JP2010039605A - Person search system, person search method, program and recording medium - Google Patents

Abstract

To provide a personal search system, a personal search method, a program, and a recording medium for searching for an individual by using a plurality of types of information in a cross-sectional manner.
A search server includes a database that registers documents created in the past and index data for indexing document contents, and a document management unit that calculates similarity between documents stored in the database. A personal information management unit 126 that calculates a matching score for skill level, role, and attribute value of the department to which the user belongs, and a schedule management unit 128 that calculates a matching score by comparing the personal schedule with the review deadline of the created document, A reviewer ranking unit 130 that registers individuals related to similar documents with high similarity to the search string in the reviewer set list, and ranks the individuals as reviewers according to the context set in the context management unit 124. including.
[Selection] Figure 1

Description

  The present invention relates to an individual search, and more particularly to an individual search system, an individual search method, a program, and a recording medium that search for an individual by using a plurality of types of information in a transverse manner.

  In the work centered on knowledge and knowledge based on documents such as the R & D field, the marketing field, and the news report field, review work by an individual who has sufficient knowledge in a field related to the document, that is, an expert, is usually performed. The review work is required to guarantee the quality of the intellectual output and to provide a short period until the review results are provided. An application used for document review up to now supports document review work in the form of establishing a workflow for review and managing the established workflow.

  In addition, an expert search system that uses a document database, a personal profile (skill profile), schedule information, and the like has been known so far. However, the known expert search methods use information created by independent applications in tandem, and API (Application Program Interface) must be generated between each application. There was a problem.

  In addition, when deciding an expert, the attribute values that should be used for selecting an expert are usually different in response to specific requests from companies, organizations, government agencies, etc. There was also the inconvenience of having to adapt it. Furthermore, in the conventional system, even though each information described above is provided, each information is managed by a different computer or program, so it is necessary to create an API, and further, an OS between computers, a programming language In order to offset these differences, a relatively large-scale system using CORBA (Common Object Resource Broker Architecture) or the like has been required.

  For these reasons, it is difficult to efficiently search the above-described information for searching for an expert, and system maintenance also wastes time and cost.

  So far, systems for searching for individuals are known in terms of managing workflows. For example, in Japanese Patent Application Laid-Open No. 2008-071082 (Patent Document 1), information stored in a user terminal is appropriately handled as task related information (information related to a task stored in the user terminal is stored, searched, and browsed). A workflow management system that can be used is described.

  Japanese Patent Application Laid-Open No. 2008-066584 (Patent Document 2) is a workflow management system that dynamically configures a workflow model when a workflow is executed, and an arbitrary user including the target user is assigned to a target user. Means for searching for a task, an attached document and attached information using history information of access to the task, attached document and attached information, and means for presenting the search result to the target user A workflow management system characterized by the above is disclosed.

  Patent Document 1 describes performing workflow management using a plurality of types of information stored in a user terminal, but does not disclose a process for searching for an optimum individual for reviewing a specific document. . Also, in Patent Document 2, in order to enable a workflow model to be dynamically configured when a workflow is executed, a task or document is searched using the user's past history information, and the search result is searched for the user. Although it is disclosed, it is not disclosed that a person in charge who is most suitable for reviewing a document or the like is searched by using a plurality of pieces of information.

  In addition, US Pat. No. 2005 / 0182743A1 (Patent Document 3) discloses an expert matching system. However, the expert matching system described in Patent Document 3 does not disclose what kind of criteria is used to specify an expert related to an object such as a specific document. Searching in association with the similarity of is not a problem to be solved.

  Furthermore, US Publication No. US2006 / 0248076A1 (Patent Document 4) discloses a technique for searching for an author from an input phrase. In Patent Literature 4, an author is searched from an enormous field collection document by searching and ranking an author related to the input phrase from the input phrase. In the technique described in Patent Document 4, an expert is searched by searching for and ranking the author of a document related to an input phrase. However, it is not an issue for the expert to determine the best reviewer candidates for review including whether or not other documents can be reviewed.

In other words, it can be used as an expert related to a specific field by using information on related documents, skills, schedules, roles, departments, etc., and is required by specific groups such as companies, organizations, and government offices. There was a need for a technique for searching for individuals while flexibly responding to attributes.
JP 2008-071082 A JP 2008-066584 A US Published Patent US2005 / 0182743A1 Specification US Published Patent US2006 / 0248076A1 Specification

  The present invention uses information such as document contents, skills, schedules, roles, departments, etc. in a cross-sectional manner in order to efficiently conduct a review related to a specific document. It is an object to provide a personal search system, a personal search method, a program, and a recording medium for determining recommended reviewers.

