JP2012141928A - Computer, document presentation method and document presentation program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a computer for presenting a document to which a title as a valid determination material of the selection of information is applied.SOLUTION: In a computer connected to a document database in which a document is stored, a processor is configured to extract an inter-word causal relation included in the document stored in the document database, create causal network data expressing the extracted inter-word causal relation with a network structure, extract a word having the causal relation with an interesting word as a word which interests a user from the causal network data, create a title of the document based on a first word and a second word if defining the interesting word as the first word and the extracted word as the second word, and create data for presenting the document to which the created title is applied via an output device to the user.

Description

  The present invention relates to a computer, a document presentation method, and a document presentation program, and to a technique for presenting a document to a computer user.

  A technique for supporting work by presenting a document (auxiliary material) that assists the work to a worker who works using the computer on the computer is known (see Patent Document 1). . Japanese Patent Application Laid-Open No. H10-228561 discloses a technique that supports selection of an operator and grasping an outline of information by presenting a document with a headline.

JP 2010-117790 A

  However, the headline given to the document in the technique disclosed in Patent Document 1 does not consider the work status of the worker. For this reason, there is a problem that is not appropriate as a judgment material for selecting information.

  An object of the present invention is to provide a computer that presents a document with a heading that is an effective determination material for selecting information, in consideration of the above-described problems.

  In order to solve the above problems, for example, the configuration described in the claims is adopted.

  The present application includes a plurality of means for solving the above problems. To give an example, a processor that executes a program, a memory that stores a program executed by the processor, and an interface connected to the processor, A computer connected to a document database for storing documents, wherein the processor extracts a causal relationship between words contained in the document stored in the document database, and the extracted words Causal network data representing a causal relationship in a network structure is created, and from the causal network data, an interesting word that is a word that a user is interested in is extracted and a word that is causal is extracted. Word, based on the first word and the second word when the extracted word is the second word Te, create a heading documents, a document the title that was created is assigned, and generating the data for presentation to a user via the output device.

  According to the present invention, it is possible to present a document with a heading that is an effective determination material for selecting information.

  Problems, configurations, and effects other than those described above will be clarified by the following description of embodiments.

It is a flowchart which shows the operation | movement outline | summary of the user of the computer system of embodiment of this invention. It is a figure explaining the structural example of the computer system of embodiment of this invention. It is a figure which shows the hardware structural example of the computer system of embodiment of this invention. It is a figure which shows the example of the personal work status data of embodiment of this invention. It is a figure which shows the example of the personal work log | history data of embodiment of this invention. It is a figure which shows the example of the personal profile data of embodiment of this invention. It is a flowchart which shows the whole process of the information provision server of embodiment of this invention. It is a flowchart which shows the control logic of the skill level determination part of embodiment of this invention. It is a flowchart which shows the control logic of the ambiguity resolution part of embodiment of this invention. It is a flowchart which shows the control logic of the causal extraction part of embodiment of this invention. It is a figure explaining the example of the causal network data produced by the causal extraction part of embodiment of this invention. It is a flowchart which shows the control logic of the interested word extraction part of embodiment of this invention. It is a flowchart which shows the control logic of the headline preparation part of embodiment of this invention. It is a figure which shows an example of the headline produced by the headline preparation part of embodiment of this invention. It is a flowchart which shows the control logic of the heading list display part of embodiment of this invention. It is a figure which shows the example of the display screen by the heading list display part of embodiment of this invention.

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

  In the computer system 1 (see FIG. 2) according to the embodiment of the present invention, a document on past trouble cases and work know-how is given to a worker (hereinafter referred to as a “user”) who uses the client computer 40 at a manufacturing site. By presenting, it is possible to prevent the recurrence of defects and improve work efficiency.

  FIG. 1 is a flowchart showing an outline of the operation of the user of the computer system 1 according to the embodiment of this invention. An outline of user operations will be described with reference to FIG.

  First, the user logs in to the business management system 201 via the business management system client 402 and inputs a user ID (S101). Thereafter, since a list of tasks is presented on the output unit (display or the like) of the client computer 40, the user selects a task to be executed and starts the task (S102). Then, a heading list of information related to the work status of the selected job is displayed on the client computer 40 (S103).

