JP6978735B2 - Document search device, document search method, and document search program - Google Patents

Description

The present invention relates to a document retrieval device, a document retrieval method, and a document retrieval program, and in particular, utilizing the correlation between documents in a document set to be searched, a text contained in one document and the one. The present invention relates to a document retrieval device that generates an extended document set based on a document that correlates with a document and reduces search omissions for search conditions.

In recent years, devices for evaluating inventions before filing have been developed. Patent Document 1 discloses a patent specification evaluation / preparation work support device that stores information on intellectual property, evaluates application documents before filing, and supports preparation. In the evaluation of an invention before filing an application, a search process such as a prior art document search or a clearance search is usually performed in order to determine whether or not the substantive requirements for registration such as novelty and inventive step are satisfied and whether or not there is infringement. Is required to do.

Japanese Unexamined Patent Publication No. 2010-224984

Conventionally, in the search process, there is a method of reconstructing a search query by expanding keywords in a search text input as a search condition (search query) with synonyms using a synonym dictionary or the like. However, it is difficult to properly create a dictionary with the above method, and there is a problem that the expanded (added) word as a synonym becomes noise and searches for documents that are not related to the search query. be.

In addition, so-called similar documents are searched by calculating the degree of similarity (distance) in a vector space that abstracts the meaning of words instead of character strings of words by using distributed expressions and vector expressions of words. A concept search is being performed. In the concept search, there are a method of expanding the word itself in the input search text with similar words in the vector space, and a method of converting the document into the vector space in advance and having it as a feature quantity. In either case, vectorization is generally performed by machine learning, but the similarity between completely unrelated documents or words becomes high, which may cause noise. In general, words often have different meanings depending on the target field, and it is necessary to construct a vector space for each target field.

Further, there is a method of classifying (clustering) the similarity between documents by words included in the document and using the peripheral documents of the document by keyword matching as the search result. In this method, it is necessary to calculate the similarity between documents in advance, but the method of calculating the similarity at that time may differ from the method of calculating the similarity required when actually searching, so it is completely different. There was a problem of searching for unrelated documents.

The present invention has been made in view of the above, and provides a document search device or the like that reduces search omissions for search conditions.

The document retrieval device according to the embodiment of the present invention is
A document search device equipped with a search unit that searches for documents corresponding to the received search content.
A correlation information storage unit that stores correlation information about one document included in a document set and a document that is correlated with the one document.
Based on the correlation information, a document that is correlated with the one document is used to generate one extended document that is an extension of the one document, and the correlation for a plurality of documents included in the document set. A document extension unit that generates an extended document about the document based on the information.
The search unit searches for the extended document corresponding to the search content from the extended document for a plurality of documents, and outputs one document corresponding to the searched extended document as a search result.

In the document retrieval apparatus according to the embodiment of the present invention.
The search unit receives the search text as the search content and receives the search text.
The extended document whose matching degree in graph units of the search text is equal to or more than a predetermined value is searched, and the search is performed.
One document corresponding to the searched extended document may be output as a search result.

In the document retrieval apparatus according to the embodiment of the present invention.
The search unit receives the search text as the search content, searches for the extended document containing the keyword included in the search text, and outputs one document corresponding to the searched extended document as a search result. May be good.

In the document retrieval apparatus according to the embodiment of the present invention.
The document set is a set of patent documents having bibliographic information, and the correlation information storage unit contains correlation information between the one document and the plurality of other documents based on the bibliographic information. It may be accumulated.

In the document retrieval apparatus according to the embodiment of the present invention.
The document set is a set of patent documents having bibliographic information.
Correlation information between the one document and the plurality of other documents may be accumulated in the correlation information storage unit based on the information regarding the examination or trial of the one document.

In the document retrieval apparatus according to the embodiment of the present invention.
The document set is a set of patent documents having bibliographic information and non-patent documents having bibliographic information, and the correlation information storage unit is a non-patent with correlation based on the bibliographic information of the one document. It may further include a correlation information document database in which information related to the document is further accumulated and document data of the non-patent document is accumulated.

In the document retrieval apparatus according to the embodiment of the present invention.
The document extension unit generates an extended document of the one document by adding a part or all of the graph structure of the document correlated with the one document to the graph structure of the one document. May be good.

