JP6084311B1 - Document information providing device - Google Patents

Abstract

[PROBLEMS] When a new technical document (other company's document) of another company is released, the company's technical document (in-house document) related to the other company's document and the technical field is automatically extracted, and the other company's document is sent to the person in charge of the company's document. In order to provide this information, it is necessary to appropriately set the extraction condition in advance, and there is a possibility that the operation man-hours for appropriately setting the extraction condition may increase. By extracting an in-house document composed of a vocabulary similar to a set of words (ie, vocabulary) included in another company's document as a document related to the other company's document and the technical field, the extraction condition is not set in advance. , It is possible to distribute the information of other company documents to the person in charge of the company document. [Selection] Figure 3

Description

  The present invention relates to a document information providing apparatus for providing information on a new technical document to a person in charge of an existing technical document related to the technical field when the new technical document is acquired.

  Information distribution promotion that distributes intellectual property right information to a predetermined user when newly published intellectual property right information (for example, published patent gazette and international publication gazette) meets preset extraction conditions A system (hereinafter also referred to as “conventional system”) is described in Patent Document 1. The extraction condition is set by registering words included in the intellectual property right information, classification information (for example, IPC and FI), and the like in the conventional system. According to the conventional system, developers and researchers in various fields can automatically receive the technology they are developing and the technical information related to the researched field.

Japanese Patent Laying-Open No. 2015-99583

  However, even if the new intellectual property information matches the extraction condition (that is, even if it contains a preset word or the classification information matches), the intellectual property information is not displayed in the conventional system. May not belong to the relevant field for users.

  If the user of the conventional system sets the extraction conditions in detail (for example, if more words are added to the extraction condition), it is possible to avoid the intellectual property right information that the conventional system does not belong to the related field. There is sex. On the other hand, if the extraction conditions are set too finely, there is a risk that the intellectual property right information in the related field will not match the extraction conditions. In other words, the man-hours for appropriately adjusting the extraction conditions may be excessive.

  Accordingly, one of the objects of the present invention is that when a new technical document (for example, intellectual property rights information such as a published patent gazette, an international published gazette, and an academic paper) is acquired, a technical field related to the technical document. It is to provide a document information providing apparatus capable of realizing information related to the technical document to a person in charge by a simple operation operation.

To achieve the above object, a document information providing apparatus according to the present invention (hereinafter also referred to as “present apparatus”) includes a document acquisition unit, a document storage unit, a related document extraction unit, and an information provision unit. I have.
The document acquisition unit acquires a classified technical document.
The document storage unit stores information on a plurality of reference technical documents each assigned a person in charge.

When the classified technical document is acquired by the document acquisition unit, the related document extracting unit extracts one or a plurality of the reference technical documents configured by vocabulary similar to the vocabulary configuring the classified technical document. A related document extraction process for extracting as a related document is executed.
The information providing unit provides information related to the classified technical document to each person in charge of the related document extracted by the related document extracting unit.

  For example, a corporation that uses the device of the present invention registers the specification, etc. relating to its own patent application in the device of the present invention as a reference technical document, and the person who invented each of the patent applications is responsible for the reference technical document. You can register as a person. In addition, the apparatus of the present invention can set the apparatus of the present invention so that the document acquisition unit acquires a published patent publication related to a patent application filed by another company (a person other than the company) as a classified technical document. .

  In this case, when the patent gazette of another company is newly published, the device of the present invention provides information on the patent gazette to the person in charge of the company's patent application composed of vocabulary similar to the vocabulary constituting the patent gazette. Can be provided.

  There is a high possibility that a technical document including a patent publication contains technical terms specific to the technical field to which the document belongs. Therefore, a set (namely, vocabulary) of words (specifically, nouns) included in each of two technical documents in which technical fields are related to each other is often similar to each other. Therefore, there is a high possibility that the patent publications of other companies provided to the person in charge by the device of the present invention belong to the technical field that the person in charge is in charge of.

  In other words, if the reference technical document is registered in the device of the present invention together with the person in charge, when the classified technical document is acquired, the device of the present invention is related to the technical document related to the classified technical document. Information about the classified technical document can be provided. At this time, it is not necessary to set extraction conditions in the apparatus of the present invention in advance. Therefore, according to the apparatus of the present invention, it is possible to provide information related to the classified technical document to the person in charge of the reference technical document related to the classified technical document and the technical field by a simple operation operation.

In one aspect of the device of the present invention,
The related document extraction unit includes:
(1) calculating a “first vocabulary distribution” determined based on the appearance frequency of each word included in a technical document in the technical document for each of the classified technical document and the plurality of reference technical documents;
(2) The “first vocabulary distribution distance”, which becomes smaller as the first vocabulary distributions of the two technical documents are similar to each other, is set to an arbitrary one of the classified technical document and the plurality of reference technical documents. Calculated for the combination of two technical documents
(3) generating a predetermined number of “neighboring document groups” by aggregating each of the technical documents included in the combination of the technical documents in which the first vocabulary distribution distance is smaller than a predetermined value;
(4) extracting the reference technical document included in the neighboring document group including the classified technical document as the related document;
It is preferable that the related document extraction process is executed.

  Technical terms specific to the technical field to which the technical document belongs are likely to be used repeatedly in the technical document. Therefore, there is a high possibility that the technical fields of the technical documents included in the same neighboring document group are related to each other. Therefore, according to this aspect, it is possible to accurately extract the reference technical document related to the classified technical document and the technical field.

In another embodiment of the device of the present invention,
The device of the present invention
Each of the plurality of reference technical documents is classified into a plurality of “reference technical document groups” that are collections of technical documents related to each other in the technical field, and a “reference technical document classification unit” is stored in the document storage unit. ,
The reference technical document classification unit includes:
When adding the reference technical document to the document storage unit,
It is determined that the reference technical document to be added belongs to the reference technical document group that is a set of the reference technical documents that are most closely related to the technical field of the reference technical document to be added among the reference technical document groups. Execute "Reference technical document addition process"
The related document extraction unit includes:
Among the reference technical document groups, the reference technical documents included in the reference technical document group that is a set of the reference technical documents most closely related to the classified technical document and the technical field are extracted as the related documents. The related document extraction process is executed by
It is preferable to be configured as described above.

  The configuration of the reference technical document group is defined when the reference technical document addition process is executed. Therefore, for example, even when a plurality of classified technical documents are acquired (that is, when the related document extraction process is executed a plurality of times), a combination of reference technical documents included in each of the reference technical document groups ( That is, the structure of the reference technical document group does not change. Therefore, according to this aspect, it is possible to provide information to a person in charge regarding a plurality of classified technical documents without changing the configuration of the reference technical document group.

  For example, when the operator of the device of the present invention registers a reference technical document in the document storage unit with the start of use of the device of the present invention, if a reference technical document group is set, then the reference technical document to be added is Belongs to any existing reference technical document group. Therefore, according to this aspect, it is possible to add a reference technical document after fixing each technical field of the reference technical document group.

The aspect of the device of the present invention described above is
The reference technical document classification unit
(1) A “second vocabulary distribution” determined based on the appearance frequency of each word included in a technical document in each technical document is added to each of the added reference technical document and the other reference technical documents. Calculate
(2) The “second vocabulary distribution distance”, which becomes a smaller value as the second vocabulary distributions of the two technical documents are similar to each other, is set to the added reference technical document and the other reference technical documents. Calculate for a combination of any two of these technical documents,
(3) generating a predetermined number of the reference technical document groups by aggregating each of the technical documents included in the technical document combination in which the second vocabulary distribution distance is smaller than a predetermined value;
(4) It is determined that the added reference technical document belongs to the reference technical document group including the added reference technical document among the generated reference technical document groups.
Accordingly, the reference technical document adding process may be executed.

  In this aspect, the reference technical document belongs to a reference technical document group having a vocabulary distribution similar to that of the added reference technical document. Therefore, a set of reference technical documents and other reference technical documents to be added can be accurately divided into a plurality of reference technical document groups configured by reference technical documents having technical fields related to each other.