  The user determines whether there is a headline that attracts interest in the displayed heading list (S104). When there is a headline that attracts interest (YES in S104), the user selects this headline, and acquires information useful for business from the information providing server 10 (S105). On the other hand, when there is no headline that attracts interest (NO in S104), the user continues the business (S106).

  Thereafter, when the job is in progress (NO in S107), the process returns to step 103. In this case, the displayed heading list is updated according to the work status of the task at that time. On the other hand, when the business is completed (YES in S107), the user logs out from the business management system 201 (S108).

  Hereinafter, the process of step 103, that is, the heading list display method will be described.

  FIG. 2 is a diagram illustrating a configuration example of the computer system 1 according to the embodiment of this invention.

  In FIG. 2, blocks (skill level determination unit 101 and the like) indicated by solid line rectangles are program modules. A block (user ID 111 or the like) indicated by a parallelogram is data. A block (business management system 201 or the like) indicated by a broken-line rectangle is an independent external system separate from the information providing server 10. A block (business information DB 50 or the like) indicated by a cylinder is a database.

  The computer system 1 includes an information providing server 10, a business management server 20, a search server 30, a client computer 40, a business information DB 50, a personal information DB 60, and a document DB 70.

  The information providing server 10 is a computer device that provides document data 118 with a header to the client computer 40. The information providing server 10 includes a skill level determination unit 101, an interesting word extraction unit 102, an ambiguity resolution unit 103, a causal extraction unit 104, and a headline creation unit 105. Each functional unit will be described in detail later.

  Based on the user ID 111 received from the business management server 20, the information providing server 10 receives the personal work status data 112 and the personal profile from the business information DB 50 and the personal information DB 60 with or without the business management server 20. Data 114 is input.

  The business management server 20 is a computer device including a business management system 201 that provides a service to the business management system client 402 in response to a request from the client computer 40. The business management server 20 receives from the client computer 40 the user ID 411 of the user and data 412 relating to the work status of the business.

  The search server 30 is a computer device that includes a search engine 301 that searches for information. The search server 30 receives the interesting word data 117 from the information providing server 10, performs a search on the document data recorded in the document DB 70 based on the interesting word data 117, and is extracted as a search result. The document data 311 is transmitted to the information providing server 10.

  The client computer 40 is a computer device that includes a heading list display unit 401 and a business management system client 402. The user executes his / her business using the client computer 40.

  The business information DB 50 is a database in which information related to business performed by each user of the client computer 40 (business name, related document list, related word list, etc., details will be described later) is recorded. The personal information DB 60 is a database in which personal profile data 114 of each user of the client computer 40 is recorded. The document DB 70 is a database in which various document data are recorded.

  FIG. 3 is a diagram illustrating a hardware configuration example of the computer system 1 according to the embodiment of this invention. In FIG. 3, the same reference numerals are given to portions that perform the same functions as those in FIG. 2, and redundant descriptions are omitted as appropriate.

  The computer system 1 shown in FIG. 3 includes an information providing server 10, a business management server 20, a search server 30, and a client computer 40 that are connected to each other via a network 80. The network 80 is a global network such as a LAN (Local Area Network), a WAN (Wide Area Network), or the Internet.

  Hereinafter, the hardware configuration of each apparatus will be described.

  The information providing server 10 includes a network interface 11, a CPU (Central Processing Unit) 12, a main storage device 13, a secondary storage device 14, and a bus 15 that interconnects them.

  The network interface 11 is an interface for the information providing server 10 to connect to the network 80. The CPU 12 is an arithmetic processing device that executes a program stored in the main storage device 13. The main storage device 13 is a storage device such as a RAM (Random Access Memory) that stores a program executed by the CPU 12 and data necessary for executing the program. The program is, for example, an OS (Operating System) (not shown). The secondary storage device 14 is a magnetic storage medium such as a hard disk device.

  The hardware configurations of the business management server 20, the search server 30, and the client computer 40 are the same as those of the information providing server 10, and a description thereof is omitted here.

  FIG. 4A is a diagram illustrating an example of the personal work status data 112 according to the embodiment of this invention.

