analysis unit 111 of the one-to-multiword extraction unit 103, the grouping process based on the part of speech information obtained by the morpheme analysis is performed. . Specifically, if the suffix is followed by a noun, the process is considered as a single morpheme,
(1) [Self, noun]
(2) [Immunity, Noun]
(3) [Gender, suffix]
(4) [Thyroid, noun]
(5) [Fire, Suffix]
These morphemes,
(A) "Self"
(B) “Immunity”
(C) “Thyroiditis”
The example which performs the process put together as these three morphemes was demonstrated.

  This morpheme summarization processing is an example of summarization processing by part-of-speech analysis. However, by analyzing the bilingual pairs registered in the technical term bilingual dictionary storage means, a plurality of minimum morphemes (for example, the morphemes (1) to (5) above) are analyzed. ) Of a plurality of combination sequences (for example, “autoimmunity” of (1) to (3)) is counted, and a morpheme sequence having an appearance count equal to or greater than a preset threshold is determined as one morpheme. It is good also as a structure which performs a process. That is, when the threshold is set to “2”, “autoimmunity” is added to the morphological analysis dictionary when an example in which “autoimmunity” is summarized twice as in the above example is detected. Set as morpheme data. A morpheme string that has been collected only once is not added.

  In addition, instead of analyzing the bilingual pair registered in the technical term bilingual dictionary storage means, for example, the number of appearances is similarly counted by analyzing the stored document in the document database of the specialized field, and the summary processing is performed based on the comparison with the threshold value. May be executed. By performing such summarization processing, it is possible to realize more accurate extraction of technical term analysis morphemes.

(2-3) Modified Example 3
FIG. 4 shows a configuration example of the language analysis system 300 according to the third modified example. As shown in FIG. 4, the language analysis system 300 according to the modified example 3 includes an English cobus storage unit 301. The English corbus storage means 301 stores a text set in the English medical field.

  In the present configuration example, the one-to-many word extraction unit 103 selects the English in the English corpus storage unit 301 among two (or more) consecutive English words in the English technical term in the technical term bilingual dictionary storage unit 102. An English word sequence that frequently appears in the text set is specified, and the corresponding English morpheme sequence is extracted by regarding the English word sequence as one English word.

  That is, in the above-described basic embodiment, the process of setting a morpheme sequence corresponding to one English word as one morpheme is performed. However, in this configuration example, even if there are a plurality of English words, the English corpus storage means In the case of an English word sequence frequently appearing in the English text set in 301, the English word sequence is regarded as one English word, and a Japanese morpheme sequence corresponding to the English word sequence is extracted. The morpheme is determined as a morpheme, output to the morpheme analysis dictionary generation means 104 and registered in the morpheme analysis dictionary 112.

  According to this configuration example, extraction of technical term analysis morphemes with less leakage can be realized. In the figure, although shown as the English corbus storage means 301, it is possible to apply a language other than English as described above, and other languages other than Japanese, for example, German, French, etc. A corpus storage unit corresponding to a language with a clear word break may be set, and a bilingual dictionary of the same language may be set for each bilingual dictionary storage unit.

  Finally, a hardware configuration example of the information processing apparatus that configures the language analysis system that executes the above-described processing will be described with reference to FIG. A CPU (Central Processing Unit) 501 executes processing corresponding to an OS (Operating System), one-to-multiword extraction processing, morpheme analysis processing, morpheme analysis dictionary generation processing, and the like described in the above embodiments. These processes are executed according to a computer program stored in a data storage unit such as a ROM or a hard disk of each information processing apparatus.

  A ROM (Read Only Memory) 502 stores programs used by the CPU 501, calculation parameters, and the like. A RAM (Random Access Memory) 503 stores programs used in the execution of the CPU 501, parameters that change as appropriate during the execution, and the like. These are connected to each other by a host bus 504 including a CPU bus.

  The host bus 504 is connected to an external bus 506 such as a PCI (Peripheral Component Interconnect / Interface) bus via a bridge 505.

  A keyboard 508 and a pointing device 509 are input devices operated by the user. The display 510 includes a liquid crystal display device, a CRT (Cathode Ray Tube), or the like, and displays various types of information as text and images.

  An HDD (Hard Disk Drive) 511 includes a hard disk, drives the hard disk, and records or reproduces a program executed by the CPU 501 and information. The hard disk is used for storing various dictionaries, that is, general term bilingual dictionaries, technical term bilingual dictionaries, morphological analysis dictionaries described in the above-described embodiments, and further stores various computer programs such as data processing programs. Is done.

  The drive 512 reads data or a program recorded on a removable recording medium 521 such as a magnetic disk, an optical disk, a magneto-optical disk, or a semiconductor memory, and the data or program is read out from the interface 507 and the external bus 506. , And supplied to the RAM 503 connected via the bridge 505 and the host bus 504.

