JP4368550B2 - Document search apparatus, document search method, and program causing computer to execute the method - Google Patents

Description

[0001]
BACKGROUND OF THE INVENTION
The present invention relates to a document search apparatus, a document search method, and a program for causing a computer to execute a method for searching an electronic document described in a language different from a search condition described in a certain language.
[0002]
[Prior art]
Conventionally, it is possible to search a document that is called “multilingual document search”, etc., in which the language of the search target document is different from the language of the search condition, for example, a document described in English from a Japanese keyword. .
[0003]
As an example, a search service for WEB documents in various languages has already been put into practical use under the name of “TITAN” by Cyberspace Laboratory and “AltaVista” by AltaVistaCompany. Research papers include “Multilingual Internet Information Retrieval Joint Research Project in AMF” (NTT NEWS RELEASE 1999/02/24).
[0004]
[Problems to be solved by the invention]
However, in each of the above prior arts, either the search condition is converted into the language of the search target document, or conversely, the search target document is converted into the language of the search condition, and both languages are made common once and then searched. I was doing.
[0005]
This conversion requires a translation system between multiple languages and a language replacement system at least at the word level, which requires high preparation and refinement costs for data such as translation dictionaries. Such a load also increases. In addition, since ambiguous words often cause a shift in meaning and concept during the translation process, search results can easily contain noise unintended by the operator, and the search accuracy can be matched to the complexity and complexity of the processing. There was a problem that could not be obtained.
[0006]
An attempt has been made to identify correspondence between words in a plurality of languages without using a dictionary by statistically processing a large amount of documents (Japanese Patent Laid-Open No. 2001-43236), but complicated and large-scale processing is performed. In addition, the current technical level cannot be expected to be as accurate as a manually created dictionary.
[0007]
The present invention has been made in view of the above-described problems of the prior art, and is a document search apparatus and document that can perform a search of an arbitrary language document from an arbitrary language with a simple process and with high accuracy. It is an object to provide a search method and a program for causing a computer to execute the search method.
[0008]
[Means for Solving the Problems]
In order to solve the above-described problems and achieve the object, the document retrieval apparatus according to the first aspect retrieves an electronic document described in the second language from the retrieval condition described in the first language. In the document search apparatus, the first search means for searching the electronic document described in the first language for the electronic document that matches the search condition described in the first language, and the first A second retrieval unit that retrieves an electronic document that is a parallel translation of the electronic document retrieved by the retrieval unit from among the electronic documents described in the second language, and the retrieval is performed according to the second retrieval condition. each Electronic document The keywords included in are extracted, and the extracted keywords Generating means for generating a search condition described in the second language on the basis of an electronic document that matches the search condition generated by the generating means from among the electronic documents described in the second language And a third search means for searching.
[0009]
According to the first aspect of the present invention, while performing multilingual document search, complicated work (for example, machine translation) for eliminating the language difference between the search condition and the search target document is not performed. I can't.
[0010]
A document search apparatus according to a second aspect of the present invention is the first searcher according to the first aspect of the present invention. Steps In the electronic document, an electronic document whose text is semantically similar to a natural sentence constituting the search condition is searched as an electronic document that matches the search condition.
[0011]
According to the second aspect of the present invention, only documents in which the meaning of the entire text, such as fields and topics, matches the search conditions are searched.
[0012]
According to a third aspect of the present invention, there is provided the document retrieval apparatus according to the first or second aspect, wherein the second retrieval unit is retrieved by the first retrieval unit. For all of the above, the electronic document which is the parallel translation is searched from the electronic document described in the second language.
[0013]
According to the third aspect of the invention, the search condition in the second language is generated using the bilingual translation of the document obtained by the search in the first language without omission.
[0014]
According to a fourth aspect of the present invention, there is provided the document search apparatus according to the first or second aspect, wherein the second search means is an electronic document searched by the first search means. Some electronic documents An electronic document having a degree of matching equal to or higher than a predetermined degree of matching with the search condition based on the result of the first search means The electronic document that is the parallel translation is searched from the electronic document described in the second language.
[0015]
According to the invention described in claim 4, the search condition in the second language is only the parallel translation of the document obtained by the search in the first language, for example, which has a particularly high degree of matching with the search condition. Selectively used and generated.