  The personal work status data 112 is data indicating the work status of each user. The personal work status data 112 includes a business ID 112A, a document being created 112B, a document being browsed 112C, and a search engine status 112D.

  The business ID 112A is an identifier that uniquely identifies the business that the user is engaged in. This work ID 112A is assigned in advance to all work in the factory or manufacturing process. The document 112B being created is a document created on the client computer 40 by the user during work, such as a design drawing or a specification. The document 112B being created is acquired by monitoring with resident software or by establishing rules for uploading to the business management server 20 at a certain timing. The browsing document 112C is a document that the user is browsing on the client computer 40 during working hours, such as a specification, a procedure manual, a Web page, and the like. This browsing document 112C is acquired by monitoring with resident software, extracting browser history, or the like. The search engine state 112D is a word or a search option input to the search engine 301. This search engine state 112D is acquired by monitoring with resident software or extracting information from a browser.

  FIG. 4B is a diagram illustrating an example of the personal work history data 113 according to the embodiment of this invention.

  The personal work history data 113 is data indicating a work state history of each user. The personal work history data 113 includes a created document history 113A and a browsing document history 113B.

  The created document history 113A is obtained by adding time information to a document 112B being created with the same job ID in the document 112B being created of the personal work status data 112. The browsing document history 113B is obtained by adding time information to the browsing document 112C having the same job ID among the browsing documents 112C of the personal work state data 112.

  FIG. 4C is a diagram illustrating an example of the personal profile data 114 according to the embodiment of this invention.

  The personal profile data 114 is data indicating the profile of each user. The personal profile data 114 includes responsible business data 114A, affiliation department data 114B, past created document data 114C, and past browse document data 114D.

  The assigned job data 114A is data related to the user's assigned job. The assigned job data 114A is referred to as a job name of the assigned job, a list of documents related to the assigned job (hereinafter referred to as “related documents”) and related coefficients, and a word related to the assigned job (hereinafter referred to as “related word”). ) List and related coefficients, a list of work IDs related to the work in charge (hereinafter referred to as “related work ID”), and a charge period of the work in charge. The related coefficient is calculated by using the appearance frequency of words in the browsed documents and created documents of all users engaged in the business. The assigned business data 114A can be read using the business ID.

  The belonging department data 114B is data relating to the user's belonging department. The belonging department data 114B includes the department name of the belonging department, a list of related business IDs and related coefficients of the belonging department, a list of belonging members of the belonging department, and a belonging period in the belonging department. The past created document data 114C is obtained by adding time information to all past created documents regardless of the business ID. The past browsing document data 114D is obtained by adding time information to all past browsing documents regardless of the business ID.

  FIG. 5 is a flowchart showing the overall processing of the information providing server 10 according to the embodiment of this invention. Here, the overall processing of the information providing server 10 will be described with reference to FIG. 2 as appropriate.

  First, when a user logs in to the business management system 201 using the client computer 40, the information providing server 10 receives the user ID 111 (S501). Thereafter, the information providing server 10 reads each piece of information using the user ID 111 received in step 501 as a key through the processing of steps 502 to 504.

  In step 502, the information providing server 10 receives the business ID of the business in charge of the user from the business management system 201 (S502). Thereafter, in step 503, the causal network data 106 is read (S503). The causal network data 106 is data having a network structure representing the causal relationship of events included in a plurality of documents in the document DB 70. Details will be described later with reference to FIG.

  Thereafter, in step 504, the information providing server 10 reads the personal profile data 114 from the personal information DB 60 (S504). Thereafter, in step 505, the information providing server 10 (skill level determination unit 101) calculates the skill level of the user based on the personal profile data 114 read in step 504 (S505). The processing in step 505 will be described later in detail with reference to FIG.

  Thereafter, in step 506, the information providing server 10 receives the personal work status data 112 at that time from the business management system 201 (S506).

  Thereafter, in step 507, the information providing server 10 updates the personal work history data 113 (S507). Specifically, the creating document 112B and the browsing document 112C included in the personal work status data 112 received in step 506 are added to the personal work history data 113. Thereafter, in step 508, the interested word extracting unit 102 extracts the user's interested words (S508). The processing in step 508 will be described later in detail with reference to FIG.