  In addition, each time a reference technical document is added, a reference technical document group is newly generated. Therefore, according to this aspect, even when many reference technical documents are added, a state in which the reference technical document groups are configured by the reference technical documents related to the technical field is maintained.

Alternatively, the aspect of the device of the present invention described above is
The related document extraction unit
(1) The same reference is made to the “first word content rate”, which is the ratio of the number of the reference technical documents including a certain word in the reference technical document group to the number of the reference technical documents included in the reference technical document group. Calculate for each technical document group and for each word,
(2) The value becomes higher as the classified technical document includes more words having the higher first word content rate, and the classified technical document includes fewer words having the lower first word content rate. The “first document relevance”, which becomes a larger value as
(3) extracting the reference technical document included in the reference technical document group having the highest value of the first document relevance among the reference technical document groups as the related document;
Accordingly, the related document extraction process may be executed.

  The first document relevance value is a large value if the classified technical document includes “words included in many reference technical documents included in a certain reference technical document group”. On the other hand, the first document relevance level is a small value if the classified technical document includes “a word that is not included in many of the reference technical documents included in the reference technical document group”. In addition, the first document relevance value is a large value if the classified technical document does not include “a word not included in many of the reference technical documents included in the reference technical document group”.

  In other words, the first document relevance is determined based on “words not included” in addition to “words included” in each of the classified technical document and the reference technical document. Therefore, according to this aspect, it is possible to accurately extract a reference technical document group including a reference technical document related to a classified technical document and a technical field.

In addition, the aspect of the device of the present invention described above is
The reference technical document classification unit
(1) The same reference is made to the “second word content ratio” which is a ratio of the number of the reference technical documents including a word in the reference technical document group to the number of the reference technical documents included in the reference technical document group. Calculate for each technical document group and for each word,
(2) The reference technical document to be added has a higher value as the number of words having a higher second word content rate is increased, and the reference technical document to be added has a lower word content rate. Calculating “second document relevance” for each of the reference technical document groups, which becomes a larger value as the content of the reference technical document group is reduced.
(3) It is determined that the added reference technical document belongs to the reference technical document group having the highest value of the second document relevance among the reference technical document groups.
Accordingly, the reference technical document adding process may be executed.

  Similar to the first document relevance level, the second document relevance level is determined based on “words not included” in addition to “words included” in each of the added reference technical document and other reference technical documents. Is done. For this reason, in this aspect, it is possible to accurately extract a reference technical document group including reference technical documents to be added and other reference technical documents related to the technical field. As a result, according to this aspect, even when many reference technical documents are added, a state in which the reference technical document groups are constituted by the reference technical documents related to the technical field is maintained.

Furthermore, the aspect of the device of the present invention described above is
The reference technical document classification unit
(1) The number of persons in charge is included in both the set of persons in charge of each of the reference technical documents included in the reference technical document group and the group of persons in charge of the added reference technical document. The “Responsibility of person in charge”, which becomes a larger value, is calculated for each of the reference technical document groups
(2) It is determined that the added reference technical document belongs to the reference technical document group having the highest relevance level among the reference technical document groups.
Accordingly, the reference technical document adding process may be executed.

  In general, each person in charge is familiar with a specific technical field and performs his / her duties. Therefore, multiple reference technical documents related to the same person may have technical fields related to each other. high. If multiple representatives are involved in a reference technical document, other reference technical documents related to a "collection of multiple representatives" whose members are similar to "collection of those multiple representatives" The reference technical document and the technical field are likely to be related to each other.

  Therefore, it is considered that the technical fields of the two reference technical documents are related to each other as the degree of relevance of the person in charge regarding the two reference technical documents is larger. For this reason, in this aspect, it is possible to accurately extract a reference technical document group including reference technical documents to be added and other reference technical documents related to the technical field. As a result, according to this aspect, even when many reference technical documents are added, a state in which the reference technical document groups are constituted by the reference technical documents related to the technical field is maintained.

It is the schematic showing the structure of the document information provision apparatus (1st apparatus) which concerns on 1st Embodiment of this invention. It is the table | surface showing the vocabulary distribution of the reference technical document and the classified technical document which a 1st apparatus calculates. It is the graph which represented the vocabulary distribution of the reference technical document and the classified technical document in the two-dimensional orthogonal coordinate system. It is a flowchart showing the reference technical document addition process which a 1st apparatus performs. It is a flowchart showing the related document extraction process which a 1st apparatus performs. It is a flowchart showing the classified technical document notification process which a 1st apparatus performs. It is the table | surface showing the word content rate computed for every reference technical document group and every word. It is an example of the word content rate and vocabulary distribution shown in order to demonstrate the calculation method of a document relevance degree. It is the table | surface showing the document relevance calculated for every reference technical document group. It is a flowchart showing the reference technical document addition process which the document information provision apparatus (2nd apparatus) which concerns on 2nd Embodiment of this invention performs. It is a flowchart showing the related document extraction process which a 2nd apparatus performs. It is a flowchart showing the reference technical document registration process which the document information provision apparatus (3rd apparatus) which concerns on 3rd Embodiment of this invention performs. It is a flowchart showing the reference technical document addition process which a 3rd apparatus performs. It is the table | surface showing the vocabulary set calculated for every reference technical document, and a person in charge relevance degree. It is the table | surface showing the reference technical document rearranged in the order of the person-in-charge relevance degree. It is a flowchart showing the reference technical document registration process which the document information provision apparatus (4th apparatus) which concerns on 4th Embodiment of this invention performs. It is a flowchart showing the reference technical document addition process which a 4th apparatus performs. It is the table | surface which contrasted the classification method of each reference technical document from the 1st apparatus to the 4th apparatus, and the classification method of a to-be-classified technical document.

The document information providing apparatus according to each embodiment of the present invention will be described below with reference to the drawings.
<First Embodiment>
FIG. 1 shows a schematic configuration of a document information providing apparatus 11 (hereinafter also referred to as “first apparatus”) according to the first embodiment of the present invention. The document information providing apparatus 11 is a general-purpose computer, and includes a CPU 21, a RAM 22, a hard disk drive (HDD) 23, a network interface 24, and an operation interface 25.

  The CPU 21 reads data, performs numerical calculations, outputs calculation results, and the like by sequentially executing a predetermined program. The RAM 22 temporarily stores data. The HDD 23 stores programs executed by the CPU 21, a database (DB), and the like.

  The network interface 24 can communicate with an external document publishing server 42 and an electronic mail server 43 via a known network 41 (for example, the Internet). The document publishing server 42 is a Web server (http server) that publishes a new published patent gazette as needed and maintains the gazette in a state where it can be acquired via the network 41. For example, the document publication server 42 is a “publication publication site” published by the Patent Office and a Web server for a patent information service provided by a private company.

  The e-mail server 43 transmits an e-mail to a desired destination (e-mail address) in response to a request (for example, an SMTP request) from the document information providing apparatus 11 (specifically, the network interface 24). Can do.

  The operation interface 25 can communicate with an input device 26 and an output device 27 connected to the document information providing device 11. The input device 26 includes a keyboard and a mouse, and the output device 27 includes a display device. In addition, the operation interface 25 includes a USB port (not shown), and can read data from a USB memory connected to the USB port and write data to the USB memory. The operator of the document information providing device 11 operates the document information providing device 11 using the input device 26 and acquires the operation result via the output device 27.

(Database)
The DB stored in the HDD 23 includes a reference technical document DB 31 and a classified technical document DB 32. The reference technical document DB 31 includes information (including a specification attached to the application) relating to a patent application filed by a specific corporation (hereinafter also referred to as “own company”), and is hereinafter also referred to as “reference technical document”. Is registered). In the reference technical document DB 31, the application number is used as the document ID (identifier) of the reference technical document.

  In the reference technical document DB 31, information on one or more persons in charge for each of the reference technical documents is registered. As a general rule, the person in charge is the inventor of the invention described in the reference technical document (that is, the description relating to the patent application, etc.) and the assistant who is involved in the formation process of the invention. If the person in charge has changed, the successor is registered in the reference technical document DB 31 as the person in charge of the reference technical document. The information regarding the person in charge registered in the reference technical document DB 31 includes the e-mail address of each person in charge.