  In addition, the present invention can be recognized as an expert related to a specific field by using information such as documents, skills, schedules, roles, departments, etc., and is required by a specific group such as a company, an organization, a public office, etc. An object of the present invention is to provide a personal search system, a personal search method, a program, and a recording medium that search for individuals while flexibly corresponding to attributes.

  In the present invention, a search server is connected to a network and receives a search request for an individual who requests a review from a client computer (hereinafter simply referred to as a client). The search request includes a document fragment that is a digest of a part of the document created by the client as a search string.

  The search server includes a database, and the database registers the ready-made document created in the past and index data for indexing the content of the ready-made document, and the document, document review history, schedule, skill level, role, and Department attributes are integrated and managed by making it possible to refer to individuals related to ready-made documents similar to search strings.

  The search server includes context management means for managing score functions to rank individuals as reviewers of documents to be targeted. The search server identifies individuals associated with similar documents that are similar to ready-made documents for which reviewers should be determined. Then, a recommended score related to the document to be reviewed is calculated using the past created document, review history, schedule, skill level, role, and attribute value of the department to which the identified individual is identified. An individual is specified by identification information such as a user ID and a user name that uniquely identify the individual.

  The extracted individual identification information is registered as an instance of the reviewer set list by the reviewer ranking means. Then, a matching score is calculated for each attribute value for the individual registered in the reviewer set list, and a recommended score is calculated according to the context set in the reviewer set list and set in the context management means. A score including a matching score for all the attribute values to be considered is set as a recommended score, and the ranking of the individual related to the similar document is calculated as a reviewer using the value of the recommended score.

  In the present invention, it is preferable that the context management unit manages a score function including the matching score as a variable. Furthermore, it is preferable that the personal search means of the present invention ranks individuals by equivalently processing attribute values of a similar document creation history or a past review history of similar documents. The personal search system of the present invention can be implemented in a search server as reviewer determination means for creating a document.

  Hereinafter, although this invention is demonstrated with embodiment, this invention is not limited to embodiment mentioned later. FIG. 1 shows functional blocks of a search server 110 included in the personal search system of this embodiment. The personal search system 100 is preferably configured as a web server from the viewpoint of troublesome client application installation and maintainability.

  The search server 110 can be accessed by a client user from multiple clients (not shown) via a network 150 including the Internet, a local area network (LAN) using wired or wireless communications, or a wide area network (WAN). A reviewer search request is received that includes a search string generated from a document fragment of the generated document. In this embodiment, the document fragment is defined as a created document or a summary created from the document, a title, a sentence in a specific area, and the like. The document fragment can be created on the client side and sent to the search server 110. However, the document specified by the search string by a text editor included in the search server 110 using the search string received from the client. The search server 110 can be created for each reviewer search request from the above area.

  In the embodiment described below, when the client issues a reviewer search request, the title of the document for which the reviewer is to be determined is input, and the title is included in the search string and provided to the reviewer search. It will be explained as a thing.

  The search server 110 manages user data in a database. The user of the search server 110 means a user of a client and a registrant who can access the search server 110 and is an individual who is a target of a reviewer search. Specifically, the individual in the present embodiment means an employee of a specific company, a member of a specific group, a public servant of a public office, and the like, and hereinafter simply referred to as a user unless otherwise specified.

  When receiving the reviewer search request, the search server 110 passes the reviewer search request to the information integration management module 120 via the network interface unit 112 and the input interface unit 114. The network interface unit 112 processes a reviewer search request at the physical layer and data link layer levels using a low-level language level program such as a network interface card (NIC) and an assembler, and sends a reviewer search request to the TCP / IP layer. Will be sent. In the TCP / IP layer, the TCP / IP header of the reviewer search request packet is analyzed, the transaction protocol type is determined, and if the application is responsible for the port specified in the TCP / IP header, for example, the HTTP protocol, A reviewer search request is passed to port number = 80.

  Further, the information integration management module 120 described above can be implemented as a database application. Examples of database applications include relational databases and object-oriented databases (OODB). In the present embodiment, in response to a reviewer search request, information stored in the database 140 is used in a cross-sectional manner and memory capacity is not consumed unnecessarily, so that information integration is performed using OODB. The management module 120 is preferably configured.