  Thereafter, in step 509, the information providing server 10 searches the document DB 70 for a document related to the interesting word based on the interesting word extracted in step 508 (S509). Here, the search is performed using the existing search engine 301. A document related to an interesting word is a document having a keyword that has a causal relationship with the interesting word. Thereafter, in step 510, the ambiguity resolution unit 103 resolves the ambiguity of the interested word (S510). The processing of step 510 will be described later in detail with reference to FIG.

  Thereafter, in step 511, the headline creation unit 105 creates a headline to be assigned to the document searched in step 509 (S511). The processing in step 511 will be described later in detail with reference to FIG.

  Thereafter, in step 512, the heading list display unit 401 presents the heading created in step 511 to the user by displaying it on the output unit (display or the like) (S512).

  Thereafter, in step 513, the information providing server 10 accepts user input (S513). The user input here is a user browsing request for the headline presented in step 512. Thereafter, in step 514, the information providing server 10 determines whether there is a document browsing request (S514).

  If there is a document browsing request (YES in S514), the information providing server 10 presents a document corresponding to the request (S515), returns to step 512, and repeats the process. On the other hand, when there is no document browsing request (NO in S514), the information providing server 10 enters a standby state and determines whether or not the user has logged out (S516).

  When logging out (YES in S516), the standby state of the information providing server 10 ends, and the process ends. On the other hand, if the user has not logged out (NO in S516), the information providing server 10 determines whether or not the personal work status data 112 has been changed (updated) (S517). If the personal work status data 112 has changed (YES in S517), the process returns to step 506 and the process is repeated. If the personal work status data 112 has not changed (NO in S517), the process returns to step 516 to repeat the process.

  Hereinafter, the processing flow of the skill level determination unit 101, the interested word extraction unit 102, the ambiguity resolution unit 103, the causal extraction unit 104, the headline creation unit 105, and the headline list display unit 401 will be described.

  FIG. 6 is a flowchart illustrating the control logic of the skill level determination unit 101 according to the embodiment of this invention. Here, the processing of step 505 in FIG. 5 will be described with reference to FIG. 2 as appropriate.

  First, the skill level determination unit 101 reads the personal profile data 114 from the personal information DB 60 using the user ID 111 as a key (S601). Next, a business ID is received from the business management system 201, and a business related word list is read from the business information DB 50 using the received business ID as a key (S602).

  Thereafter, in step 603, the skill level determination unit 101 determines that the past-created document data 114C in the personal profile data 114 read in step 601 includes the business-related words in the business-related word list read in step 602. The appearance frequency of the business related word is acquired (S603). Thereafter, in step 604, if the past browsing document data 114D in the personal profile data 114 read in step 601 includes the business-related words in the business-related word list read in step 602, the appearance frequency of this business-related word. Is acquired (S604).

  Thereafter, in step 605, the skill level determination unit 101 acquires a list of related business IDs and a belonging period in the belonging department data 114B in the personal profile data 114 (S605). Thereafter, in step 606, the related work ID and the charge period are acquired from the work data 114A in the personal profile data 114 (S606). Thereafter, in step 607, the user's skill level for the work indicated by the work ID is calculated according to the formula (1) based on the acquired various data, and is output as the user skill level data 115 (S607).

  Expression (1) is an expression for calculating the skill level of the user for the business. In equation (1), the degree of department relevance (business ID, department ID) is the degree of knowledge acquisition of related work (depth of relation between business and department) according to the period of belonging to the department. L is a proficiency function that indicates the correlation between the period in which a department or business has been engaged and the proficiency level of related knowledge. L (Affiliation period (department ID)) is the proficiency level for the work by the period of belonging to the department. L (in charge period (business ID)) is a proficiency level for the business during the business period. p (w) is a frequency of appearance of business related words in a document created in the past (past created document data 114C). D (w) is the word specialty based on the related word list of the business ID. In other words, Σ (p (w) × D (w)) is a proficiency level for work estimated from documents created in the past. It is determined that the proficiency level is higher as the documents created in the past contain more business related words. q (w) is a frequency of appearance of business related words in a document browsed in the past (past browsed document data 114D). In other words, Σ (q (w) × D (w)) is a proficiency level for work estimated from documents browsed in the past. In addition, it is judged that the proficiency is higher as the documents browsed in the past contain more business-related words. Α, β, γ, and δ are constants of α ≧ 0, β ≧ 0, and γ ≧ δ ≧ 0, respectively.