  The classified technical document DB 32 relates to a published patent publication (hereinafter also referred to as “classified technical document”) relating to a patent application filed by a person other than the company (hereinafter also referred to as “other company”). Information is registered. In the classified technical document DB 32, the application number is used as the document ID (identifier) of the classified technical document.

(Outline of processing)
The processing executed by the CPU 21 (hereinafter also simply referred to as “CPU”) includes reference technical document addition processing, related document extraction processing, and classified technical document notification processing. When the operator operates the input device 26 to register a new reference technical document (hereinafter also referred to as “new reference technical document”) in the document information providing apparatus 11, the CPU executes a reference technical document addition process. The contents of the new reference technical document and the document ID are added to the reference technical document DB 31.

  When a new classified technical document (hereinafter also referred to as “new classified technical document”) is published by the document publication server 42, the CPU executes a related document extraction process and executes a known method (for example, A new classified technical document is acquired (downloaded) by an http request. Further, when executing the related document extraction process, the CPU extracts a set of reference technical documents related to the new classified technical document and the technical field. Next, the CPU registers the content and document ID of the new classified technical document and the document ID of the reference technical document related to the new classified technical document and the technical field in the classified technical document DB 32.

  When information related to the new classified technical document is registered in the classified technical document DB 32, the CPU executes a classified technical document notification process, and takes charge of the new classified technical document and the reference technical document related to the technical field. An e-mail is sent to the person in charge to notify the person in charge that the new classified technical document has been released and the contents of the new classified technical document.

(Reference technical document addition process)
At the time of executing the reference technical document addition process, the CPU executes a morphological analysis process (so-called text mining process) on the new reference technical document by a well-known method to thereby execute a word (specifically, a word included in the new reference technical document). Is a noun) and the number of occurrences of each word. A set of combinations of a word included in a document and the number of occurrences of the word is also referred to as a “vocabulary distribution”.

  An example of the vocabulary distribution of each reference technical document registered in the reference technical document DB 31 is shown in the table of FIG. In the table of FIG. 2, the documents from document 1 to document (N-1) are reference technical documents that have already been registered in the reference technical document DB 31 at the time of registration in the reference technical document DB 31. This is a new reference technical document that is newly added to the reference technical document DB 31. Each of the reference technical documents from document 1 to document N includes some (or all in some cases) of the M types of words from word 1 to word M.

(Related document extraction process-explanation using simplified example)
When executing the related document extraction process, the CPU extracts a vocabulary distribution related to the new classified technical document. The calculated vocabulary distribution of the new classified technical document is shown in FIG. Further, the CPU collects a set of documents composed of the N reference technical documents registered in the new classified technical document and the reference technical document DB 31 by collecting the documents related to the technical field by a predetermined value K. Classify into groups. Therefore, the CPU sets “the vocabulary distribution distance D representing the degree to which the technical fields of the two documents are related” regarding the combination of any two documents included in the set of (N + 1) documents. Is calculated based on each vocabulary distribution. The value of the vocabulary distribution distance D decreases as the technical fields of the two documents are more closely related.

  A method for calculating the vocabulary distribution distance D will be described with reference to a simplified example. FIG. 3 shows an example of the vocabulary distribution for each document when it is assumed that there are only two words, ie, word 1 (for example, “catalyst”) and word 2 (for example, “temperature”). Shown in the figure.

3, the number of occurrences of the word 1 is represented by x ₁ axis, the number of occurrences of the word 2 is represented by x ₂ axis. The x ₁ axis and _{x 2} axes orthogonal to each other. In the distribution diagram of FIG. 3, each of the reference technical documents is represented by a white circle (“◯”), and one new classified technical document is represented by a black square (“■”). The black square representing the new classified technical document is also referred to as point P0.

  The vocabulary distribution distance D relating to a combination of two documents is calculated as a distance between two points (white circles or black squares) corresponding to each of the documents on the distribution map. If the distance between the two points is short, the vocabulary distribution distance D becomes small. Specifically, the difference between the number of occurrences of word 1 in one of the two documents and the number of appearances of word 1 in the other is small, and the difference between the number of occurrences of word 2 in one and the number of occurrences of word 2 in the other Is smaller, the vocabulary distribution distance D is smaller. In other words, if the vocabulary distributions of the two documents are similar, the vocabulary distribution distance D becomes small. If the vocabulary distribution distance D is small, the two documents corresponding to the vocabulary distribution distance D are considered to be related to each other in the technical field.

For example, the reference technical document represented by the point P1 in FIG. 3 contains the word 1 containing x ₁ 1 (ie x ₁ = x ₁ 1) and the word 2 containing x ₂ 1 (ie x ₂ = x ₂ 1). In this case, the coordinates of the point P1 can be represented as a point P1 (x ₁ 1, x ₂ 1). Similarly, the reference technical document represented by the point P2 in FIG. 3 is represented as a point P2 (x ₁ 2, x ₂ 2), and the reference technical document represented by the point P3 in FIG. ₁ 3, x ₂ 3).

If the distance between the points P1 and P2 in the distribution diagram of FIG. 3 (ie, the vocabulary distribution distance D) is expressed as D (1,2), the vocabulary distribution distance D (1,2) can be expressed by the following formula (1 ). Similarly, the vocabulary distribution distance D (2, 3) between the points P2 and P3 is calculated by the following equation (2).

  When the vocabulary distribution distance D is calculated for a combination of two arbitrary documents, a group composed of documents (reference technical document and new classified technical document) whose vocabulary distribution distance D is equal to or less than the first threshold Dth1 is generated. The For example, the vocabulary distribution distance D (1,2) and the vocabulary distribution distance D (2,3) are both equal to or less than the first threshold value Dth1. Therefore, the documents represented by the points P1, P2, and P3 are included in the same group (also referred to as “group 1” for convenience).

  On the other hand, the vocabulary distribution distance D (3,4) regarding the combination of the two documents represented by the points P3 and P4 is larger than the first threshold value Dth1. Therefore, the document represented by the point P4 is not included in the group 1. The group including the document represented by the point P4 is also referred to as “group 2” for convenience.

  In this example, each of the documents is divided into five groups (that is, group 1 to group 5). In FIG. 3, points (white circles or black squares) representing documents included in each of the five groups are surrounded by closed curves C1 to C5, respectively. The point P0 representing the new classified technical document is included in the group 1. Therefore, it can be understood that each of the reference technical documents included in the group 1 is related to the new classified technical document and the technical field.

  In this example, the first threshold value Dth1 is set so that the number of generated groups is “5”. For example, when documents are grouped based on a second threshold value Dth2 (that is, Dth1> Dth2) shorter than the first threshold value Dth1, group 4 (a group constituted by documents surrounded by a closed curve C4) is represented by a broken line C4a and a broken line. Divided into two groups surrounded by each of C4b.

  That is, the smaller the threshold value of the vocabulary distribution distance D is, the more document groups are generated. In other words, as the threshold value of the vocabulary distribution distance D becomes smaller, a document group composed of documents with more closely related technical fields is generated.

(Related Document Extraction Process-Actual Process)
In the example shown in FIG. 3, the distribution map is shown by a two-dimensional orthogonal coordinate system (that is, each of the reference technical document and the new classified technical document includes words other than word 1 and word 2). The actual document contains a large number of words. Therefore, the actual vocabulary distribution distance D is obtained by plotting points representing each document in accordance with the vocabulary distribution in a multi-dimensional orthogonal coordinate system having a number of axes orthogonal to each other. It is calculated based on the distance between them.

If the reference technical document and the new classified technical document are composed of part (or all in some cases) of M types of words, the point Pa (x ₁ a, x ₂ a in the M-dimensional multi-dimensional orthogonal coordinate system is used. ,..., X _M a) and the vocabulary distribution distance D (Pa, Pb) between the points Pb (x ₁ b, x ₂ b,..., X _M b) are calculated by the following equation (3).