[0016]
According to a fifth aspect of the present invention, there is provided the document retrieval apparatus according to the first aspect, further comprising: the electronic document retrieved by the first retrieval unit. It is characterized by comprising display means for displaying information that can be specified and information that can specify the electronic document searched by the third search means.
[0017]
According to the fifth aspect of the present invention, the search result in the first language and the search result in the second language are displayed together on the screen.
[0018]
According to a sixth aspect of the present invention, there is provided a document retrieval method comprising: retrieving an electronic document described in a second language from a search condition described in the first language. Using computer In the document search method to search, The computer is An electronic document that matches the search condition described in the first language , Written in the first language , Previously stored in a predetermined storage area Electronic document group A first search step for searching from among the electronic documents, and an electronic document that is a parallel translation of the electronic document searched in the first search step , Written in the second language , Previously stored in a predetermined storage area Electronic document group The second search step for searching from the list and the second search condition each Electronic document The keywords included in are extracted, and the extracted keywords Generating a search condition described in the second language based on the electronic document, and an electronic document that matches the search condition generated in the generation process , Written in the second language , Previously stored in a predetermined storage area Electronic document group A third search step for searching from within Execute It is characterized by that.
[0019]
According to the invention described in claim 6, while performing multilingual document search, complicated work (for example, machine translation) for eliminating the language difference between the search condition and the search target document is not performed. I can't.
[0020]
Further, the program according to the invention described in claim 7 executes the method described in claim 6. Above The computer is executed.
[0021]
According to the invention described in claim 7, the method described in claim 6 is executed by a computer.
[0022]
DETAILED DESCRIPTION OF THE INVENTION
Exemplary embodiments of a document search apparatus, a document search method, and a program for causing a computer to execute the method according to the present invention will be explained below in detail with reference to the accompanying drawings.
[0023]
(Basic principle of the invention)
Prior to the description of specific embodiments, first, the basic principle of the present invention will be briefly described. FIG. 1 is an explanatory view schematically showing the basic principle of the present invention. The search target document group in the present invention is assumed to be a large number of WEB documents collected from the Internet, for example. In the figure, a document with an ID starting with J indicates a document described in Japanese, and a document with an ID starting with E indicates a document described in English.
[0024]
Here, homepages of companies operating on a global scale or companies opening in e-marketplaces may be created in a plurality of language versions such as a Japanese version and an English version. In the figure, document J-1 shows a Japanese version of such a home page, and document E-1 shows an English version of the same page. The document J-1 and the document E-1 have the same description content, but differ only in the language for describing the content.
[0025]
A Japanese document having a faithful English translation such as this document J-1 or an English document having a faithful Japanese translation such as document E-1 will hereinafter be collectively referred to as “a document (group) with a translation”. . On the other hand, the documents J-2 and J-3 that have only the Japanese version, or the documents E-2, E-3, and E-4 that have only the English version, are referred to as “documents without translation (group)” below. Collectively.
[0026]
If all the search target documents are bilingual documents, it is possible to easily search for documents in other languages using an arbitrary search condition. That is, for example, in order to be able to search an English document from a Japanese search condition, an English document corresponding to the Japanese document is used as a search result instead of a Japanese document that matches the Japanese search condition. Return it.
[0027]
However, as shown in the figure, an actual search target document includes a document with a translation and a document without a translation. At least for the latter, the language on the search condition side is matched with the language on the search target document side as described above. Or, conversely, the conventional technique is to perform a search after matching the language on the search target document side with the language on the search condition side.
[0028]
In contrast, in the present invention, as will be described in detail below, the text of an English document corresponding to a Japanese document that matches a Japanese search condition is regarded as an English search condition, and an English document that matches the search condition is searched. By returning as a result, an English document is searched from a Japanese search condition without going through complicated and difficult language conversion processing and statistical processing.
[0029]
That is, as schematically shown in FIG. 1, first, a natural sentence input in Japanese by an operator is used as a search condition, and the condition is met (which may be said to be similar to the natural sentence as a whole). Search for Japanese documents ((1) in the figure).
[0030]
Assuming that the Japanese documents J-1 and J-2 meet the above conditions, the corresponding English document E-1 exists in the Japanese document J-1 among these applicable documents. The English natural sentence that is the body of the English document E-1 is used as a new search condition, and an English document that matches the condition (which may be said to be similar to the natural sentence as a whole) is searched (see FIG. Medium (2)).