  The reviewer search request is sent by designating the HTTP protocol (port = 80), and then passed to the input interface unit 114 implemented using CGI (Common Gateway Interface) or the like. The input interface unit 114 analyzes the received packet, calls an application corresponding to the request, and executes processing. The execution result of the application is sent as a search result to the output interface unit 116 configured as a CGI output module, sent to the network 150 via the network interface unit 112, and the search result is sent to the requesting client. Will be sent.

  In order to execute the above-described processing, the search server 110 uses a CISC architecture microprocessor such as a PENTIUM (registered trademark), a PENTIUM (registered trademark) compatible chip, or a RISC architecture microprocessor such as a POWER PC (registered trademark). It is mounted as a single core or a multi-core, and execution of an application is controlled by an operating system (OS) such as WINDOWS (registered trademark) 200X, UNIX (registered trademark), or LINUX (registered trademark).

  The application is implemented as a server program such as CGI, servlet, or APACHE, which is implemented using a programming language such as C ++, JAVA (registered trademark), JAVA (registered trademark) SCRIPT, PERL, or RUBY. Process reviewer search requests from.

  The information integration management module 120 includes a plurality of database objects. Each object is realized as a function unit and a function unit on the search server 110 when the CPU executes the program using the program and data read into the RAM or the like.

  More specifically, the information integrated management module 120 includes a document management unit 122, a context management unit 124, and a personal information management unit 126.

  The document management unit 122 manages documents, emails, comments, review contents, and other ready-made documents that have already been created, and executes a search for similar documents similar to the ready-made documents in order to search for reviewers. To do. The document management unit 122 calculates the similarity between the ready-made document stored in the database 140 and the search string using the document, sentence, and other fragments that are the search string included in the reviewer search request. Then, the similarity is compared with a set similarity threshold, and a ready-made document having a similarity equal to or higher than the similarity threshold is determined as a similar document. Thereafter, user identification information related to the determined similar document is acquired and passed to the reviewer ranking unit 130. Note that the user to be extracted is referred to by the identification information, and the document management unit 122 creates the document similar to the document to be reviewed or equivalently identifies the identification information of the user who has a history of reviewing the similar document before. The identification information of the user is searched by weighting and passed to the reviewer ranking unit 130. For this reason, the document management unit 122 can extract a review expert who is a user who has never created a document directly and specializes in review, and enables more flexible expert search.

  In the present embodiment, the similarity calculation can be used by any known method. The method for executing the similarity search is not particularly limited in the present embodiment, but when registering the parent document, a keyword included in the parent document is used to generate a TF-IDF (Term Frequency-Inverted Document). Data mining is performed by indexing documents using the (Frequency) method, the search string included in the review request is checked against the index data, and the keyword contained in the search string and the frequency of occurrence of the keyword in the index data, etc. A similar search method using can be used. Since the similarity search method is not the gist of the present invention, further detailed description is omitted.

  In this embodiment, the context management unit 124 is an object that manages a score function for calculating a reviewer's recommended score by the reviewer ranking unit 130 described later. The context management unit 122 is called from the reviewer ranking unit 130, reads each matching score registered in the reviewer set list, substitutes it for the score function managed by the context management unit 124, and calculates a recommended score. Thereafter, the context management unit 124 returns the calculated recommended score to the reviewer ranking unit 130 as a return value.

  The personal information management unit 126 manages the user by associating a user ID, a user name, or an e-mail address if necessary as identification information, and responds to a search request using the identification information as a search key. Then, it is referenced across the search. Also, the personal information management unit 126 manages the user's schedule, skill level, role of whether or not the person is a reviewer, department, etc., and each attribute is assigned according to a call from the reviewer ranking unit 130. A value matching score is calculated, and each score value is returned to the reviewer rank attaching unit 130 as a return value.

  The information integrated management unit 120 of the search server 110 illustrated in FIG. 1 illustrates a schedule management unit 128 that is a method of the personal information management unit 126, for example. The schedule management unit 128 manages schedule data for each user together with a schedule management method and a schedule matching method.

  The schedule management method of the schedule management unit 128 searches for a schedule sent from a specific user registered in the database 140 and notifies the user of the schedule. In addition, the schedule matching method of the schedule management unit 128 compares the designated user's schedule with the review deadline in response to the call from the reviewer ranking unit 130, calculates the matching score for the schedule attribute, The matching score is returned to the reviewer rank attaching unit 130 as a return value.