  When the vocabulary distribution distance D is calculated for a combination of two arbitrary documents, the document is obtained by aggregating the reference technical documents and the new classified technical documents included in the range where the vocabulary distribution distance D is equal to or less than the distance threshold Dth. Are grouped. In addition, the distance threshold Dth is adjusted so that the number of groups to be generated is a predetermined value K. Of the set of documents divided into K sets, each of the reference technical documents (also referred to as “related documents” for convenience) included in the set to which the new classified technical document belongs is referred to as the new classified technical document. And technical fields are related.

(Classified technical document notification process)
When the related document extraction process is completed, the CPU executes a classified technical document notification process. That is, an e-mail is transmitted to each person in charge of the related document extracted by the related document extraction process.

(Specific operation-reference technical document addition processing)
Next, a specific operation of the CPU will be described with reference to the flowcharts of FIGS. As described above, when the operator of the document information providing apparatus 11 operates the operation interface 25 to register a new reference technical document, the CPU executes a reference technical document addition process. Specifically, the CPU starts the process from step 400 in FIG. 4 and proceeds to step 405.

  In step 405, the CPU makes a new reference technical document stored in the USB memory connected to the USB port of the operation interface 25 (specifically, an HTML file such as a patent application application and a specification attached to the application). And the contents are stored in the RAM 22.

  Next, the CPU proceeds to step 410 to extract a vocabulary distribution from the new reference technical document. Further, the CPU proceeds to step 415 to register the contents of the new reference technical document and the vocabulary distribution of the new reference technical document in the reference technical document DB 31. Thereafter, the CPU proceeds to step 495 to end the present routine.

(Specific operation-related document extraction processing)
On the other hand, the CPU executes the related document extraction process represented by the flowchart in FIG. 5 every time a predetermined time elapses. Therefore, when the appropriate timing is reached, the CPU starts processing from step 500 in FIG. 5 and proceeds to step 505.

  In step 505, the CPU determines whether or not the new classified technical document is published on the document publication server 42. Specifically, the CPU confirms whether or not there is a classified technical document published after the previous execution of this routine on the document publication server 42. If the new classified technical document is published, the CPU makes a “Yes” determination at step 505 to proceed to step 510 to acquire (download) the new classified technical document from the document publication server 42.

  Further, the CPU proceeds to step 515 to extract a vocabulary distribution from the acquired new classified technical document. Thereafter, the CPU proceeds to step 520 to calculate the vocabulary distribution distance D based on the above equation (3) for a combination of two arbitrary documents among the new classified technical document and the reference technical document.

  Next, the CPU proceeds to step 525 to determine whether or not the reference technical document exists in a range where the vocabulary distribution distance D from the new classified technical document is shorter than the predetermined reference distance Dr. The reference distance Dr is set so that if the vocabulary distribution distance D is greater than the reference distance Dr, the two technical documents related to the vocabulary distribution distance D are determined to be not related to each other. Yes.

  If there is a reference technical document in which the vocabulary distribution distance D to the new classified technical document is shorter than the reference distance Dr, the CPU makes a “Yes” determination at step 525 to proceed to step 530 to determine the distance. The threshold value Dth is set as the reference distance Dr. Further, the CPU proceeds to step 530 to group documents (reference technical document and new classified technical document) by collecting documents whose vocabulary distribution distance D is equal to or smaller than the distance threshold Dth.

  Next, the CPU proceeds to step 540 to determine whether or not the number of generated document groups is equal to or less than a predetermined value K. If the number of document groups is larger than the predetermined value K, the CPU makes a “No” determination at step 540 to proceed to step 545 to decrease the value of the distance threshold Dth by the predetermined value Dc. Next, the CPU proceeds to step 535.

  The CPU repeatedly executes the processing from step 535 to step 545 until the number of generated document groups becomes equal to or less than the predetermined value K. If the number of generated document groups is equal to or less than the predetermined value K, the CPU makes a “Yes” determination at step 540 to proceed to step 550, where the reference included in the same document group as the new classified technical document. It is determined that the technical document is a document related to the new classified technical document and the technical field (that is, a related document). Further, the CPU proceeds to step 555 to register the contents of the new classified technical document and the document ID of the related document in the classified technical document DB 32.

  Next, the CPU proceeds to step 560 to determine whether or not all new classified technical documents published on the document publication server 42 have been processed. If an unprocessed new classified technical document remains, the CPU makes a “No” determination at step 560 to proceed to step 510.

  On the other hand, if the processing described above has been completed for all new classified technical documents, the CPU makes a “Yes” determination at step 560 to proceed to step 595 to end the present routine.

  In addition, if there is no reference technical document in which the vocabulary distribution distance D to the new classified technical document is shorter than the reference distance Dr, the CPU makes a “No” determination at step 525 to proceed to step 565. It is determined that the related document does not exist. Further, the CPU proceeds to step 555 to register the contents of the new classified technical document and the absence of the related document in the classified technical document DB 32.

  If the new classified technical document is not published on the document publication server 42, the CPU makes a “No” determination at step 505 to directly proceed to step 595.

(Specific operation-classified technical document notification process)
When the related document extraction process described above ends (that is, after proceeding to step 595 in FIG. 5), the CPU executes a classified technical document notification process. Specifically, the CPU starts the process from step 600 in FIG. 6 and proceeds to step 605. In step 605, the CPU determines whether or not there is a classified technical document (unprocessed document) that has not been processed by this routine after being registered in the classified technical document DB 32.

  If an unprocessed document exists, the CPU makes a “Yes” determination at step 605 to proceed to step 610 to determine whether or not a related document is registered for the unprocessed document. If the related document is registered, the CPU makes a “Yes” determination at step 610 to proceed to step 615 to extract each person in charge of the related document from the reference technical document DB 31.

  Next, the CPU proceeds to step 620 and requests the email server 43 to send an email to the person in charge. Further, the CPU proceeds to step 625 to determine whether or not processing has been completed for all unprocessed documents. If processing has been completed for all unprocessed documents, the CPU makes a “Yes” determination at step 625 to proceed to step 695 to end the present routine.

  On the other hand, if an unprocessed document remains, the CPU makes a “No” determination at step 625 to proceed to step 610. If no related document is registered for the unprocessed document, the CPU makes a “No” determination at step 610 to proceed directly to step 625. In addition, if there is no unprocessed document at the start of execution of this routine, the CPU makes a “No” determination at step 605 to directly proceed to step 695.

As described above, the first device (document information distribution device 11)
A document acquisition unit (such as a network interface 24) that acquires classified technical documents;
A document storage unit (HDD 23) for storing information on a plurality of reference technical documents each assigned a person in charge;
When the classified technical document is acquired by the document acquisition unit, a related document that extracts one or a plurality of the reference technical documents configured by vocabulary similar to the vocabulary configuring the classified technical document as a related document A related document extraction unit (CPU 21 or the like) that executes extraction processing;
An information providing unit (such as a network interface 24) that provides information on the classified technical document to each person in charge of the related document extracted by the related document extracting unit;
It has.

In addition, the related document extraction unit
A first vocabulary distribution determined based on the appearance frequency of each word included in a technical document is calculated for each of the classified technical document and the plurality of reference technical documents (FIGS. 2 and 4). Step 410 and step 515 of FIG. 5),
The first vocabulary distribution distance (vocabulary distribution distance D), which becomes smaller as the first vocabulary distributions of the two technical documents are similar to each other, is determined from among the classified technical document and the plurality of reference technical documents. Calculate for any two technical document combinations (step 520 in FIG. 5);
A predetermined number (K) of neighboring document groups are generated by aggregating each of the technical documents included in the technical document combination in which the first vocabulary distribution distance is smaller than a predetermined value (distance threshold Dth) (FIG. 5). Steps 535 to 545),
Extracting the reference technical document included in the neighboring document group including the classified technical document as the related document (step 550 in FIG. 5);
Thus, the related document extraction process is executed.

  According to the first apparatus, when a classified technical document is newly published by the document publication server 42, the person in charge of the reference technical document related to the classified technical document and the technical field can obtain information on the classified technical document. Can be received automatically. In this case, the person in charge and the operator do not need to register the extraction conditions (for example, the search word used for word search and the classification information of the published patent publication) in advance in the first device. Therefore, according to the first device, it is possible to realize information distribution of new classified technical documents related to the technical field in charge of each person in charge by a simple operation operation.