[0031]
Then, the document that satisfies the new condition, that is, the document E-1 itself, and the documents E-2 and E-3 similar to the document E-1, are presented to the operator as the final search results.
[0032]
In other words, in the present invention, the original Japanese search condition is not translated into English by itself, but a Japanese document similar to the Japanese (document J-1 in this example) has already been created manually. For the sake of convenience, an accurate English translation (document E-1) corresponds to an English translation of the above condition, and an English document is searched using this as a new search condition.
[0033]
A faithful English translation of a Japanese document that matches a Japanese search condition is similar to the natural sentence that is the search condition in terms of content and language, but similar in content. At least, there is less semantic and conceptual deviation from the original search condition written in Japanese than the English translation automatically generated by the conventional machine translation technology. By performing a search using this as a search condition, Noise included in the final search result can be reduced.
[0034]
Also, manual translations are more natural than automatically generated translations (that is, they have a quality that is uncomfortable even if they are read by native speakers), and word usage and logic development methods are also in line with the language. Since it is used, it is considered that the similarity with the search target document described in the same language can be determined more accurately by adopting this as a search condition. However, the translation used as a stepping board does not necessarily have to be created manually. If the deviation from the original text is not large, it may be a machine-translated document. It ’s just that there ’s an advantage like that).
[0035]
As described above, according to the present invention, (1) a search for a Japanese document input by an operator using a Japanese natural sentence as a search condition, and (2) an English document corresponding to the Japanese document obtained by the search. Search for documents written in English from the search conditions written in Japanese by chaining the search of English documents using the natural language of English as the search condition. Is possible.
[0036]
(Embodiment of the Invention)
Next, FIG. 2 is an explanatory diagram showing a hardware configuration of the document search apparatus according to the embodiment of the present invention. In the figure, 201 indicates a CPU that controls the entire apparatus, 202 indicates a ROM that stores basic input / output programs, and 203 indicates a RAM that is used as a work area of the CPU 201.
[0037]
Reference numeral 204 denotes an HDD (hard disk drive) that controls reading / writing of data with respect to the HD (hard disk) 205 under the control of the CPU 201, and 205 denotes an HD that stores data written according to the control of the HDD 204. Yes.
[0038]
Reference numeral 206 denotes an FDD (floppy disk drive) that controls reading / writing of data with respect to the FD (floppy disk) 207 according to the control of the CPU 201, and 207 denotes a removable FD that stores data written according to the control of the FDD 206. Respectively.
[0039]
Reference numeral 208 denotes a cursor, menu, window, or display that displays various data such as characters and images, and 209 is connected to a network such as a LAN via a communication cable 210 and functions as an interface between the network and the CPU 201. Each network I / F is shown.
[0040]
Reference numeral 211 denotes a keyboard having a plurality of keys for inputting characters, numerical values, and various instructions. Reference numeral 212 denotes a mouse for selecting and executing various instructions, selecting a processing target, moving a cursor, and the like. ing. Reference numeral 213 denotes a CD-ROM which is a detachable recording medium, 214 denotes a CD-ROM drive for controlling reading of data to the CD-ROM 213, and 200 denotes a bus or cable for connecting the above-described units. ing.
[0041]
Next, FIG. 3 is an explanatory diagram showing a functional configuration of the document search apparatus according to the embodiment of the present invention. As shown, the document search apparatus according to the present invention includes a document storage unit 300, a Japanese search condition input unit 301, a Japanese document search unit 302, an English search condition generation unit 303, an English document search unit 304, and a search result display unit 305. It is the structure containing.
[0042]
First, the document storage unit 300 is a functional unit that holds a document group to be searched by a Japanese document search unit 302 and an English document search unit 304 described later. Here, it is assumed that the documents in the document storage unit 300 are a large number of WEB documents collected from the Internet, and those written in Japanese are written in the Japanese document storage unit 300a, and those written in English are English. Assume that each document is stored in the document storage unit 300b.
[0043]
As described above, some of the documents in the document storage unit 300 are documents with translations and the rest are documents without translations. The bilingual document holds, as its attribute information (attached information), the ID of the bilingual document described in another language (not limited to the ID, any information that can identify the document). Yes.
[0044]
For example, if the Japanese document J-1 and the English document E-1 are in a parallel translation relationship, the former attribute information is the latter ID “E-1”, and the latter attribute information is the former ID. A certain “J-1” is written in advance.