  Further, the information integration management module 120 includes a reviewer ranking unit 130 and a data management unit 132. Upon receiving the reviewer search request, the reviewer ranking unit 130 creates and initializes a reviewer set list as a subclass of the personal information management object. Then, the reviewer ranking unit 130 sequentially adds the user identification information extracted as a result of the similarity search executed by the document management unit 122 as an instance. After that, the context management unit 124 is called to obtain each matching score registered at that time, calculate a recommended score, and receive a recommended score as a return value. Then, the reviewer ranking unit 130 can refer to the searched user's schedule, role, department, etc., and renew the recommended score recursively until there are no attribute values that can be used for judgment, and can recommend it as a reviewer. Search results are created in descending order and sent to the reviewer search request sender.

  Each time the data used by each function processing unit described above is sent from a specific user, the data management unit 132 writes data in an appropriate data area of the database 140, and data specified in response to a request. And the storage and deletion of data to be processed by each function processing unit are managed in an integrated manner.

  In the preferred embodiment, the information integration management module 120 reduces the amount of data such as a table, and further inherits data already created by another management unit, thereby enabling reduction in memory consumption and the like while improving efficiency. Integrated processing is possible.

  FIG. 2 is a flowchart of processing in the personal search method according to this embodiment. The process shown in FIG. 2 starts from step S200. In step S201, the user creates a document on the client and stores it in the database 140. In step S202, the search server 110 displays a reviewer search page on the client in response to a request from the user, downloads the search form, and displays it on the desktop screen of the client. Then, the user designates a document or a document title to be reviewed by the reviewer on the search form.

  In step S203, it is determined whether a search command has been issued. During this time, the client receives and buffers an input of an additional input item when executing a reviewer search input by the user. The additional input items are determined based on whether or not there is an input in an input field provided on a GUI (Graphical User Interface) displayed as a search form.

  If an additional input item is described in the input field instead of receiving a search request in step S203 (no), input of a review deadline, skill information, additional keywords, etc. is received from the user on the search form in step S209. Then, after registering in the buffer memory, the process is branched again to step S203. On the other hand, if a search request is input in step S203 (yes), in step 204, a search request is issued to the search server 110 in response to a search command by clicking a “Search” button or the like. The search request is generated including a search string that enables specification of a document fragment such as a document title or a document extraction range input by the user, a user ID, and the like, and is addressed to the search server 110 using the POST method or the GET method. Sent

  In step S205, the search server 120 receives a reviewer search request, and executes a similar search of ready-made documents. In the similar search in step S205, the search server 110 obtains a search string in a range specified from a document to be reviewed by a reviewer or generates a specific range of a document specified by the search string as a document fragment. A document similarity search is performed by matching the search string with the index data.

  The document similarity search can be executed by decomposing the search string into units of words and using the fact that the word is registered in the index data of the ready-made document registered in the database and its appearance frequency. In the present embodiment, the specific method for executing the similarity search of documents is not particularly limited.

  Further, in the ranking process of step S205, the search server 110 configures the reviewer set list as a subclass of the personal information management object that configures the personal information management unit 122. The user registered in the reviewer set list is a user (individual) whose user has been registered as a creator of a similar document extracted as a result of a similar search or whose review history of the similar document has been previously registered. Is done. As a result, attribute values such as skill information, schedule information, role information, and department information managed by the personal information object are inherited, a reviewer set list can be generated efficiently, and transaction management corresponding to reviewer search requests can be performed. Efficiency can be improved.

  Further, in step S205, the reviewer ranking unit 130 sequentially scores each attribute using the extracted skill information, schedule information, role information, department information, etc. of the user, and obtains the scoring result. Use individuals to calculate a recommended score and rank users with the calculated recommended score.

  In step S206, a reviewer list is generated as a structured document in which the extracted users are displayed in descending order of the score. In step S207, the structured document is sent to the client as an HTTP response. A series of processing of the search method ends.

  FIG. 3 shows a detailed flowchart of the process of step S205 among the processes shown in FIG. The process shown in FIG. 3 starts in step S300 when the process is passed from step S204. In step S301, a reviewer set list is generated, and the attributes of similarity, score, skill, schedule suitability, role, department, recommended score, and review request destination are initialized. In step S302, a document is identified from the database based on the document title or user ID included in the reviewer search request, the similarity is calculated using the search string, and the similarity is equal to or higher than the set similarity threshold. Identification information such as a user name and a user ID of a user related to a similar document is acquired as a reviewer candidate and registered in the reviewer set list. Further, reviewers are ranked in descending order of similarity.