  In addition, since the first device accurately extracts the related document based on the vocabulary distribution of the reference technical document and the classified technical document, the information about the classified technical document that is not related to the technical field that the person in charge is in charge of is distributed. Can be avoided.

Second Embodiment
Next, the document information distribution apparatus 12 (hereinafter also referred to as “second apparatus”) according to the second embodiment of the present invention will be described. The first device divides a set of reference technical documents and new classified technical documents into groups of K predetermined values at the time of executing the related document extraction process, and each of the reference technical documents belonging to the same group as the new classified technical documents. Was determined to be a related document. On the other hand, the second device divides the reference technical document into K groups when the reference technical document addition process is executed, and which reference technical document group is the new classified technical document when the related document extraction process is executed. The related document is extracted by determining whether it belongs to. Hereinafter, this difference will be mainly described.

(Reference technical document addition process)
The CPU 21 of the document information distribution apparatus 12 (hereinafter also simply referred to as “CPU”) includes a reference including a new reference technical document by grouping the documents based on the vocabulary distribution distance D described above when the reference technical document addition process is executed. The set of technical documents is divided into K “sets of reference technical documents in which technical fields are related to each other” (hereinafter also referred to as “reference technical document group”).

  Further, the CPU, for each reference technical document group, the ratio of “the number of reference technical documents including a certain word in the reference technical document group” to “the number of reference technical documents included in the reference technical document group”. Is calculated for each word. That is, the word content rate θ is a value determined for each reference technical document group and for each word. The word content rate θ is referred to in a related document extraction process described later.

  The CPU does not calculate the word content rate θ for words that are not included in any of the reference technical documents included in the reference technical document group. Accordingly, the word content θ is a value greater than 0 and less than or equal to 1 (that is, 0 <θ ≦ 1).

  For example, when a certain reference technical document group includes 10 reference technical documents, and four of the reference technical documents include the word 1, the word content rate θ of the word 1 in the reference technical document group Is 4/10 = 0.4.

  An example of the reference technical document group and the word content rate θ is shown in FIG. FIG. 7 shows a reference technical document group a and a reference technical document group b. Group a includes Na reference technical documents from document a1 to document aNa, and group b includes Nb reference technical documents from document b1 to document bNb.

  Each of the reference technical documents included in the group a includes some (in some cases, all) of the Ma words from the word a1 to the word aMa. Each of the reference technical documents included in the group b includes a part (or all in some cases) of the Mb words from the word b1 to the word bMb.

  In FIG. 7, whether or not each of the words is included in each of the reference technical documents is represented by characters “1” and “0”. For example, since the word a1 is included in the document a1 included in the group a, “1” is described in the corresponding part of the table of FIG. On the other hand, since the word a4 is not included in the document a1, “0” is described in the corresponding portion of the table of FIG. A set of words included in a document represented by “1” and “0” in the table of FIG. 7 is also referred to as a “vocabulary set”.

  In addition, the table of FIG. 7 shows the word content rate θa of group a and the word content rate θb of group b for each word. For example, the ratio of the number of reference technical documents including the word a1 to the number of reference technical documents (Na) included in the group a (namely, the word content θa) is “0.15”.

(Related document extraction process-explanation using simplified example)
When executing the related document extraction process, the CPU sets the document relevance level P for each of the reference technical document groups based on the “word content rate θ of the reference technical document group” and the “vocabulary set of new classified technical documents”. calculate. The document relevance P is a value representing the degree to which the technical field of the new classified technical document and the technical field of the reference technical document included in a certain reference technical document group are related, and is more closely related. The bigger you are. Therefore, the CPU determines that the new classified technical document belongs to the reference technical document group having the maximum document relevance value P.

  A method of calculating the document relevance level P will be described with reference to a simplified example. An example of a word content rate θ related to a reference technical document group and a vocabulary set of a new classified technical document when the number of words included in the reference technical document constituting a reference technical document group is four is shown in FIG. 8 (A) and (B) respectively.

  8A and 8B, whether or not the new classified technical document includes each of the four words (that is, the vocabulary set) is represented by the characters “1” and “0”. . In general, in the example of FIG. 8A, the new classified technical document includes a word having a high word content rate θ and does not include a word having a low word content rate θ. On the other hand, in the example of FIG. 8B, the new classified technical document includes words with a low word content rate θ and does not include words with a high word content rate θ.

  For each of the four words, the document relevance level P is obtained by obtaining “a word content rate θ of a certain word and a coefficient R determined based on whether or not the word is included in the new classified technical document”. Is calculated by multiplying each coefficient R. The coefficient R is equal to the word content rate θ of a word if the word is included in the new classified technical document. If the word is not included in the new classified technical document, the coefficient R is 1 to A value obtained by subtracting the word content rate θ (that is, (1−θ)).

  For example, in FIG. 8A, the word content rate θ of the word 1 is “0.7”, and the new classified technical document includes the word 1, so the coefficient R (1) of the word 1 is “0”. .7 ". On the other hand, since the word content rate θ of the word 3 is “0.3” and the new classified technical document does not include the word 1, the coefficient R (3) of the word 3 is 1−0.3 = 0.7. It is. The document relevance P is calculated by multiplying these four coefficients (that is, P = R (1) × R (2) × R (3) × R (4)).

  According to the example of FIG. 8A, the document relevance P is “0.3136”. On the other hand, according to the example of FIG. 8B, the document relevance P is “0.0036”. As can be understood from the examples of FIGS. 8A and 8B, the document relevance P becomes higher as the new classified technical document includes a word having a higher word content rate θ, and includes the word. The higher the value, the lower the rate θ is in the new classified technical document.

ここで、θ（ｉ）は参照技術文書グループに含まれるｉ番目の単語に関する
単語含有率θであり、
Ｅ（ｉ）はｉ番目の単語が新被分類技術文書に含まれているか否かを表し、
ｉ番目の単語が新被分類技術文書に含まれていれば「１」となり、
ｉ番目の単語が新被分類技術文書に含まれていなければ「０」となる。 (Related Document Extraction Process-Actual Process)
In the example of FIGS. 8A and 8B, the reference technical document group includes only four words, but the actual reference technical document group includes a large number of words. If M words are included in the reference technical document group, the document relevance P is calculated by the following equation (4).

Here, θ (i) relates to the i-th word included in the reference technical document group.
The word content θ,
E (i) represents whether or not the i-th word is included in the new classified technical document,
If the i-th word is included in the new classified technical document, it will be “1”.
If the i-th word is not included in the new classified technical document, “0” is set.

  The CPU calculates the document relevance P for each of the reference technical document groups. An example of the document relevance level P for each reference technical document group is shown in the table of FIG. As can be understood from FIG. 9, the document relevance level P of group b has the largest value among the K reference technical document groups. Therefore, the new classified technical document has a technical field related to each of the reference technical documents included in the group b compared to the reference technical documents included in the other reference technical document groups (that is, to the group b). It is understood that the reference technical document included is a related document).

(Specific operation-reference technical document addition processing)
Next, a specific operation of the CPU during the execution of the reference technical document addition process and the related document extraction process will be described with reference to the flowcharts of FIGS. Steps shown in the flowchart of FIG. 10 where the same processing as the steps shown in the flowchart of FIG. 4 is executed are denoted by the same step symbols as in FIG. In addition, the steps shown in the flowcharts of FIGS. 10 and 11 in which the same processes as the steps shown in the flowchart of FIG. 5 are executed are denoted by the same step symbols as in FIG. . Regarding the classified technical document notification process, the operation of the second device is the same as the operation of the first device, and thus the description thereof is omitted.

  When executing the reference technical document addition process, the CPU starts the process from step 1000 in FIG. 10 and proceeds to step 1015 through the processes in steps 405 and 410. That is, at 1015, the CPU calculates the vocabulary distribution distance D based on the above equation (3) for any two combinations of reference technical documents including the new reference technical document.