[0045]
Next, FIG. 4 is a flowchart showing a document search processing procedure of the document search apparatus according to the embodiment of the present invention. In the following, the functions of the remaining units shown in FIG. 3 will be described in order according to the procedure shown in FIG.
[0046]
In step S401, the Japanese search condition input unit 301 of the document search apparatus according to the present invention displays a search condition input screen as shown in FIG. 5 on the display 208 and waits for input from the operator. Then, characters input from the keyboard 211 or the like are sequentially displayed in the search condition input area 500.
[0047]
As shown in the figure, a natural sentence consisting of a plurality of sentences is input here as a search condition. However, a natural sentence consisting of a single sentence, a single or a plurality of keywords, etc. may be used as a Japanese character string. Anything may be used.
[0048]
Next, when it is detected in step S402 that an instruction to execute the search is input, that is, when the search execution button 501 shown in FIG. 5 is clicked with the mouse 212 (step S402: Yes), the Japanese language search is performed. The condition input unit 301 delivers the character string in the input area 500 at that time as a search condition to the Japanese document search unit 302 described later.
[0049]
In response to this, the Japanese document search unit 302 searches the Japanese document storage unit 300a based on the search conditions in step S403. This search method by the Japanese document search unit 302 can search a document described in a language from search conditions described in a certain language (even if it can search a document in a single language). However, here, a method generally called “vector space method” is adopted.
[0050]
The “vector space method” calculates the cosine distance between the feature vector of a search condition and the feature vector of each document to be searched, and uses a document whose absolute or relatively small distance as a search condition. It is presented to the operator as a matching document that matches.
[0051]
The feature vector here is an n-dimensional vector composed of n element values corresponding to n keywords (vocabulary), and each element value is most simply determined by the appearance frequency of the corresponding keyword. It is determined. For example, the feature vector of a document that contains only one keyword in the body has only the value of the element corresponding to the keyword, such as (0, 1, 0, 0,...), And the remaining n−1. A vector whose element values are all zero.
[0052]
In this vector space method, a document whose tendency as a whole of the keywords appearing in the text is similar to the search condition has a smaller distance from the search condition, and therefore is more likely to be a conforming document. For this reason, there is an advantage that there is less noise in the search result compared to a Boolean search (general keyword search) in which documents are simply selected based on whether or not a specific keyword in the search condition is included.
[0053]
The Japanese document search unit 302 ranks or scores the documents to be searched on the basis of the above distance, and from the highest rank / highest score document to a certain number of documents, or a predetermined rank / all higher than a predetermined score Documents with a high degree of match with the search condition, such as documents of No. Then, the IDs of these conforming documents (not limited to IDs, any information that can identify the document) are delivered to an English search condition generation unit 303 and a search result display unit 305, which will be described later.
[0054]
Returning to the procedure of FIG. 4, next, in step S404, the English search condition generation unit 303 to which the search result is transferred from the Japanese document search unit 302 sets the attribute information of each document specified by the transferred ID to Japanese. Read from the word document storage unit 300a. Then, it is determined whether or not at least one ID of an English document is included therein, that is, whether or not at least one document with a translation is included in a Japanese document picked up by the search.
[0055]
If even one translated document is included in the result (step S404: Yes), then in step S405, the text of the English document corresponding to each translated document, that is, the attribute read above. The text of the English document specified by each ID in the information is sequentially read from the English document storage unit 300b. Then, a search condition to be given to the English document search unit 304 described later is generated from these English natural sentences.
[0056]
Here, the English search condition generation unit 303 reads out the corresponding bilingual documents for all the bilingual documents in the search result by the Japanese document search unit 302, but selects some bilingual documents and selects the bilingual documents. It is also possible to read out only.
[0057]
For example, the translation is limited to documents that have a particularly high degree of match with the search conditions, such as a certain number of documents from the highest ranking / highest score among all conforming documents, or all documents with a predetermined ranking / predetermined score. Adopt as an English search condition. In other words, even if the document is a conforming document and has a low degree of matching with the search condition, the existence of the parallel translation is ignored.
[0058]
In this way, by adopting bilingual translation of a particularly high-level document among conforming documents, the Japanese search condition input from the Japanese search condition input unit 301 and the English generated by the English search condition generation unit 303 are used. There is less deviation from the search condition, and the final search result is narrowed down and has a high relevance rate. On the other hand, since the recall rate is low, if a search without omission is required, all the translations that exist for Japanese conforming documents should be adopted as English search conditions as in this example. That's fine.