  Thereafter, in step S303, a reviewer is extracted from the reviewer set list, the extracted reviewer's skill information is referred, a matching score is calculated, and registered in the score attribute of the reviewer set list. Then, the score of the document similarity is included, and the reviewer is re-ranked.

  In step S304, the reviewer's schedule information and review deadline are inquired, the schedule suitability is calculated, registered in the schedule suitability attribute, and re-ranked by the schedule. Note that the schedule conformance is determined by whether or not the reviewer's schedule to be inspected for schedule conformance has already been registered including the review deadline. Degree = 0, and if not registered, goodness of fit = 1.

  The user is then re-ranked with the recommended score using the value of each attribute already acquired, including the schedule fit. In step S305, the role information of the reviewer is inquired to determine the presence or absence of review authority, the score value of the role attribute is calculated, registered in the role attribute, and the user is re-ranked including the value of the role attribute. . Note that the value of the role attribute is set as roll value = 1 if the reviewer in process is authorized to review, and role value = 0 if not authorized.

  In step S306, the department information of the reviewer is inquired, registered as a department attribute value, and the user is re-ranked including the department attribute value. The value of the department attribute is determined by the expertise of the document to be reviewed, and the same product name, project name, or category name specified by the department or comment that is the same as or related to the user requesting the review Assign a score = 1 for reviewers who have sex and score = 0 for reviewers who have no relevance.

  In step S307, it is determined whether or not the inquiry has been completed for all the instances in the reviewer set list. If the inquiry has not been completed (no), the process returns to step S303 to repeat the processes up to step S307. On the other hand, if it is determined in step S307 that all the stances have been inquired (yes), in step S308, an identification value such as the email address of the recommended reviewer is registered in the review request destination attribute of the instance of the reviewer set list The attribute to be presented to the client user is extracted to generate a search result, and the process ends in step S309.

  FIG. 4 shows a data structure and a schematic search schema 400 accessed by the reviewer ranking unit 130 in the personal search processing executed by the search server 110 of the present embodiment. When the search server 110 receives a reviewer search request from a client, the reviewer ranking unit 130 declares and initializes the reviewer set list 480. Thereafter, the search server 110 calls the document management unit 122 and issues a search request for document similarity search to the database 140. In the embodiment to be described, the index data 410 is used to search for similar documents and search for similar documents. The document 420 is extracted, and the extraction of the user related to the similar document 420 is instructed.

  The document management unit 122 performs a similar search on the index data registered in the database 140 using a search string included in the search request, and selects a document whose similarity is equal to or higher than the set similarity threshold. The identification information of the user extracted as 420 is passed to the reviewer ranking unit 130 as a return value, and added to the instance of the reviewer set list. In another embodiment of the similar search, the full text search may be executed using the document itself registered in the database 140.

  When the reviewer ranking unit 130 obtains the user identification information related to the extracted similar document, the reviewer ranking unit 130 registers it as an instance of the reviewer set list 480. The reviewer ranking unit 130 instructs the user to refer to each registered attribute value, and each attribute value such as the extracted skill information 440, schedule information 450, role information 460, department information 470 of the user is extracted. And the score value calculated from each attribute value is acquired sequentially, and the process of registering in the reviewer set list 480 is executed.

  Each registered attribute value is used to sequentially calculate a reviewer's recommended score, and in the described embodiment, attribute values for department information 470 are obtained and scored for all attribute values. When the is completed, the final reviewer's recommendation score is determined and the recommendation ranking is completed. The reviewer rank attaching unit 130 acquires attribute values and score values to be presented as search results from the instances of the reviewer set list 480, and attaches appropriate tags such as XML and HTML to the structured document. Create The prepared structured document is sent to the output interface 116, and the search result is sent to the client as an HTTP response in the form of a table, thereby providing the recommended reviewer information to the client.

  5 to 7 show data structures of document information, personal information, and skill information used by the search server 110 of this embodiment. The document information shown in FIG. 5 specifies a document abstract, a parent document, or a user ID of the user who created the document, specifying the document abstract, the parent document, or a document created accompanying the parent document. Creator information such as user name, and when review of the document is completed, reviewer information that is identification information of the user who performed the review, reviewer information is posted and registered in response to the completion of the review Review history, review period, comments attached to the document, review status indicating the execution status of the review work, and a URI (Uniform that registers the index or index and document used to search for the document corresponding to the instance. Register Resource Identifier).