In the document retrieval apparatus according to the embodiment of the present invention.
The document extension unit is a graph structure of the one document, which is a tree structure graph structure utilizing the dependency structure of the text, and a graph structure of the document having a correlation with the one document, and is a text. By adding a part or all of the tree structure graph structure using the dependency structure, an extended document of the above-mentioned one document may be generated.

According to one embodiment of the invention
This is a document search method that searches for documents that correspond to the received search content.
A step of extracting a document that is correlated with one document included in the document set from the correlation information storage unit, and one extended document that is an extension of the one document by using the correlated document. Is executed, and the document expansion step creates the extended document for the plurality of documents based on the correlation of the correlation information storage unit.
The computer searches for the extended document corresponding to the search content from the extended document for the plurality of documents, and executes a search step of outputting one document corresponding to the searched extended document as a search result. ..

According to one embodiment of the invention
A document search program that searches for documents that correspond to the received search content, and has a function to extract correlation information about documents that correlate with one document included in the document set from the correlation information storage unit, and the correlation information. To realize a document extension function that generates one extended document that is an extension of the one document, and the document extension function is based on the correlation information stored in the correlation information storage unit. Create the extended document for multiple documents and
Further, the computer has a document search function that searches for an extended document corresponding to the search content from the extended documents for the plurality of documents and outputs one document corresponding to the searched extended document as a search result. make it happen.

According to one embodiment of the present invention, it is possible to provide a document retrieval device or the like that reduces search omissions with respect to search conditions.

It is an example of the functional block diagram of the document retrieval apparatus which concerns on one Embodiment of this invention. This is an example of the hardware configuration of the document retrieval device (computer) according to the embodiment of the present invention. This is an example of a document database. (A) is an example of generating an extended document, and (b) is an example of an extended document database. It is a flow chart of the search process which concerns on one Embodiment of this invention. It is a schematic diagram of the intellectual property creation support system configuration which concerns on one Embodiment of this invention.

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

<Hardware configuration>
First, the hardware configuration of the document retrieval device (computer) 100 according to the embodiment of the present invention will be described with reference to FIG. The document retrieval device 100 includes a processor 101, a memory 102, a storage 103, an input / output interface (I / F) 104, and a communication I / F 105, and is described in the present embodiment by their cooperation. Realize functions and methods. For example, the function or method of the present disclosure is realized by the processor 101 executing an instruction included in a program read in the memory 102.

The processor 101 executes a function and / or a method realized by a code or an instruction contained in a program stored in the storage 103. The processor 101 is, for example, a central processing unit (CPU), an MPU (Micro Processing Unit), a GPU (Graphics Processing Unit), a microprocessor (microprocessor), a processor core (processor core), a multiprocessor (multiprocessor), and an ASIC (Application-). Specific Integrated Circuit), FPGA (Field Programmable Gate Array), etc. are included, and each is implemented by a logic circuit (hardware) or a dedicated circuit formed in an integrated circuit (IC (Integrated Circuit) chip, LSI (Large Scale Integration)), etc. Each process disclosed in the form may be realized. Further, these circuits may be realized by one or a plurality of integrated circuits, and a plurality of processes shown in each embodiment may be realized by one integrated circuit. In addition, LSI may be referred to as VLSI, super LSI, ultra LSI, or the like depending on the degree of integration.

The memory 102 temporarily stores the program loaded from the storage 103 and provides a work area to the processor 101. Various data generated while the processor 101 is executing the program are also temporarily stored in the memory 102. The memory 102 includes, for example, a RAM (Random Access Memory), a ROM (Read Only Memory), and the like.

The storage 103 stores the program. The storage 103 includes, for example, an HDD (Hard Disk Drive), an SSD (Solid State Drive), a flash memory, and the like.

The communication I / F 105 is implemented as hardware such as a network adapter, communication software, and a combination thereof, and transmits / receives various data via the network NET. The communication may be executed by wire or wirelessly, and any communication protocol may be used as long as mutual communication can be executed. The communication I / F 105 executes communication with, for example, another information processing device such as Yuna of the user terminal via the network NET. The communication I / F 105 transmits various data to other information processing devices according to instructions from the processor 101. Further, the communication I / F 105 receives various data transmitted from other information processing devices and transmits them to the processor 101.