[0059]
When there are a plurality of English documents adopted as search conditions, the English search condition generation unit 303 combines the texts of the respective documents into a continuous natural sentence, and delivers it to an English document search unit 304 described later. In response to this, the English document search unit 304 searches for an English document similar to the natural sentence as a whole.
[0060]
Of course, the text may be regarded as one search condition for each adopted document, and each document may be separately transferred to the English document search unit 304, and the same search may be repeated for the number of documents. In this case, the search result display unit 305, which will be described later, distinguishes and displays English documents searched according to each condition, or collects and displays the union of English documents searched according to each condition. It is possible.
[0061]
It should be noted that keywords commonly used in each of the documents adopted above (or at least many of them) and keywords that are included only in the adopted documents and not included in other documents (adopted) It is also possible to specify only keywords that characterize the document group that has been used for the entire document group that is the parent document, and generate search conditions from these keywords.
[0062]
Next, the English document search unit 304 handed over the search condition from the English search condition generation unit 303 searches the English document storage unit 300b in step S406, and the ID (not limited to the ID) of the English document that matches the above condition. Any information can be used as long as it can identify the document), and the search result display unit 305 described later is handed over.
[0063]
The search of the English document by the English document search unit 304 is performed by the vector space method here, similarly to the search of the Japanese document by the Japanese document search unit 302. However, it is not necessarily limited to this method, and the search by both functional units does not have to be essentially the same. For example, the Japanese document search unit 302 searches for a document that is substantially similar to a natural sentence that is a search condition by the above-described vector space method, and the English document search unit 304 uses an English document that is adopted as a search condition by a Boolean search. You may make it search only the document which contains the keyword in the inside reliably.
[0064]
Next, in step S407, the search result display unit 305 is handed over each ID of the Japanese document as a search result from the Japanese document search unit 302 and each ID of the English document as a search result from the English document search unit 304. A search result display screen as shown in FIG.
[0065]
In the figure, each headline of a Japanese document searched by the Japanese document search unit 302 is displayed in the Japanese document display frame 600, and each headline of the English document searched by the English document search unit 304 is displayed in the English document display frame 601. Are displayed in descending order of the degree of match with the search condition. When this heading is clicked with the mouse 212, the text of the document having the heading can be displayed. In the figure, the document ID is also displayed in parentheses next to the heading, but this may or may not be displayed.
[0066]
According to the embodiment described above, a document described in a certain language can be searched with a search condition in another language, but complicated translation processing and statistical processing are not performed in the process. It does not occur and the load on the system is much smaller than in the prior art.
[0067]
In addition, it is not necessary to prepare a dictionary for translation between languages, which has conventionally required a lot of time and labor, and it is only necessary to collect a group of documents including partially translated documents. In recent years, since there are not a few electronic documents that have been created with translation in advance, including WEB documents, this collection is also easy. In other words, the present invention pays attention to the fact that the collected documents often include bilingual documents, and uses this situation to get an idea that a multilingual document search can be realized as a foothold. It is.
[0068]
In the above-described embodiment, the search for the WEB document on the Internet is taken as an example. However, in addition to the search for the FAQ document in the internal network of the international company, the search for academic papers in various languages in various research / education institutions, etc. The present invention is also applicable.
[0069]
In the embodiment described above, an English document is searched from Japanese, but it goes without saying that a Japanese document can be searched from English. In addition, as schematically shown in FIG. 7, for example, an English document having a translation relationship is obtained from a Japanese document searched in Japanese, and a German document having a translation relationship is acquired from the English document searched in the text of the English document. If a German document is retrieved and a German document retrieved from the body of the German document is taken as a search result, the German document is retrieved from the Japanese search condition as a result. It is possible to search for documents in any language from any language by tracing the documents in a bilingual relationship to an expression that spans multiple languages.
[0070]
In FIG. 1 and FIG. 7, the bilingual document and the document group searched by the bilingual document are drawn as if they existed in the same database, but this is not necessarily the base of the bilingual document extraction. It is not necessary to perform a search for the document group using the corresponding bilingual document. That is, for example, a bilingual translation of a Japanese document searched under a Japanese search condition may be acquired from the database A, and a database B different from that may be searched using the text as an English search condition.