  When the reviewer search request is received, the reviewer ranking unit 130 calls the document management unit 122 to access the URI or the like managed by the document management unit 122 of the database 140, and uses the search string for the index or the document body. And execute a query to calculate the document similarity. The document management unit 122 acquires identification information of a user registered for a document whose calculated similarity is equal to or greater than the similarity threshold, or a user registered in the review history, and passes it to the reviewer ranking unit 130 for review. Registered in the user group list 480.

  The reason for extracting similar document creators and users who have similar document review histories as recommended reviewers is that users who have created similar documents can recommend them as reviewers. This is because a user who exclusively reviews a document can estimate a certain reviewer's accuracy for a document to be reviewed.

  FIG. 6 shows the data structure of the personal information 600 registered in the database 140 in this embodiment in the form of a table. The personal information 600 according to the present embodiment manages a plurality of attribute values, and enables the user attribute of the user, the role attribute of the user, the administrator attribute designating the administrator for the user, and the schedule contents to be referred to. Register schedule attributes. Furthermore, the personal information 600 includes a skill attribute that allows the user to refer to skill information unique to the user, a created document attribute that registers a document ID and the like to enable reference of a document created by the user, In order to refer to the review history, a reviewed document attribute for registering the document ID of the document that has been reviewed and a department attribute that allows the department to which the user belongs to be registered.

  After extracting similar documents, the reviewer ranking unit 130 extracts the user IDs of creators of similar documents or users who have a review history, and the extracted user ID values are added to the reviewer set list 480 as reviewer candidates. to add. Further, the reviewer ranking unit 130 calls the context management unit 124, sequentially refers to the role attribute, schedule attribute, skill attribute, and department attribute specific to the user, and uses the score function set from the designated viewpoint. Then, the score calculation is performed, and the instances of the reviewer set list 480 are sequentially sorted in the calculated score order to generate the reviewer set list 480 sorted by the final recommended score.

  FIG. 7 shows an embodiment of skill information 700 for providing skill levels in this embodiment in a table format. The skill level is registered for each specific user using the skill content information 710 for registering the skill category such as Language, Technical Skill, Teamwork, and the skill name such as English, Japanese, C ++, and XML. The The skill level 720 of a specific user is assigned an index such as SKI_A (A is a value that uniquely identifies the user, such as the user ID of the user), and when the user is specified, the user ID of the user, etc. The skill level 720 is searched by using the identification information as an index.

  In the skill level 720, a skill name, a category, and a skill level are registered. In the embodiment to be described, the skill level is calculated in the order of increasing value, and the score is calculated for the skill. Skill information scoring can use various calculation methods. For example, a skill vector is defined with the skill name as an orthogonal axis, and the inner product of the skill level sent from the user is calculated. A score can be calculated.

  In addition, role information, schedule information, and department information are also registered as role names, categories, dates and schedule contents, department names, etc., and each score is calculated in response to reference by the reviewer ranking unit 130. Used for.

  The recommended score calculation for determining the recommended reviewer is registered by the context management unit 124 and can be executed using a score function given by the following equation (1).

In the above formula (1), W _p is a recommended score, SCORE (attribute) is a score for each attribute such as similarity, skill level, schedule suitability, role suitability, department suitability, and f _attribute Is a weight factor for how much the score for each attribute value is reflected in the recommended reviewer score. In the present embodiment, when there is no value for SCORE (attribute), a recommended reviewer score is calculated by implementing the context management unit 124 so as to return value = 0 instead of returning an exception error, Correct the rank of recommended reviewers sequentially.

  For this reason, when applying to a specific company, for example, when the department attribute is not adopted, the skill level is not adopted, and the schedule is not adopted, the processing of the context management unit 124 and the reviewer ranking unit 130 is performed. It is possible to cope with the minimum cost of customizing each data according to the request of a specific company without correcting the data.

  FIG. 8 shows an embodiment of a reviewer search form 800 sent by a client from the search server 110 of this embodiment. The reviewer search form 800 provides a plurality of input fields in the display window 810, and first, it is possible to specify a document to be a reviewer search target as “Document name”. Note that the “Document name” can be entered by the client user himself / herself, and if the target document is specified through another GUI (Graphical User Interface), the document's similarity, such as its title, can be specified. It is also possible to adopt a configuration for automatically acquiring the minimum data necessary for determining the degree. In the embodiment shown in FIG. 8, Document name = Solution Strategy is set, and the value is sent to the search server 110 as a search string.