The input / output I / F 104 includes an input device for inputting various operations to the document retrieval device 100 and an output device for outputting the processing result processed by the document retrieval device 100. The input / output I / F 104 may be integrated with the input device and the output device, or may be separated into the input device and the output device. The input device is realized by any one of all kinds of devices capable of receiving an input from a user and transmitting information related to the input to the processor 101, or a combination thereof. The input device includes, for example, a hardware key such as a touch panel, a touch display, and a keyboard, a pointing device such as a mouse, a camera (operation input via an image), and a microphone (operation input by voice). The output device outputs the processing result processed by the processor 101. The output device includes, for example, a touch panel, a speaker, and the like.

<Functional configuration>
Next, the functional configuration of the document retrieval apparatus 100 according to the embodiment of the present invention will be described with reference to FIG. It should be noted that each functional unit shown in FIG. 1 is not essential, and other functional units may be provided. Further, the functions or processes of each functional unit may be realized by machine learning or AI (Artificial Intelligence) to the extent feasible.

The document retrieval device 100 includes a correlation information storage unit 110, a document expansion unit 120, a search unit 130, a correlation information document database 140, and an extended document database 150. The correlation information document database 140 and the extended document database 150 may be provided outside the document retrieval device 100, for example, on the cloud, and may be accessible via a network. According to one embodiment of the present invention, the document retrieval device 100 expands the document data of the document database 200 or the correlation information document database 140 to be searched to generate the extended document database 150, and creates the extended document database 150. By making it a search target, it is possible to reduce search omissions.

The correlation information storage unit 110 includes correlation information between one document included in the document set (document database 200) and a plurality of other documents different from the one document included in the document set. Further, as will be described later, the correlation information storage unit 110 can also include correlation information regarding a non-patent document that correlates with one document included in the document set, and document data of the non-patent document. Here, when the document database 200 is, for example, the database of the Japan Patent Office that stores the patent documents, the bibliographic information in the patent documents can be used as the correlation information between the documents. For example, the correlation information storage unit 110 can store correlation information created based on, for example, the degree of agreement of the inventor, the degree of agreement of patent classifications such as IPC, and the applicant. Further, the correlation information storage unit 110 includes a document and another document created based on the number of shares of the keywords in the text included in one document and the keywords in the text included in the other document. Correlation information may be accumulated.
Further, the correlation information storage unit 110 may store information regarding the cited document of the notice of reasons for refusal cited by the examiner in the examination of one document. Further, the correlation information storage unit 110 may store information on the documents cited in the trial case of one document. In addition, the correlation information storage unit 110 is associated with the degree of agreement of the inventor, the degree of agreement of patent classifications such as IPC, the applicant, and the like in advance, and correlates with non-patent documents (technical papers, newspapers, press releases of companies). Information) information may be accumulated. At that time, the document data of the document can also be stored in the correlation information storage unit 110.
The correlation information is not limited to the above-mentioned information, and information created based on an existing method for evaluating the correlation between documents may be accumulated.
Further, although the embodiment in which the correlation information is accumulated in advance is described in the correlation information storage unit, as another possible embodiment, the extraction function unit equipped with the existing technology for evaluating the correlation between documents is provided in the correlation information storage unit. By providing information on correlated documents from the document database and correlation information document database based on the text content and bibliographic information of the search content at each search timing, the correlation information is stored in the correlation information storage unit each time. It is also possible to.

Here, an example of the document database 200 will be described with reference to the figures. FIG. 3 is a conceptual diagram of the document database 200, and the document database 200 can be considered to be associated with a set of a plurality of texts (text group) for each identifier (document ID) that identifies each document. For example, a document with a document ID "ID_X" is associated with "text a", "text b", "text c", "text d", and the like.

The document extension unit 120 uses a similar document in which the correlation information for one document is equal to or more than a predetermined value among a plurality of other documents different from one document included in the document database 200, and one document is used. Generates one extended document that extends.
Further, the document expansion unit 120 expands one document by using the document data of the non-patent document included in the correlation information document database 140 based on the correlation information about one document included in the document database 200. You can also generate an extended document for.

The expansion process by the document expansion unit 120 will be specifically described with reference to FIGS. 4A and 4B. The document extension unit 120 converts the text a included in the document ID “ID_X” of FIG. 3 and the text q included in the document ID “ID_Y” into graph sets Ga and Gq by Bag Of Words, respectively. In the example of FIG. 4A, the text a is "By the way, today's weather is sunny and then cloudy", and the text q is "Tomorrow's weather is sunny and then rainy due to the approaching typhoon." ..