  Next, the CPU proceeds to step 1035 through the processing of step 530, groups the reference technical documents including the new reference technical document by aggregating documents whose vocabulary distribution distance D is equal to or smaller than the distance threshold Dth, and proceeds to step 540. move on.

  When the CPU determines “Yes” in step 540 (ie, when the reference technical document group to which the new reference technical document belongs is determined), the CPU proceeds to step 1045 to extract a vocabulary set of the new reference technical document. Next, the CPU proceeds to step 1050 to calculate the word content rate θ for each word included in each reference technical document group.

  Further, the CPU proceeds to step 1055 to register the contents of the new reference technical document, the reference technical document group to which the new reference technical document belongs, the word content rate θ, and the like in the reference technical document DB 31. Thereafter, the CPU proceeds to step 1095 to end the present routine. If the CPU determines “No” in step 540, it proceeds to step 1035 through step 545.

(Specific operation-related document extraction processing)
When executing the related document extraction process, the CPU starts the process from step 1100 in FIG. 11 and proceeds to step 505. When the CPU determines “Yes” in step 505, it proceeds to step 1115 through the process of step 510 and extracts a vocabulary set of the new classified technical document. Next, the CPU proceeds to step 1120 to calculate the document relevance level P for each of the reference technical document groups of the new classified technical document based on the above equation (4).

  Further, the CPU proceeds to step 1125 to determine whether or not the calculated maximum value of the document relevance level P is greater than or equal to a predetermined relevance level threshold Pth. If the document relevance level P for a certain reference technical document group is smaller than the relevance level threshold Pth, the relevance threshold Pth is related to the technical field of the new classified technical document “reference technical document included in the reference technical document group”. It is set to a value that is determined not to be.

  If the calculated maximum value of the document relevance level P is greater than or equal to the relevance level threshold value Pth, the CPU makes a “Yes” determination at step 1125 to proceed to step 1130, where the value of the document relevance level P is the maximum. Extract technical document groups.

  Next, the CPU proceeds to step 560 through steps 550 and 555. When the CPU makes a “Yes” determination at step 560, the CPU proceeds to step 1195 to end the present routine.

  On the other hand, if the calculated maximum value of the document relevance P is smaller than the relevance threshold Pth, the CPU makes a “No” determination at step 1125 to proceed to step 560 through the processing of step 565 and step 555. When the CPU makes a “No” determination at step 505, the CPU proceeds directly to step 1195. In addition, when the CPU makes a “No” determination at step 560, the CPU proceeds to step 510.

As described above, the second device (document information distribution device 12)
Each of the plurality of reference technical documents is classified into a plurality of reference technical document groups that are collections of technical documents related to each other in the technical field, and a reference technical document classification unit (CPU 21 or the like) is stored in the document storage unit. ,
The reference technical document classification unit includes:
When adding the reference technical document to the document storage unit, the reference technical document is a set of the reference technical documents that are most closely related to the added technical technical field in the reference technical document group. A reference technical document addition process (FIG. 10) for determining that a reference technical document to be added to the document group belongs is executed.
The related document extraction unit includes:
Among the reference technical document groups, the reference technical documents included in the reference technical document group that is a set of the reference technical documents most closely related to the classified technical document and the technical field are extracted as the related documents. The related document extraction process (FIG. 11) is executed by
It is configured as follows.

In addition, the second device
The reference technical document classification unit
Calculating a second vocabulary distribution determined based on the appearance frequency of each word included in a technical document for each of the added reference technical document and the other reference technical documents;
The second vocabulary distribution distance (vocabulary distribution distance D), which becomes a smaller value as the second vocabulary distributions of the two technical documents are similar to each other, is added to the reference technical document and the other reference technical documents. Is calculated for a combination of two arbitrary technical documents (step 1015 in FIG. 10),
A predetermined number of the reference technical document groups are generated by aggregating each of the technical documents included in the technical document combination in which the second vocabulary distribution distance is smaller than a predetermined value (Step 1035, Step 540, Step in FIG. 10). 545 and step 1045),
It is determined that the added reference technical document belongs to the same reference technical document group including the added reference technical document among the generated reference technical document groups (step 1055 in FIG. 10).
Accordingly, the reference technical document adding process is executed.

Furthermore, the second device is
The related document extraction unit
The first word content rate (word content rate θ), which is the ratio of the number of the reference technical documents including a word in the reference technical document group to the number of the reference technical documents included in the reference technical document group, is the same. For each reference technical document group and for each word (step 1055 in FIG. 10),
The higher the value of the classified technical document, the higher the first word content rate, the larger the value, and the lower the classified technical document, the lower the first word content rate. A first document relevance level (document relevance level P) having a large value is calculated for each of the reference technical document groups (step 1120 in FIG. 11),
The reference technical document included in the reference technical document group having the highest value of the first document relevance among the reference technical document groups is extracted as the related document (step 1130 in FIG. 11).
Thus, the related document extraction process is executed.

  According to the second apparatus, even if the related document extraction process is executed a plurality of times, the combination of the reference technical documents included in each of the reference technical document groups (that is, the configuration of the reference technical document group) does not change. Therefore, it is avoided that the configuration of the reference technical document group changes depending on the contents of the classified technical document, and the combination of persons in charge who receive information on the classified technical document is changed. Specifically, the occurrence of an event that information related to one classified technical document is distributed to the person in charge A and the person in charge B and information related to another classified technical document is distributed to the person in charge A and the person in charge C is avoided. Is done.

  In addition, since the configuration of the reference technical document group is determined based on the vocabulary distribution distance D, according to the second device, the reference technical document group is obtained after accurately collecting the reference technical documents related to each other in the technical field. Can be generated. Further, since the reference technical document group is newly generated every time the reference technical document adding process is executed, the second technical apparatus relates the technical fields to each other even if many reference technical documents are added. The state constituted by the reference technical document is maintained.

  In addition, the second device extracts a related document based on the document relevance level P when executing the related document extraction process. In other words, the second device determines the related document based on “words not included” in addition to “words included” in each of the classified technical document and the reference technical document. Therefore, according to the second apparatus, it is possible to accurately extract a reference technical document group including a reference technical document related to a classified technical document and a technical field.

<Third Embodiment>
Next, a document information distribution apparatus 13 (hereinafter also referred to as “third apparatus”) according to a third embodiment of the present invention will be described. The second device divides the reference technical document into a predetermined number of reference technical document groups based on the vocabulary distribution distance D when the reference technical document addition process is executed. That is, the second apparatus regroups the reference technical documents every time a new reference technical document is registered. On the other hand, when the third device adds a new reference technical document, to which group of the reference technical document groups the new reference technical document belongs (ie, which group and the technical field are related). Is determined based on the document relevance P. Hereinafter, this difference will be mainly described.

(Reference technical document registration process)
The operator of the document information distribution apparatus 13 classifies the reference technical documents into predetermined value K groups (reference technical document groups) in advance according to the technical field, and registers them in the document information distribution apparatus 13. At this time, the CPU 21 of the document information distribution apparatus 13 (hereinafter also simply referred to as “CPU”) executes a reference technical document registration process, and each of the reference technical documents is referred to for each reference technical document group. Register with.

(Reference technical document addition process)
Further, when the operator adds a reference technical document, the CPU executes a reference technical document addition process. When executing the reference technical document addition process, the CPU calculates a document relevance level P representing the relevance level with the new reference technical document for each of the reference technical document groups, and determines the reference technical document group having the maximum document relevance level P. It is determined that the new reference technical document belongs.

(Specific operation-reference technical document registration process)
Next, a specific operation of the CPU at the time of executing the reference technical document registration process and the reference technical document addition process will be described with reference to the flowcharts of FIGS. The steps shown in the flowchart of FIG. 13 and the same steps as those shown in the flowcharts of FIGS. 4, 10, and 11 are the same as those in FIGS. 4, 10, and 11, respectively. The step code is attached.

  The operation of the third device is the same as the operation of the second device for the related document extraction process, and the operation of the third device is the same as the operation of the first device for the classified technical document notification process. The description of the process is omitted.