  Through the processing described above, the skill level determination unit 101 can determine and quantify the user's skill level for business.

  FIG. 7 is a flowchart illustrating the control logic of the ambiguity resolution unit 103 according to the embodiment of this invention. Here, the processing of step 510 in FIG. 5 will be described with reference to FIG. 2 as appropriate.

  First, the ambiguity resolution unit 103 reads the personal work history data 113, the causal network data 106, and the interesting word data 116 (S701 to S703). Next, in step 704, it is determined whether or not the input interesting word data 116 appears in two or more nodes in the causal network data 106 (see FIG. 9) (S704).

  If it appears in one or less nodes (NO in S704), since this interesting word data 116 is a new word or a word that has only one meaning, the process proceeds to step 705, and the ambiguity resolution unit 103 determines this interest word. The word data 116 is output as it is (S705), and the process is terminated. On the other hand, when the node appears in two or more (plural) nodes (YES in S704), the node's context suitability is calculated for each of the plurality of nodes by Expression (2) (S706). The context suitability is an index indicating the suitability of an interesting word in a document.

Expression (2) is an expression for calculating the context suitability of the node. In Expression (2), d (w, i) is a (semantic) distance on the causal network data 106 between the word w appearing in the personal work history data 113 and the interest word i indicated by the interest word data 116. It is. This d (
w, i) becomes smaller as the causal relationship between the node and the word is stronger. t (w) is the elapsed time (temporal distance within the context) since the word w last appeared in the context. C1 and C2 are constants. It is to be noted that the more words that are located in the vicinity of the relevant interesting word node appear on the personal work history data 113 and the newer the appearance time on the personal work history data 113, the more the calculation is performed according to the equation (2) The degree of context suitability is high.

  Thereafter, in step 707, the ambiguity resolution unit 103 compares the context suitability of each node calculated in step 706, and adopts a node having a high context suitability (S707). Thereafter, in step 708, the node (interested word with limited meaning) adopted in step 707 is output (S708).

  By the processing described above, the ambiguity resolution unit 103 can resolve this ambiguity, limit the meaning to the most suitable interest word, and output this interest word when the interest word has a plurality of meanings. it can.

  FIG. 8 is a flowchart illustrating the control logic of the causal extraction unit 104 according to the embodiment of this invention. Here, the processing of step 510 in FIG. 5 will be described with reference to FIG. 2 as appropriate.

  First, the causal extraction unit 104 reads one unprocessed document (processing target document) from the document DB 70 (S801). In step 802, the causal extraction unit 104 extracts one word pair having a causal relationship from the document read in step 601 (S802).

  Thereafter, in step 803, the causal extraction unit 104 determines whether or not two words constituting the extracted word pair are both registered as nodes in the causal network data 106 (S803).

  If the two words are registered together (YES in S803), the process proceeds to step 804, and the causal extraction unit 104 connects the two registered nodes with a link (link strength 1) (S804). The link strength is an index indicating the strength of the causal relationship between nodes. In step 804, if two nodes are already connected by a link, 1 is added to the link strength. On the other hand, if both words are not registered (NO in S803), the process proceeds to step 805, and the causal extraction unit 104 determines whether one of the words is registered as a node in the causal network data 106. (S805).

  If only one word is registered (YES in S805), the process proceeds to step 806, where the causal extraction unit 104 registers the unregistered word as a new node, and links the two nodes (link strength 1). ) (S806). On the other hand, if neither of the two words is registered (NO in S805), the process proceeds to step 807, where the causal extraction unit 104 registers the two words as new nodes, and links the two nodes (link strength 1). (S807).

  Thereafter, in step 808, the causal extraction unit 104 determines whether there is another word pair having a causal relationship (S808). If there is another word pair (YES in S808), the process returns to step 802 and the process is performed. repeat. In step 809, the causal extraction unit 104 determines whether there is an unprocessed document in the document DB 70 (S809). If there is another unprocessed document (YES in S809), the process returns to step 801. Repeat the process.