[0071]
Note that the above-described Japanese document search unit 302 corresponds to “first search means” in the claims, and the processing to be performed corresponds to “first search step” in the claims. In addition, the English search condition generation unit 303 serves as the “second search unit” and the “generation unit” in the claims, and the processing to be performed includes the “second search step” and the “generation step” in the claims. It is. Further, the English document search unit 304 corresponds to “third search means” in the claims, and the processing performed by the English document search unit 304 corresponds to “third search step” in the claims. Further, the search result display unit 305 corresponds to “display means” in the claims.
[0072]
The above-described Japanese search condition input unit 301 to search result display unit 305 realize the functions of the respective units when the CPU 201 executes command processing in accordance with program commands read from the HD 205 or the like to the RAM 203. Is. In particular, the Japanese document search unit 302 and the English document search unit 304 are realized by “ConceptBase Search” which is a document search engine manufactured and sold by the present applicant.
[0073]
In addition to the HD 205, the above program can be stored in various recording media such as the FD 207, the CD-ROM 213, or the MO, and can be distributed via this medium, and can also be distributed via a network.
[0074]
【The invention's effect】
As described above, the invention according to claim 1 is described in the first language in the document search apparatus for searching for an electronic document described in the second language from the search condition described in the first language. A first search means for searching for an electronic document that matches the searched search condition from among the electronic documents described in the first language, and an electronic document that is a parallel translation of the electronic document searched by the first search means A second retrieval means for retrieving the document from the electronic document described in the second language, and the document is retrieved by the second retrieval condition. each Electronic document The keywords included in are extracted, and the extracted keywords Generating means for generating a search condition described in the second language on the basis of an electronic document that matches the search condition generated by the generating means from among the electronic documents described in the second language And a third search means for searching, so that while performing multilingual document search, no complicated work is performed to eliminate language differences between search conditions and search target documents. Thus, there is an effect that a document search apparatus capable of searching a document in an arbitrary language from an arbitrary language with a simple process is obtained.
[0075]
The invention according to claim 2 is the invention according to claim 1, in which the first searcher is used. Steps In the electronic document, an electronic document whose text is semantically similar to the natural text that constitutes the search condition is searched as an electronic document that matches the search condition. Only documents that match the search condition are searched, and this makes it possible to obtain a document search apparatus that can perform a search for a document in an arbitrary language from an arbitrary language with a simple process and with high accuracy. Play.
[0076]
According to a third aspect of the present invention, in the first or second aspect of the invention, the second search unit is configured to search for all electronic documents searched by the first search unit. Since the electronic document which is the parallel translation is searched from the electronic documents described in the second language, the parallel translation of the document obtained by the search in the first language is used as the search condition in the second language. As a result, a document search apparatus capable of performing a search for a document in an arbitrary language from an arbitrary language with a simple process and with high accuracy (specifically, a high reproduction rate) is obtained. There is an effect that is.
[0077]
According to a fourth aspect of the present invention, in the first or second aspect of the present invention, the second search unit includes a part of the electronic document searched by the first search unit. Electronic document An electronic document having a degree of matching equal to or higher than a predetermined degree of matching with the search condition based on the result of the first search means Therefore, the electronic document that is the parallel translation is searched from the electronic documents described in the second language. Therefore, the search condition in the second language is, among the documents obtained by the search in the first language, For example, only bilingual translations that have a particularly high degree of match with the search conditions are selectively used and generated. This makes it easy to search for documents in any language from any language with high accuracy (specifically In particular, it is possible to obtain a document retrieval apparatus that can be performed).
[0078]
Further, the invention according to claim 5 is the invention according to any one of claims 1 to 4, further comprising: information that can specify the electronic document searched by the first search means; and Since the display means for displaying the information that can identify the electronic document searched by the third search means is provided, the search result in the first language and the search result in the second language are displayed together on the screen, Thus, there is an effect that a document search apparatus capable of easily presenting the result of multilingual document search including the intermediate result to the operator can be obtained.
[0079]
According to a sixth aspect of the present invention, an electronic document described in a second language is retrieved from a search condition described in the first language. Using computer In the document search method to search, The computer is An electronic document that matches the search condition described in the first language , Written in the first language , Previously stored in a predetermined storage area Electronic document group A first search step for searching from among the electronic documents, and an electronic document that is a parallel translation of the electronic document searched in the first search step , Written in the second language , Previously stored in a predetermined storage area Electronic document group The second search step for searching from the list and the second search condition each Electronic document The keywords included in are extracted, and the extracted keywords Generating a search condition described in the second language based on the electronic document, and an electronic document that matches the search condition generated in the generation process , Written in the second language , Previously stored in a predetermined storage area Electronic document group A third search step for searching from within Execute Therefore, while realizing multilingual document search, no complicated work is performed to eliminate the language difference between the search condition and the search target document. There is an effect that a document search method capable of performing a search by a simple process is obtained.