  When executing the reference technical document registration process, the CPU starts the process from step 1200 in FIG. 12, proceeds to step 1205, and is pre-grouped (that is, classified into the reference technical document group) stored in the USB memory. A plurality of reference technical documents are read and the contents are stored in the RAM 22. Next, the CPU proceeds to step 1210 and extracts each vocabulary set of the read reference technical document.

  Further, the CPU proceeds to step 1215 to calculate the word content rate θ for each word included in each reference technical document group. Thereafter, the CPU proceeds to step 1220 to register each content of the reference technical document and each word content rate θ of the reference technical document group in the reference technical document DB 31. At this time, the CPU deletes information related to the reference technical document already registered in the reference technical document DB 31 from the reference technical document DB 31. Next, the CPU proceeds to step 1295 to end the present routine.

(Specific operation-reference technical document addition processing)
When executing the reference technical document addition process, the CPU starts the process from step 1300 in FIG. 13, and proceeds to step 1350 through the processes in step 405, step 1120, step 1130, and step 1045.

  In step 1350, the CPU adds a new reference technical document to the reference technical document group (that is, the group to which the new reference technical document belongs) extracted in step 1130 (ie, updated reference technology). The word content rate θ of each word included in the document group is calculated. Next, the CPU proceeds to step 1355 to register information related to the new reference technical document and information related to the updated reference technical document group in the reference technical document DB 31. Further, the CPU proceeds to step 1395 to end this routine.

As described above, the third device (document information distribution device 13)
The reference technical document classification unit
The second word content rate (word content rate θ), which is the ratio of the number of the reference technical documents containing a word in the reference technical document group to the number of the reference technical documents included in the reference technical document group, is the same. For each reference technical document group and for each word (step 1215 in FIG. 12),
The added reference technical document has a higher value as it contains more words with a higher second word content, and the added reference technical document has fewer words with a lower second word content. A second document relevance level (document relevance level P) that increases as it is included is calculated for each of the reference technical document groups (step 1120 in FIG. 13).
It is determined that the added reference technical document belongs to the reference technical document group having the highest value of the second document relevance among the reference technical document groups (step 1130 in FIG. 13).
Accordingly, the reference technical document adding process is executed.

  According to the third apparatus, a combination of reference technical documents included in each of the reference technical document groups (that is, the configuration of the reference technical document group) is determined based on the document relevance level P. In other words, the structure of the reference technical document group is determined based on “words not included” in addition to “words included” in each of the reference technical documents. Therefore, according to the third device, even when many reference technical documents are added, the state in which the reference technical document groups are constituted by the reference technical documents related to the technical field is maintained.

<Modification of Third Embodiment>
Next, a document information distribution apparatus (hereinafter also referred to as “this modification apparatus”) according to a modification of the third embodiment of the present invention will be described. The third device registers a plurality of reference technical documents classified in advance by the reference technical document registration process in the reference technical document DB 31, and further adds a new reference technical document to the reference technical document DB 31 by the reference technical document addition process. To which reference technical document group the new reference technical document belongs. On the other hand, the document information distribution apparatus according to this modification does not execute the reference technical document addition process.

  For this reason, when a new reference technical document is registered (that is, when a reference technical document is added to the reference technical document DB 31), the operator of the deforming apparatus executes a reference technical document registration process again. In other words, the operator needs to determine which reference technical document group the new reference technical document belongs to when registering the new reference technical document in the deformation apparatus.

<Fourth embodiment>
Next, a document information distribution apparatus 14 (hereinafter also referred to as “fourth apparatus”) according to a fourth embodiment of the present invention will be described. The third device determines which reference technical document group the new reference technical document belongs to based on the document relevance P when the new reference technical document is further added to the set of reference technical document groups classified in advance. It was. On the other hand, when the fourth device further adds a new reference technical document to the set of reference technical document groups classified in advance, the fourth device is based on the person-in-charge relevance level Q determined according to each person in charge of the reference technical document. Determine which reference technical document group the new reference technical document belongs to. Therefore, this difference will be mainly described below.

(Reference technical document addition process)
The CPU 21 of the document information distribution apparatus 14 (hereinafter also simply referred to as “CPU”) calculates the person-in-charge relevance level Q for each reference technical document when executing the reference technical document addition process. The person-in-charge relevance level Q is similar to the set of persons in charge of the new reference technical document and the set of persons in charge of the reference technical documents (other reference technical documents) already registered in the reference technical document DB 31. It is a value that represents the degree to which

  Specifically, the value of person-in-charge relevance Q is a person in charge included in the intersection (ie, logical product) of a set of persons in charge of the new reference technical document and a group of persons in charge of other reference technical documents. Equal to the number of people. Therefore, the value of the person-in-charge relevance Q increases as the two sets of persons in charge are similar to each other and the number of persons in charge included in these sets increases. It is considered that the technical fields of the two reference technical documents are related to each other as the person-in-charge relevance level Q regarding the two reference technical documents is larger.

  When executing the reference technical document addition process, the CPU calculates the person-in-charge relevance level Q for each combination of the new reference technical document and each of the other reference technical documents. An example of the calculated person-in-charge association degree Q is shown in FIG. In FIG. 14, the persons in charge of the N reference technical documents from the reference technical document 1 to the reference technical document N already registered in the reference technical document DB 31 and the new reference technical document are “1” and “ It is represented by the character “0”.

  For example, since the person in charge of the reference technical document 1 includes the person in charge 1, “1” is described in the corresponding part of FIG. On the other hand, since the person in charge of the reference technical document 1 does not include the person in charge 4, “0” is described in the corresponding part of FIG. 14. A set of individual persons in charge included in a person in charge of a reference technical document is also referred to as a “person in charge group”.

  In addition, the table in FIG. 14 corresponds to each of the reference technical document groups (any one of K groups of group a, group b,..., Group K) to which each of the reference technical documents belongs, and each of the reference technical documents. The value of the person-in-charge relevance Q is shown.

  An example of a table obtained by rearranging the reference technical documents in descending order of the person-in-charge relevance level Q is shown in FIG. The CPU extracts L reference technical documents (six person-in-charge documents) in the descending order of the person-in-charge relevance level Q, and the most similar person-in-charge documents in the reference technical document group. Is determined to be a group to which the new reference technical document belongs.

  In this example, among the six related reference technical documents, there are three reference technical documents included in group a, two reference technical documents included in group b, and reference technologies included in group c. There is one document. Since the number of related documents included in group a is the largest compared to reference technical documents included in other groups, the CPU determines that the new reference technical document is included in group a.

(Specific operation-reference technical document registration process)
Next, a specific operation of the CPU during execution of the reference technical document registration process and the reference technical document addition process will be described with reference to the flowcharts of FIGS. 16 and 17. Steps shown in the flowcharts of FIGS. 16 and 17 in which the same processing as the steps shown in the flowcharts of FIGS. 4, 10, 12, and 13 are executed are shown in FIGS. The same step codes as those in FIGS. 12 and 13 are given.

  As for the related document extraction process, the operation of the fourth device is the same as the operation of the second device, and for the classified technical document notification processing, the operation of the fourth device is the same as the operation of the first device. The description of the process is omitted.

  When executing the reference technical document registration process, the CPU starts the process from step 1600 of FIG. 16, and proceeds to step 1620 through the processes of steps 1205 to 1215. In step 1620, the CPU extracts each person-in-charge set of the reference technical document.

  Next, the CPU proceeds to step 1625 to register each content of the reference technical document, each word content θ of the reference technical document group, each person in charge of the reference technical document, and the like in the reference technical document DB 31. At this time, information regarding the reference technical document already registered in the reference technical document DB 31 is deleted from the reference technical document DB 31. Next, the CPU proceeds to step 1695 to end the present routine.

(Specific operation-reference technical document addition processing)
When executing the reference technical document addition process, the CPU starts the process from step 1700 of FIG. 17, proceeds to step 1710 through the process of step 405, and extracts a person-in-charge set from the read new reference technical document.