  Through the processing described above, the causal extraction unit 104 extracts the causal relationship between words in all documents in the document DB 70 and creates the causal network data 106.

  In particular, by extracting causal relationships from technical documents at the manufacturing site, past documents can be used for the purposes of design support, investigation of the cause of defects, prevention of recurrence of defects, and the like.

  FIG. 9 is a diagram illustrating an example of the causal network data 106 created by the causal extraction unit 104 according to the embodiment of this invention. The causal extraction unit 104 creates causal network data 106 as shown in FIG. 9 based on the accident / defect report at the manufacturing site stored in the document DB 70.

  In FIG. 9, the rectangle indicates a node. An arrow represents a link, that is, a causal relationship between nodes. For example, “friction 901” and “heat generation 902” indicate that “friction 901” is a causal relationship that causes “heat generation 902”. Further, “heat generation 902” and “component deformation 903” indicate that “heat generation 902” is a causal relationship that causes “component deformation 903”.

  FIG. 10 is a flowchart illustrating the control logic of the interesting word extraction unit 102 according to the embodiment of this invention. Here, the processing of step 508 in FIG. 5 will be described with reference to FIG. 2 as appropriate.

  First, the interested word extraction unit 102 reads the personal work state data 112 (S1001). Next, in step 1002, the user skill level data 115 is read (S1002). Thereafter, in step 1003, the personal work history data 113 is read (S1003). Thereafter, in step 1004, the interested word candidate is initialized (S1004). An interesting word candidate is an interesting word group which becomes a candidate for the interesting word extraction part 102 to extract an interesting word.

  Thereafter, in steps 1005 to 1008, the interested word extracting unit 102 searches the related word list in the personal profile data 114, the feature words in the document 112B being created in the personal work status data 112, the feature words in the document 112C being browsed, and the search. A word included in the search expression input to the search engine 301 in the engine state 112D is added to the candidate for interesting words (S1005 to 1008). Here, the characteristic words in the document are words that are highly related to the document, such as being included in many in the document. For example, the word is characterized by a word having a high TF (Term Frequency) -IDF (Inverse Document Frequency). Can be treated as a word. TF (Term Frequency) -IDF (Inverse Document Frequency) is a value representing the likelihood of a feature word in a document, and is obtained by dividing TF by DF (Document Frequency). TF is the frequency of appearance of words in the document. DF is the appearance frequency of words in all documents. Thereafter, in step 1009, the interested word extraction unit 102 calculates the interest levels of all the interested word candidates according to the equation (3) (S809). The degree of interest is an index indicating the degree of interest of the user with respect to the interesting word.

  Expression (3) is an expression for calculating the interest level of the interesting word candidate. In Equation (3), the business related coefficient is a related coefficient of a related word called by the business ID. The user skill level is the user skill level data 115 read in step 1002. The appropriate skill level is a word specialty and a skill level of a user who is highly likely to be interested in the word. If it is assumed that a highly specialized word has a low appearance frequency, the word specialty can be calculated based on the DF.

  The importance in the search expression is the importance of the word assigned to each word in the search expression. For example, when the search formula is an AND search, importance is equally distributed to each word connected by AND. On the other hand, when the search expression is an OR search, the importance is assigned to each word connected by OR without being divided equally.

  For example, when the search formula is (A∨B) ∧C), the importance in the search formula is (A, B, C) = (0.5, 0.5, 0.5). When the search formula is (A∧B) ∨C), the importance in the search formula is (A, B, C) = (0.5, 0.5, 1.0). Α, β, γ, δ, and ε are constants of 0 or more.

  According to this equation (3), words that are closely related to the work being worked on, words that appear particularly frequently in the created document and the read document, words that have the same level of expertise as the skill level of the worker, and the worker When using a search system, the more the word included in the search formula, the higher the degree of interest.

  Thereafter, in step 810, the interested word extracting unit 102 extracts an interesting word candidate having a certain degree of interest and outputs it as interesting word data 116 (S810).

  Through the processing described above, the interested word extraction unit 102 can extract words that are likely to be of interest to the user at that time based on the personal work history data 113, the personal profile data 114, and the like.