  The connection port 514 is a port for connecting the external connection device 522 and has a connection unit such as USB or IEEE1394. The connection port 514 is connected to the CPU 501 and the like via the interface 507, the external bus 506, the bridge 505, the host bus 504, and the like. The communication unit 515 is connected to a network and executes communication with various databases and other information processing apparatuses.

  Note that the hardware configuration example of the information processing apparatus as the language analysis system shown in FIG. 5 is an example of an apparatus configured by applying a PC, and the language analysis system of the present invention is not limited to the configuration shown in FIG. Any configuration capable of executing the processing described in the above-described embodiments may be used.

  The present invention has been described in detail above with reference to specific embodiments. However, it is obvious that those skilled in the art can make modifications and substitutions of the embodiments without departing from the gist of the present invention. In other words, the present invention has been disclosed in the form of exemplification, and should not be interpreted in a limited manner. In order to determine the gist of the present invention, the claims should be taken into consideration.

  The series of processes described in the specification can be executed by hardware, software, or a combined configuration of both. When executing processing by software, the program recording the processing sequence is installed in a memory in a computer incorporated in dedicated hardware and executed, or the program is executed on a general-purpose computer capable of executing various processing. It can be installed and run.

  For example, the program can be recorded in advance on a hard disk or ROM (Read Only Memory) as a recording medium. Alternatively, the program is temporarily or permanently stored on a removable recording medium such as a flexible disk, a CD-ROM (Compact Disc Read Only Memory), an MO (Magneto optical) disk, a DVD (Digital Versatile Disc), a magnetic disk, or a semiconductor memory. It can be stored (recorded). Such a removable recording medium can be provided as so-called package software.

  The program is installed on the computer from the removable recording medium as described above, or is wirelessly transferred from the download site to the computer, or is wired to the computer via a network such as a LAN (Local Area Network) or the Internet. The computer can receive the program transferred in this manner and install it on a recording medium such as a built-in hard disk.

  Note that the various processes described in the specification are not only executed in time series according to the description, but may be executed in parallel or individually according to the processing capability of the apparatus that executes the processes or as necessary. Further, in this specification, the system is a logical set configuration of a plurality of devices, and the devices of each configuration are not limited to being in the same casing.

  As described above, according to the configuration of the present invention, from the registration data of the bilingual dictionary of Japanese and foreign languages, the translation pairs of Japanese and foreign languages are different from one word to one word. Extract bilingual pairs with registered data, perform Japanese morphological analysis in the extracted bilingual pairs to generate multiple segment words, and translate the foreign language words corresponding to the generated segment words or segment word strings It is possible to acquire based on the registered data in the dictionary, register the segment word or segment word string corresponding to the acquired foreign language word as a morpheme in the dictionary for morpheme analysis, and perform morpheme analysis based on the registered morpheme information did. According to this configuration, for example, even in terms of a specialized field such as a medical field, morpheme registration information based on the correct classification can be generated and used, and an accurate morpheme analysis process is realized.

It is a figure which shows the basic structural example of the language analysis system of this invention. It is a figure which shows the flowchart explaining the process sequence which the language analysis system of this invention performs. It is a figure which shows the structure of one Example of the language analysis system of this invention. It is a figure which shows the structure of one Example of the language analysis system of this invention. It is a figure explaining the hardware structural example of the language analysis system which concerns on one Embodiment of this invention.

Explanation of symbols

DESCRIPTION OF SYMBOLS 100 Language analysis system 101 General term bilingual dictionary storage means 102 Technical term bilingual dictionary storage means 103 One-to-many word extraction means 104 Morphological analysis dictionary generation means 105 Morphological analysis means 111 Morphological analysis means 112 Morphological analysis means 301 English corpus storage means 501 CPU ( Central Processing Unit)
502 ROM (Read-Only-Memory)
503 RAM (Random Access Memory)
504 Host bus 505 Bridge 506 External bus 507 Interface 508 Keyboard 509 Pointing device 510 Display 511 HDD (Hard Disk Drive)
512 drive 514 connection port 515 communication unit 521 removable recording medium 522 external connection device

Claims (19)