  Next, the CPU proceeds to step 1715 to calculate a person-in-charge relevance level Q between the new reference technical document and each of the reference technical documents already registered in the reference technical document DB 31. Further, the CPU proceeds to step 1720 to extract the top L reference technical documents (that is, person-in-charge similar documents) having a large person-in-charge relevance level Q.

  Next, the CPU proceeds to step 1725 to extract a reference technical document group including the most related reference technical documents from among the reference technical document groups to which each of the L person-in-charge similar documents belongs, and a new reference technical document is obtained. It is determined that it belongs to the reference technical document group. Further, the CPU proceeds to step 1795 through steps 1045, 1350, and 1355, and ends the present routine.

As described above, the fourth device (document information providing device 14)
The reference technical document classification unit
The larger the number of persons in charge included in both the set of persons in charge of the reference technical document included in the reference technical document group and the group of persons in charge of the added reference technical document, the larger the value. The person-in-charge relevance level (Q) is calculated for each of the reference technical document groups (step 1715 in FIG. 16),
It is determined that the added reference technical document belongs to the reference technical document group having the highest value of the person-in-charge relevance among the reference technical document groups (step 1720 in FIG. 16).
Accordingly, the reference technical document adding process is executed.

  According to the fourth device, a combination of reference technical documents included in each of the reference technical document groups (that is, the configuration of the reference technical document group) is determined based on the person-in-charge relationship degree Q. Therefore, “a plurality of reference technical documents in which technical fields related to the same person in charge are related to each other” is highly likely to be included in the same reference technical document group. Therefore, according to the fourth device, even when many reference technical documents are added, the state in which the reference technical document groups are configured by the reference technical documents related to the technical field is maintained.

  FIG. 18 shows a table comparing the classification method of the reference technical document and the classification method of the classified technical document from the first device to the fourth device described above.

  Although the embodiment of the document information providing apparatus according to the present invention has been described above, the present invention is not limited to the above embodiment, and various modifications can be made without departing from the object of the present invention. For example, the reference technical document in each embodiment is an in-house patent application (specifications attached to the application), and the classified technical document is a patent application (published patent publication) of another company. However, the reference technical document and the classified technical document may be technical documents other than the patent application.

  For example, the reference technical document and the classified technical document may be technical papers. Alternatively, the utility model registration application may be included in the reference technical document and the classified technical document. Further, the reference technical document and the classified technical document may be a news manuscript published on a Web server accessible via the network 41. In addition, the reference technical document and the classified technical document may be a combination of these documents (ie, patent application, utility model registration application, academic paper, and news manuscript).

  In addition, in the first embodiment and the second embodiment, the vocabulary distribution is calculated based on the number of occurrences of a certain word in a certain technical document. However, the vocabulary distribution may be calculated based on the ratio of the number of appearances of the word to the length of the technical document (text length).

  In addition, in each embodiment, the vocabulary distribution and the vocabulary set are calculated for each word appearing in a certain technical document. However, the vocabulary distribution and / or vocabulary set may be calculated after aggregating words appearing in the technical document based on a predetermined synonym dictionary. For example, when “motor”, “motor”, and “motor” are treated as synonyms and, as a result, “motor” or “motor” is included in a technical document, the lexical distribution and / or vocabulary set is calculated. In, these words may be treated as “motors”.

  In addition, in the first embodiment, if there is no reference technical document in a range where the vocabulary distribution distance D is shorter than the reference distance Dr, the first device determines that there is no related document. However, this process may be omitted. Similarly, in the second embodiment, if the document relevance level P is not equal to or higher than the relevance level threshold Pth, the second device determines that there is no related document. However, this process may be omitted.

  In addition, in each embodiment, a person in charge is assigned to each reference technical document. However, instead of the person in charge, the department in charge of the company may be assigned to each of the reference technical documents. For example, a department to which the inventor of the invention described in a reference technical document (that is, a specification related to a patent application, etc.) may be assigned to the reference technical document. In this case, the fourth device may calculate the person-in-charge relevance level Q based on the department in charge of each reference technical document.

DESCRIPTION OF SYMBOLS 11 ... Document information delivery apparatus, 21 ... CPU, 22 ... RAM, 23 ... HDD, 24 ... Network interface, 25 ... Operation interface, 26 ... Input device, 27 ... Output device, 41 ... Network, 42 ... Document publication server, 43 ... an email server.

Claims (4)

A document acquisition unit for acquiring classified technical documents;
Each of the plurality of reference technical documents to which a person in charge is assigned is classified into one of a plurality of reference technical document groups that are collections of documents related to each other in the technical field, and information about the reference technical documents is stored. A document storage unit;
When the classified technical document is acquired by the document acquisition unit, the reference technical document that is a group constituted by vocabularies similar to the vocabulary constituting the classified technical document among the plurality of reference technical document groups A related document extracting unit that executes a related document extracting process of extracting each of the reference technical documents included in the group as a related document;
With
The related document extraction unit includes:
For each reference technical document group, a first word content ratio that is a ratio of the number of the reference technical documents including a word in the reference technical document group to the number of the reference technical documents included in the reference technical document group. And for each word,
The higher the value of the classified technical document, the higher the first word content rate, the larger the value, and the lower the classified technical document, the lower the first word content rate. Calculating a first document relevance value having a large value for each of the reference technical document groups;
Extracting the reference technical document included in the reference technical document group having the highest value of the first document relevance among the reference technical document groups as the related document;
Accordingly, the document information providing apparatus configured to execute the related document extraction process. A document acquisition unit for acquiring classified technical documents;
Each of the plurality of reference technical documents to which a person in charge is assigned is classified into one of a plurality of reference technical document groups that are collections of documents related to each other in the technical field, and information about the reference technical documents is stored. A document storage unit;
When the classified technical document is acquired by the document acquisition unit, the reference technical document that is a group constituted by vocabularies similar to the vocabulary constituting the classified technical document among the plurality of reference technical document groups A related document extracting unit that executes a related document extracting process of extracting each of the reference technical documents included in the group as a related document;
With
The document storage unit
When the reference technical document is added, the reference technical document is added to the reference technical document group, which is a set of the reference technical documents most technically related to the added reference technical document. A document information providing apparatus for executing a reference technical document addition process for classifying reference technical documents,
The document storage unit
A second word content rate, which is a ratio of the number of the reference technical documents including a word in the reference technical document group to the number of the reference technical documents included in the reference technical document group, is set for each reference technical document group. And for each word,
The added reference technical document has a higher value as it contains more words with a higher second word content, and the added reference technical document has fewer words with a lower second word content. Calculating a second document relevance level that increases as it is included for each of the reference technical document groups;
Classifying the added reference technical document into the reference technical document group having the highest value of the second document relevance among the reference technical document groups;
The reference technical document addition process is executed by
Document information providing apparatus configured as described above. A document acquisition unit for acquiring classified technical documents;
Each of the plurality of reference technical documents to which a person in charge is assigned is classified into one of a plurality of reference technical document groups that are collections of documents related to each other in the technical field, and information about the reference technical documents is stored. A document storage unit;
When the classified technical document is acquired by the document acquisition unit, the reference technical document that is a group constituted by vocabularies similar to the vocabulary constituting the classified technical document among the plurality of reference technical document groups A related document extracting unit that executes a related document extracting process of extracting each of the reference technical documents included in the group as a related document;
With
The document storage unit
When the reference technical document is added, the reference technical document is added to the reference technical document group, which is a set of the reference technical documents most technically related to the added reference technical document. A document information providing apparatus for executing a reference technical document addition process for classifying reference technical documents,
The document storage unit
The larger the number of persons in charge included in both the set of persons in charge of the reference technical document included in the reference technical document group and the group of persons in charge of the added reference technical document, the larger the value. For each reference technical document group,
Classifying the reference technical document to be added to the reference technical document group having the highest relevance level among the reference technical document groups.
The reference technical document addition process is executed by
Document information providing apparatus configured as described above. The document information providing apparatus according to any one of claims 1 to 3 ,
A document information providing apparatus including an information providing unit that provides information on the classified technical document to each person in charge of the related document extracted by the related document extracting unit.

Images