Assuming that the window width is 3, the graph set Ga (by the way, today's weather is) (today's weather is sunny) (weather is sunny, later) (sunny, later, cloudy) (later cloudy, The graph set Gq is (tomorrow, due to the approach of the typhoon) (tomorrow, the weather is due to the approach) (tomorrow, the weather is sunny) (the weather is sunny, later) (sunny, later,). (Rain) (later, rain). The document extension unit 120 compares the graph sets Ga and Gq, and shares a predetermined number of elements (here, 2) of the graph sets Gq (tomorrow, the weather is sunny), (sunny, later, rain). And (later, rain) are added (integrated) to the graph set Ga to generate the extended graph set Ga'. That is, the extended graph set Ga'is (by the way, today's weather is) (today's weather is sunny) (weather is sunny, then) (sunny, later, cloudy) (after, cloudy). (Tomorrow, the weather will be sunny) (sunny, then rain) (later, rain). Then, the document expansion unit 120 stores the expansion graph set Ga'as a search feature amount of the document ID "ID_X" in the expansion document database 150.

The document extension unit 120 can create an extended document as described above for all the documents included in the document database 200 and store the extended document in the extended document database 150. Further, the document extension unit 120 can also create an extended document as described above for a document composed of all non-patent documents included in the correlation information document database 140 and store the extended document in the extended document database 150. As a result, an extended document corresponding to the search content is generated for each text included in each document, and the extended document database 150 as shown in FIG. 4B is generated.

In the above description, the window width is set to 3, but the present invention is not limited to this, and the window width may be set arbitrarily. Further, the number of shared graph elements, which is a criterion for extracting the graph to be integrated, may be arbitrarily set. For the conversion to the graph set, a tree structure graph using a text dependency structure may be used instead of Bag Of Words. Tree structure graphs can be used to extend a document with more relevant graphs by expanding the document according to the similarity of the graph structure between the texts.

When the search condition 300 is input, the search unit 130 performs a search using the graph set stored in the extended document database 150 instead of the document database 200 as a search target. The search for the extended document database 150 can be performed by graphing the search text as a search condition in the same manner as in the document extension unit 120 and determining the degree of matching of the graphs. The search may be determined according to the matching of keywords from each graph element. The search unit 130 acquires one document (that is, the original document before expansion) corresponding to the extended document searched from the extended document database 150 from the document database 200 and outputs it as a search result.

If the searched document is an extended document, the search unit 130 may weaken the weighting as a search result. This makes it possible to ensure the accuracy of the search results.

<Search process>
Next, the search process by the document retrieval device 100 will be described with reference to the flow chart of FIG.

First, the correlation information storage unit 110 extracts document information correlated with one document included in the document set from the correlation information storage unit. (Step S11). The document expansion unit 120 generates an extended document in which one document is expanded by using a similar document which is another document whose correlation information with respect to one document is equal to or higher than a predetermined value among a plurality of other documents (step). S12). If a document having correlation information of a predetermined value or more is a non-patent document, the document expansion unit 120 expands one document by using the document data of the non-patent document stored in the correlation information storage unit. Generate the extended document (step S12'). The search unit 130 accepts the search condition (step S13) and searches for the extended document corresponding to the search condition (step S14). The search unit 130 outputs the document corresponding to the searched extended document as the search result 310 (step S15).

As described above, according to one embodiment of the present invention, the probability that a keyword co-occurring with a keyword in one document or similar text is stored in the extended document database 150 and matches the search text which is a search condition. Will be higher. Further, since the extended document is generated by the correlation information storage unit 110 in advance with a document that correlates with one document, it is possible to secure a search rate higher than the accuracy of the existing method. Further, since the text to be expanded according to the graph similarity is selected from the similar documents, it is possible to expand the search range without deteriorating the search accuracy (matching rate).

It should be noted that the target for generating the extended document by the document expanding unit 120 may be only a predetermined constituent unit in one document. For example, when one document is a patent document, only the claims may be expanded, and the specification in the same document and the specification in a similar document may be targeted as the target text for acquiring the expanded graph. As a result, it becomes possible to carry out a search expanded with synonyms while narrowing down the search target.