[0080]
According to the invention described in claim 7, the method described in claim 6 is performed. Above There is an effect that a program that can be executed by a computer is obtained.
[Brief description of the drawings]
FIG. 1 is an explanatory view schematically showing the basic principle of the present invention.
FIG. 2 is an explanatory diagram showing a hardware configuration of a document search apparatus according to an embodiment of the present invention.
FIG. 3 is an explanatory diagram showing a functional configuration of a document search apparatus according to an embodiment of the present invention.
FIG. 4 is a flowchart showing a document search process performed by the document search apparatus according to the embodiment of the present invention.
FIG. 5 is an explanatory diagram showing an example of a search condition input screen in the document search apparatus according to the embodiment of the present invention.
FIG. 6 is an explanatory diagram showing an example of a search result display screen in the document search apparatus according to the embodiment of the present invention.
FIG. 7 is an explanatory view schematically showing the basic principle of another embodiment of the present invention.
[Explanation of symbols]
200 Bus or cable
201 CPU
202 ROM
203 RAM
204 HDD
205 HD
206 FDD
207 FD
208 display
209 Network I / F
210 Communication cable
211 keyboard
212 mouse
213 CD-ROM
214 CD-ROM drive
300 Document storage
300a Japanese document storage
300b English document storage
301 Japanese search condition input part
302 Japanese Document Search Department
303 English search condition generator
304 English Document Search Department
305 Search result display area

Claims (7)

In a document search apparatus for searching an electronic document described in a second language from a search condition described in a first language,
First search means for searching for an electronic document that matches a search condition described in the first language from a group of electronic documents described in the first language;
Second search means for searching an electronic document that is a parallel translation of the electronic document searched by the first search means from an electronic document group described in the second language;
A keyword included in common to each of the electronic documents retrieved by the second search condition, only included in the electronic documents retrieved by the second retrieval condition, as described by the second language Generating means for extracting keywords that are not included in other electronic documents in the electronic document group, and generating search conditions described in the second language based on the extracted keywords;
Third search means for searching for an electronic document that matches the search condition generated by the generation means from among the electronic documents described in the second language;
A document retrieval apparatus comprising:   The first search means searches for an electronic document whose body is semantically similar to a natural sentence that constitutes the search condition in the electronic document as an electronic document that matches the search condition. The document search apparatus according to claim 1.   The second search means searches for an electronic document that is a parallel translation of all the electronic documents searched by the first search means from electronic documents described in the second language. The document search apparatus according to claim 1 or 2, wherein:   The second search means is a part of the electronic documents searched by the first search means, and a predetermined match with the search condition based on a result by the first search means 3. The document search apparatus according to claim 1 or 2, wherein an electronic document that is a parallel translation of an electronic document that is greater than or equal to a predetermined degree is searched from electronic documents described in the second language.   The information processing apparatus further comprises display means for displaying information that can specify the electronic document searched by the first search means and information that can specify the electronic document searched by the third search means. The document search device according to any one of claims 1 to 4. In a document retrieval method for retrieving an electronic document described in a second language from a search condition described in a first language using a computer,
The computer is
A first search step of searching for an electronic document that matches a search condition described in the first language from a group of electronic documents described in the first language and previously stored in a predetermined storage area; ,
A second search for searching an electronic document that is a parallel translation of the electronic document searched in the first search step from an electronic document group described in the second language and stored in a predetermined storage area in advance. Process,
A keyword that is commonly included in each electronic document searched by the second search condition, is included only in the electronic document searched by the second search condition, and is described in the second language. Generating keywords that are not included in other electronic documents in a group of electronic documents stored in a predetermined storage area in advance, and generating search conditions described in the second language based on the extracted keywords Process,
A third search step for searching for an electronic document that matches the search condition generated in the generation step from an electronic document group described in the second language and stored in a predetermined storage area in advance;
The document retrieval method characterized by performing.   A program for causing a computer to execute the method according to claim 6.

Images