A bilingual dictionary storage means for storing bilingual dictionaries of Japanese and foreign languages other than Japanese,
Based on the registered bilingual dictionary data, one-word-to-multi-word bilingual pairs whose Japanese-foreign language bilingual pairs are different from one-word to one-word are extracted, and based on the Japanese morphological analysis included in the extracted bilingual pairs A one-to-many word extracting unit that generates a plurality of segment words by dividing processing, and that analyzes the foreign language word corresponding to the generated segment word or segment word string based on registration data of the bilingual dictionary;
A morpheme analysis dictionary generation unit that executes processing for registering, in the morpheme analysis dictionary, Japanese corresponding to the segmented word or segmented word string determined to correspond to the foreign language word in the one-to-many word extraction unit; ,
A language analysis system comprising: The language analysis system further includes:
2. The language analysis system according to claim 1, further comprising a morpheme analysis unit that executes a morpheme analysis to which the morpheme analysis dictionary generated by the morpheme analysis dictionary generation unit is applied. The one-to-many word extraction means includes
For a bilingual pair in which the bilingual pair of Japanese and foreign language included in the registered data of the bilingual dictionary is a one-to-one bilingual pair, the Japanese word included in the bilingual pair is output to the dictionary generating unit for morphological analysis. ,
The morphological analysis dictionary generation means includes:
The language analysis system according to claim 1, wherein the language analysis system is configured to execute a process of registering the Japanese word input from the one-to-many word extraction unit in a morphological analysis dictionary. The bilingual dictionary storage means includes:
A general term bilingual dictionary that stores a bilingual dictionary of Japanese and foreign languages for general terms, and a technical term bilingual dictionary that stores bilingual dictionaries of Japanese and foreign languages for technical terms,
The one-to-many word extraction means includes
From the data registered in the technical term bilingual dictionary, a bilingual pair of one word to many words in which a bilingual pair of Japanese and a foreign language is different from one word to one word is extracted, and the Japanese translation included in the extracted bilingual pair A plurality of segment words obtained by dividing Japanese by morphological analysis are generated, and the words in the foreign language corresponding to the generated individual segment words or segment word strings are stored in both the technical term bilingual dictionary and the general term bilingual dictionary. The language analysis system according to claim 1, wherein the language analysis system is configured to execute a process of analyzing based on registered data. The technical term bilingual dictionary is a bilingual dictionary storing technical terms in the medical field,
The one-to-many word extraction means includes
A plurality of segment words are generated by morphological analysis of Japanese which is a technical term in the medical field, and the foreign word corresponding to the generated segment word or segment word string is converted into the technical term bilingual dictionary and the general term bilingual dictionary. 5. The language analysis system according to claim 4, wherein the analysis processing is executed based on both of the registered data. The bilingual dictionary storage means is configured to store a bilingual dictionary of Japanese and English,
One-to-many word extraction means
From the data registered in the bilingual dictionary, a bilingual pair of one-word-multiple words whose Japanese-English bilingual pair is different from one-word-one-word is extracted, and by morphological analysis of Japanese included in the extracted bilingual pair It is configured to generate a plurality of segment words obtained by dividing Japanese, and to perform processing for analyzing English words corresponding to the generated segment words or segment word strings based on registration data of the bilingual dictionary The language analysis system according to claim 1. The one-to-many word extraction means includes
Generate multiple segment words by dividing Japanese by Japanese morpheme analysis, and apply segment word summary processing to the generated segment words by applying the part-of-speech information of the segment words analyzed in the morpheme analysis. The language analysis system according to claim 1, wherein the language analysis system is configured to generate a word string and analyze a foreign language word corresponding to the segmented word string. The one-to-many word extraction means includes
Generating a plurality of segment words obtained by dividing Japanese by morphological analysis of Japanese, and when generating a segment word string based on the generated segment words, analyzing the appearance frequency of the segment word string in the bilingual dictionary or other database, 2. The language analysis system according to claim 1, wherein the language analysis system is configured to execute processing for analyzing a corresponding foreign language word for a segment word string having a high appearance frequency. The language analysis system further includes:
A foreign language corpus storage means for storing the foreign language text data;
The one-to-many word extraction means includes
In the foreign language word analysis processing corresponding to the segmented word or segmented word string generated by the Japanese morphological analysis, the foreign language text in the corpus storage means is referred to, and the word string that frequently appears in the foreign language text is 2. The language analysis system according to claim 1, wherein the language analysis system is configured to perform analysis processing that identifies and regards the word string as one foreign language word. A language analysis method for executing language analysis processing in a language analysis system,
The one-to-many word extracting means refers to a bilingual dictionary storing a bilingual dictionary of Japanese and a foreign language other than Japanese, and the bilingual translation pair of Japanese and foreign language is one word to one word from the registered data of the bilingual dictionary. The bilingual pairs of different one-word vs. multi-words are extracted, a plurality of segment words are generated by the division processing based on the Japanese morphological analysis included in the extracted bi-translation pairs, and the corresponding segment words or segment word strings are A one-to-many word extraction step of analyzing a foreign language word based on registration data of the bilingual dictionary;