  FIG. 11 is a flowchart illustrating the control logic of the headline creation unit 105 according to the embodiment of this invention. Here, the processing of step 511 in FIG. 5 will be described with reference to FIG. 2 as appropriate.

  First, the headline creation unit 105 reads the causal network data 106 (S1101). Next, in step 1102, a set of document data 311 as a search result by the search engine 301 and an interesting word (hereinafter referred to as “search word”) indicated in the interesting word data 117 is read (S1102).

  Thereafter, in step 1103, the headline creation unit 105 extracts words having a causal relationship within a certain distance from the search word on the causal network data 106 (S1103). Thereafter, in step 1104, the search word is the first word, and the word extracted in step 1103 is the second word (S1104).

  Specifically, the interesting word that is the basis for the relationship between the user and the content is the first word. This first word may be plural. Words related to the business being executed by the user, words in the document being created, feature words of the browsing document history, search keywords, etc. are candidates for the first word. By making such a first word a part of the headline, the satisfaction (transparency) with respect to the presented content can be enhanced. Next, in the contents, the second word that is most closely related to the first word is determined.

  Thereafter, in step 1105, the headline creation unit 105 creates a headline for the document by combining the first word, the second word, and the interpolated expression (S1105). The interpolated expression here indicates expressions such as “by ...”, “cause of ...”, “cause ...”, which represent the causal relationship between the first word and the second word.

  In this way, the headline creation unit 105 combines the first word and the second word, and creates a headline by interpolating the expression representing the relationship between these words. For example, when handling a technical document at a manufacturing site, the technical document has a high probability of having a cause, phenomenon, and countermeasure structure. In the document, a word that is in any one of cause-phenomenon, cause-countermeasure, phenomenon-countermeasure relationship with the first word is selected as the second word.

  FIG. 12 is a diagram illustrating an example of a headline created by the headline creation unit 105 according to the embodiment of this invention.

  When the first word (interesting word) is “short circuit” and the second word is “aging operation”, the created heading is “short circuit by aging operation” as shown in FIG. Also, if the second word is extracted (“bolt loosening”, “local overheating”, “bolt energization”, “melt”, etc.), multiple headlines are created (“short circuit due to bolt loosening”, “local”). Short circuit due to overheating, "short circuit due to volt energization", "short circuit due to melt", etc.). When multiple headings are created from one document, priority is given to causal headings that frequently appear in other documents. That is, a headline (in the example shown in FIG. 12, “short circuit due to melt”) created by two words having the strongest link strength on the causal network data 106 is employed. When there is no link that directly connects the first word and the second word, the sum of the reciprocals of the strengths of all the links that pass through may be calculated, and the sum of the reciprocals may be used as an index for determining the strength of the causal relationship.

  Returning to FIG. 11, in step 1106, the headline creation unit 105 adds the headline created in step 1105 to the document and outputs it as headlined document data 118 (S1106). Thereafter, in step 1107, it is determined whether there is another document as a search result (S1107). If there is another document (YES in S1107), the process returns to step 1102 to repeat the process.

  Through the processing described above, the headline creation unit 105 creates a headline that matches the user's work situation and knowledge, and assigns the created headline to the document for output. The headline creation unit 105 may create the same headline from a plurality of contents. In this case, the plurality of contents may be grouped by giving the same heading as the same quality contents. This prevents browsing of documents with overlapping contents.

  FIG. 13 is a flowchart illustrating the control logic of the heading list display unit 401 according to the embodiment of this invention. Here, the processing of step 512 in FIG. 5 will be described with reference to FIG. 2 as appropriate.

  First, the heading list display unit 401 reads all the headlined document data 118 output from the heading creation unit 105 (S1301). Next, in step 1302, all the headings that do not overlap among the headlined document data 118 read in step 1301 are displayed (S1302). Thereafter, in step 1303, the user input is waited (S1303).

  Thereafter, in step 1304, when there is a user browsing request for the headline (YES in S1304) and there are two or more documents corresponding to the headline (YES in S1305), the heading list display unit 401 displays The document is displayed, and when there is a browsing request, another document is displayed (S1306).