The document retrieval device 100 according to the embodiment of the present invention described above may be applied to the intellectual property creation support system. Figure 6 shows a configuration example of the intellectual property creation support system. As shown in FIG. 6, the intellectual property creation support system 500 includes a document retrieval device 100, a document database (DB) 200, and a communication terminal 400 (400A to 400D) connected to each other via a network NET. include. The number of communication terminals 400 is not limited to the one shown in the figure.

The communication terminals 400A to 400D are communication terminals of users who use the intellectual property creation support service provided by the intellectual property creation support system 500. In the intellectual property creation support system 500, the document retrieval device 100 can search for prior art documents. In FIG. 6, the communication terminal 400 shows a notebook computer or a desktop computer, but the communication terminal 400 may be of any type as long as the document search service can be used via the network NET. The communication terminal 400 is, for example, a smartphone, a mobile phone (feature phone), a handheld computer device (for example, a PDA (Personal Digital Assistant), etc.), a wearable terminal (for example, a glasses-type device, a clock-type device, a head-mounted display (HMD:). A head-mounted display, etc.), another type of computer, or a communication platform may be included. The communication terminal 400 receives an input operation from a user and transmits a search condition to a document search device 100 via a network NET. ..

The document database 200 can be, for example, a database of the Japan Patent Office. The JPO database may include one or more offices. In addition, since it is possible to cover about 90% of the world's patents by including the databases of the five agencies of the United States, Europe, Japan, China, and South Korea, in order to improve the search accuracy of the prior art documents, these It is good to include the database of 5 agencies. The database is not limited to the above-mentioned one, and may be information existing on the Internet.

The network NET may include a wireless network or a wired network. Specifically, the network NET includes wireless LAN (WLAN), wide area network (WAN), ISDNs (integrated service digital networks), wireless LANs, LTE (long term evolution), LTE-Advanced, and so on. 4th generation (4G), 5th generation (5G), CDMA (code division multiple access) and the like. The network NET is not limited to these examples, and is, for example, a public switched telephone network (PSTN), Bluetooth (Bluetooth (registered trademark)), an optical line, an ADSL (Asymmetric Digital Subscriber LINE) line, and a satellite. It may be a communication network or the like. Further, the network NET may be a combination of these.

In the intellectual property creation support system 500, the user can at least perform a prior art literature search, an invalidity search, and a clearance search. The user transmits the search condition to the document retrieval device 100 via the communication terminal 400. For example, in the case of a prior art document search, the user sends, for example, an invention sentence describing the invention as information about his / her invention as a search condition. The search unit 130 extracts keywords representing the invention included in the invention text as search keywords. Then, a group of similar patent sentences is extracted from the document database 200 based on the extracted search keywords. At this time, the document search device 100 expands the document database 200 as described above, and the search unit 130 searches for the extended document database 150.

Although the present invention has been described with reference to the drawings and examples, it should be noted that those skilled in the art can easily make various modifications and modifications based on the present disclosure. Therefore, it should be noted that these modifications and modifications are included in the scope of the present invention. For example, the functions included in each component, each step, etc. can be rearranged so as not to be logically inconsistent, and a plurality of components, steps, etc. can be combined or divided into one. Is. Further, the configurations shown in the above embodiments may be appropriately combined. For example, each component described as being included in the document retrieval device 100 may be physically distributed by a plurality of computers, or may be realized as one computer.

The program of each embodiment of the present disclosure may be provided stored in a computer-readable storage medium. The storage medium can store the program in a "non-temporary tangible medium". Programs include, for example, software programs and computer programs.

When appropriate, the storage medium may be one or more semiconductor-based or other integrated circuits (ICs) (eg, field programmable gate arrays (FPGAs), application-specific ICs (ASICs), etc.), hard disks. Disk drive (HDD), hybrid hard drive (HHD), optical disk, optical disk drive (ODD), optical magnetic disk, optical magnetic drive, floppy diskette, floppy disk drive (FDD), magnetic tape, solid drive (SSD), RAM drive, secure digital card or drive, any other suitable storage medium, or any suitable combination of two or more thereof can be included. The storage medium may be volatile, non-volatile, or a combination of volatile and non-volatile, where appropriate.

Further, the program of the present disclosure may be provided to the information processing apparatus via an arbitrary transmission medium (communication network, broadcast wave, etc.) capable of transmitting the program.