The morpheme analysis dictionary generating means executes a process of registering, in the morpheme analysis dictionary, Japanese corresponding to the segmented word or segmented word sequence determined to correspond to the foreign language word in the one-to-many word extraction step. A morphological analysis dictionary generation step to perform,
A language analysis method characterized by comprising: The language analysis method further includes:
A morpheme analysis step, wherein the morpheme analysis unit executes a morpheme analysis to which the morpheme analysis dictionary generated in the morpheme analysis dictionary generation step is applied;
The language analysis method according to claim 10, comprising: The one-to-many word extraction step includes:
For a bilingual pair in which the bilingual pair of Japanese and foreign language included in the registration data of the bilingual dictionary is a one-to-one bilingual pair, the Japanese word included in the bilingual pair is output to the dictionary generating unit for morphological analysis Including steps,
The morphological analysis dictionary generation step includes:
The language analysis method according to claim 10, wherein a process of registering the Japanese word input in the one-to-many word extraction step in a morphological analysis dictionary is executed. The bilingual dictionary includes a general term bilingual dictionary storing a bilingual dictionary of Japanese and foreign languages for general terms, and a technical term bilingual dictionary storing a bilingual dictionary of Japanese and foreign languages for technical terms. ,
The one-to-many word extraction step includes:
From the data registered in the technical term bilingual dictionary, a bilingual pair of one word to many words in which a bilingual pair of Japanese and a foreign language is different from one word to one word is extracted, and the Japanese translation included in the extracted bilingual pair A plurality of segment words obtained by dividing Japanese by morphological analysis are generated, and the words in the foreign language corresponding to the generated individual segment words or segment word strings are stored in both the technical term bilingual dictionary and the general term bilingual dictionary. The language analysis method according to claim 10, wherein the language analysis method is a step of executing a process of analyzing based on registered data. The technical term bilingual dictionary is a bilingual dictionary storing technical terms in the medical field,
The one-to-many word extraction step includes:
A plurality of segment words are generated by morphological analysis of Japanese which is a technical term in the medical field, and the foreign word corresponding to the generated segment word or segment word string is converted into the technical term bilingual dictionary and the general term bilingual dictionary. The language analysis method according to claim 13, wherein the language analysis method is a step of executing a process of analyzing based on both of the registered data. The bilingual dictionary is a bilingual dictionary of Japanese and English,
The one-to-many word extraction step
From the data registered in the bilingual dictionary, a bilingual pair of one-word-multiple words whose Japanese-English bilingual pair is different from one-word-one-word is extracted, and by morphological analysis of Japanese included in the extracted bilingual pair A step of generating a plurality of segment words obtained by dividing Japanese and executing processing for analyzing English words corresponding to the generated segment words or segment word strings based on registration data of the bilingual dictionary The language analysis method according to claim 10. The one-to-many word extraction step includes:
Generate multiple segment words by dividing Japanese by Japanese morpheme analysis, and apply segment word summary processing to the generated segment words by applying the part-of-speech information of the segment words analyzed in the morpheme analysis. 11. The language analysis method according to claim 10, wherein the language analysis method is a step of generating a word string and analyzing a foreign language word corresponding to the segmented word string. The one-to-many word extraction step includes:
Generating a plurality of segment words obtained by dividing Japanese by morphological analysis of Japanese, and when generating a segment word string based on the generated segment words, analyzing the appearance frequency of the segment word string in the bilingual dictionary or other database, The language analysis method according to claim 10, which is a step of executing processing for analyzing a corresponding foreign language word for a segment word string having a high appearance frequency. The one-to-many word extraction step includes:
In the foreign language word analysis processing corresponding to the segmented word or segmented word string generated by Japanese morphological analysis, referring to the foreign language text stored in the foreign language corpus storage means storing the foreign language text data, The language analysis method according to claim 10, wherein a word string that appears frequently in the foreign language text is specified, and an analysis process is performed in which the word string is regarded as one foreign language word. A computer program that executes language analysis processing in a language analysis system,
The one-to-many word extraction means refers to a bilingual dictionary storing a bilingual dictionary of Japanese and a foreign language other than Japanese, and the bilingual translation pair of Japanese and foreign language is converted into a one-to-one word from the registered data of the bilingual dictionary. A pair of different one-word-to-multi-word translations is extracted, and a plurality of segment words are generated by a segmentation process based on Japanese morphological analysis included in the extracted pair of translation pairs, and corresponding segment words or segment word strings are generated. A one-to-many word extraction step for analyzing the foreign language word based on registration data of the bilingual dictionary;
The morpheme analysis dictionary generating means executes a process of registering, in the morpheme analysis dictionary, Japanese corresponding to the segmented word or segmented word sequence determined to correspond to the foreign language word in the one-to-many word extraction step. A morphological analysis dictionary generation step to be performed;
A computer program characterized by comprising:

Images