  On the other hand, when there is a user browsing request for a headline (YES in S1304) and there is one document corresponding to the headline (NO in S1305), the headline list display unit 401 corresponds to the headline. The document is displayed (S1307). After the document is displayed, the process returns to step 1303 to wait for user input again. In step 1304, if there is no user browsing request (NO in step S1304), the process ends.

  Through the processing described above, the heading list display unit 401 displays the document data 118 with a heading created by the heading creation unit 105.

  FIG. 14 is a diagram showing an example of a display screen 1401 by the heading list display unit 401 according to the embodiment of this invention.

  In the display screen 1401 shown in FIG. 14, a title created by the header creation unit 105 is displayed in the title 1402. A text confirmation button 1403 is a button for confirming the text of the document corresponding to the title 1402. In the business name 1404 and product name 1405, the business name and product name related to the title 1402 are displayed. In the related word 1406, a word related to the title 1402 is displayed.

  By displaying the display screen 1401 as shown in FIG. 14, the user can quickly and accurately determine whether or not the displayed information is necessary.

  As mentioned above, although embodiment of this invention was described, the said embodiment showed one of the application examples of this invention, and in the meaning which limits the technical scope of this invention to the specific structure of the said embodiment. Absent. Various modifications can be made without departing from the scope of the present invention.

1 Computer System 10 Information Providing Server 20 Business Management Server 30 Search Server 40 Client Computer 50 Business Information DB
60 Personal Information DB
70 Document DB
101 Skill Level Determination Unit 102 Interesting Word Extraction Unit 103 Ambiguity Resolution Unit 104 Causal Extraction Unit 105 Headline Creation Unit 106 Causal Network Data 112 Personal Work Status Data 113 Personal Work History Data 114 Personal Profile Data 115 User Skill Level Data 116 Interesting Words Data 118 Headlined document data 401 Headline list display section

Claims (7)

A computer comprising a processor for executing a program, a memory for storing a program executed by the processor, and an interface connected to the processor, and connected to a document database for storing documents,
The processor is
Extracting the causal relationship between words included in the document stored in the document database, creating causal network data representing the extracted causal relationship between the words in a network structure,
From the causal network data, extract words that are causal with an interest word that is a word that the user is interested in,
Based on the first word and the second word when the interesting word is the first word and the extracted word is the second word, a headline of the document is created,
A computer that generates data for presenting a created document with the heading to a user via an output device.   The processor extracts an interest word of the user based on work state data indicating a work state of the user for a predetermined job and profile data indicating a profile of the user. Item 1. The computer according to item 1. The processor creates user skill data indicating the skill level of the user for the predetermined job based on the profile data,
The computer according to claim 2, further comprising extracting an interest word of the user based on the worker skill level data.   When there are a plurality of interesting words, the processor selects an interesting word in which the words around the interesting word in the causal network data appear with the highest appearance frequency in the work state data. The computer according to claim 1, wherein the computer is an interesting word of the user.   The said processor produces the heading of the said document by the said 1st word, the said 2nd word, and the interpolation expression showing the causal relationship of the said 1st word and the said 2nd word. 1. The computer according to 1. A document presentation method in a computer comprising a processor for executing a program, a memory for storing a program executed by the processor, and an interface connected to the processor, and connected to a document database for storing the document. ,
The processor is
A procedure for extracting a causal relationship between words included in a document stored in the document database, and creating causal network data representing the extracted causal relationship between the words in a network structure;
A procedure for extracting a word that is causal with an interesting word that is a word that the user is interested in from the causal network data;
Creating a document heading based on the first word and the second word when the interesting word is the first word and the extracted word is the second word;
A procedure for generating data for presenting the created document with the heading to a user via an output device;
The document presentation method characterized by performing. A document presentation program executed on a computer connected to a document database for storing a document, comprising: a processor for executing the program; a memory for storing the program executed by the processor; and an interface connected to the processor. Because
A procedure for extracting a causal relationship between words included in a document stored in the document database, and creating causal network data representing the extracted causal relationship between the words in a network structure;
A procedure for extracting a word that is causal with an interesting word that is a word that the user is interested in from the causal network data;
Creating a document heading based on the first word and the second word when the interesting word is the first word and the extracted word is the second word;
A procedure for generating data for presenting the created document with the heading to a user via an output device;
Is executed by the computer.

Images