Each embodiment of the present disclosure may also be realized in the form of a data signal embedded in a carrier wave, in which the program is embodied by electronic transmission.

The program of the present disclosure is implemented using, for example, a script language such as JavaScript (registered trademark), Python, C language, Go language, Swift, Kotlin, Java (registered trademark), and the like.

100 Document search device 101 Processor 102 Memory 103 Storage 110 Correlation information storage unit 120 Document expansion unit 130 Search unit 140 Correlation information document database 150 Extended document database 200 Document database 300 Search conditions 310 Search results 500 Intellectual property creation support system 400 Communication terminal 400A-400D communication terminal

Claims (9)

A document search device equipped with a search unit that searches for documents corresponding to the received search content.
A correlation information storage unit that stores correlation information about one document included in a document set and a document that is correlated with the one document.
Based on the correlation information, a document that is correlated with the one document is used to generate one extended document that is an extension of the one document, and the correlation for a plurality of documents included in the document set. A document extension unit that generates an extended document about the document based on the information.
The search unit searches for the extended document corresponding to the search content from the extended document for a plurality of documents, and outputs one document corresponding to the searched extended document as a search result .
The document extension section has a graph structure of the one document.
By adding part or all of the graph structure of a correlated document to the one document above
A document retrieval device that generates an extended document of the above-mentioned one document. The search unit receives the search text as the search content and receives the search text.
The extended document whose matching degree in graph units of the search text is equal to or more than a predetermined value is searched, and the search is performed.
The document search device according to claim 1, wherein one document corresponding to the searched extended document is output as a search result. The search unit receives the search text as the search content and receives the search text.
Search for the extended document that contains the keywords contained in the search text,
The document search device according to claim 1 or 2, wherein one document corresponding to the searched extended document is output as a search result. The document set is a set of patent documents having bibliographic information.
One of claims 1 to 3, wherein the correlation information storage unit stores correlation information between the one document and the plurality of other documents based on the bibliographic information. The document retrieval device described in paragraph 1. The document set is a set of patent documents having bibliographic information.
The claim is characterized in that the correlation information storage unit stores correlation information between the one document and the plurality of other documents based on the information regarding the examination or trial of the one document. The document retrieval device according to any one of Items 1 to 4. The document set is a set of patent documents having bibliographic information and non-patent documents having bibliographic information.
In the correlation information storage unit, information on correlated non-patent documents is further accumulated based on the bibliographic information of the one document.
The document retrieval apparatus according to any one of claims 1 to 5, further comprising a correlation information document database in which document data of the non-patent document is accumulated. The document extension unit is a graph structure of the one document.
For a tree structure graph structure using the dependency structure of text,
By adding a part or all of the tree structure graph structure that is the graph structure of the document that correlates with the one document and uses the dependency structure of the text.
The document retrieval apparatus according to any one of claims 1 to 6, wherein an extended document of the above-mentioned document is generated. This is a document search method that searches for documents that correspond to the received search content.
The computer
A step to extract a document that is correlated with one document included in the document set from the correlation information storage unit, and
A document expansion step that uses the correlated document to generate an extended document that is an extension of the one document.
And run
The document expansion step creates the extended document for a plurality of documents based on the correlation of the correlation information storage unit.
The computer, from the extensive documentation for said plurality of documents, searching the extended document corresponding to the search activity, one of the document corresponding to the retrieved extended document, perform a search step of outputting a retrieval result ,
The document expansion step has a graph structure of the one document.
By adding part or all of the graph structure of a correlated document to the one document above
Performing a generation step to generate an extended document of the one document,
Document search method. A document search program that searches for documents that correspond to the received search content.
On the computer
A function to extract correlation information about documents that correlate with one document included in the document set from the correlation information storage unit, and
Using the correlation information, a document extension function for generating an extended document, which is an extension of the one document, is realized.
The document extension function creates the extended documents for a plurality of documents based on the correlation information stored in the correlation information storage unit.
A document search function is realized in which the computer searches for an extended document corresponding to the search content from the extended documents for the plurality of documents and outputs one document corresponding to the searched extended document as a search result. Let me
The document extension function is applied to the graph structure of the one document.
By adding part or all of the graph structure of a correlated document to the one document above
A document retrieval program that further realizes a generation function for generating an extended document of the above-mentioned one document.

Images