  Further, in the display window 810 shown in FIG. 8, the skill name to be requested from the reviewer can be input or designated together with the level as “English 5” or “Strategy 2”. Has been. Note that the skill name can be displayed as a pop-up list or the like instead of being manually input by the user, thereby providing an interface that can be selected by the user. In addition, the display window 810 includes a field 830 for inputting a review deadline, a field 840 for inputting an estimated time required for review, a purpose for requesting a review from the reviewer, or a specific to be added to the reviewer search. A field 850 for inputting a keyword is provided.

  Using the GUI shown in FIG. 8, when there is a skill name to be added, the user clicks the “Add Skill” button 820 to display a pop-up list, and selects the skill name to be added. Can be additionally displayed in the display window 810. After the user inputs data to be requested to the reviewer using the GUI shown in FIG. 8, when the user clicks the “Search” button 860, each data designated from the input field is displayed in the POST method. Alternatively, it is sent as a reviewer search request to the search server 110 by the GET method. The search server 110 receives the reviewer search request, executes the above-described processing, and returns the search result as an HTTP response to end the series of transactions.

  9-12 illustrate an embodiment of a search result 100 sent by the search server 110 to a client. The search result 900 is a search result that is displayed when the document similarity is regarded as important. The reviewer name, the skill suitability, the schedule tolerance, and the department suitability are displayed as the reviewer identification information. A button for instructing a process for requesting a degree and a review is registered for each user.

  The search result 900 shown in FIG. 9 ranks the reviewers calculated by the above formula (1) by assigning the recommended ranks in the recommended score order using the document similarity as the total similarity. it's shown. In FIG. 9, user A is registered higher than user B. In the search result 900 shown in FIG. 9, the user D has the same document similarity and skill suitability in comparison with the user C, but the user D has a skill suitability higher than that of the user C. As a result, the user C is registered higher in the recommended rank than the user D.

  When the user who has received the search result 900 shown in FIG. 9 requests processing from the user A, the user clicks a “request” button registered in the user A record. The “request” button is hyperlinked to the mailer by a tag such as “a href =” mailto: XXXX ”, for example, and the email address of the user A is set as the destination in response to the click of the button Then, the mailer is activated and a review request can be made.

  FIG. 10 shows another embodiment of the search result generated by the search server 110 of this embodiment. In the embodiment shown in FIG. 10, the schedule tolerance and the department are emphasized, and the user B has the same document similarity and skill level as the user A, but the schedule is not allowed. , Ranked below user A. Although user C has a document similarity higher than user B but has no schedule tolerance, user C is more than user B because the department is registered as a department not in charge of review. Ranked low. User D also has a higher document similarity than user B, but is ranked lower than user B in terms of schedule tolerance, skill level, and department.

  On the other hand, user A is registered as a recommended rank = 1 because it is acceptable in terms of schedule and has a higher document similarity than other users. The user who has received the search result 1000 shown in FIG. 10 can make a review request by clicking the “request” button registered in the record of the user A.

  FIG. 11 illustrates the implementation of search results 1100 when scoring a user using role relevance when the document similarity, skill relevance, schedule tolerance, and department are equivalent for the extracted users. It is a form. In the search result shown in FIG. 11, the user has no different attribute other than the role. In the embodiment shown in FIG. 11, the user A and the user B and the user C and the user D are assigned to the user A by the role. , User B is ranked higher. For user A and user B, for example, by evaluating the order of user IDs and user IDs, past review results, and the like, the display order can be determined and used as a search result.

  FIG. 12 shows an embodiment of the personal search system 1200 of this embodiment. In the personal search system 1200, a search server 1270 that manages a database 1280 is connected to a network 1260. The database 1280 stores various data to be used by the search server 1270 in this embodiment, and in response to a query from the search server, various reviews are implemented as functional modules of the search server 1270. It is returned to the ranking section 130.

  A plurality of clients 1210 to 1250 are connected to the network 1260. Each of the clients 1210 to 1250 is configured as a personal computer on which a CPU, RAM, ROM, hard disk drive, and the like are mounted. The clients 1210 to 1250 are Internet Explorer, Mozilla, Opera, etc. under the control of WINDOWS (registered trademark) XP, Vista, MAC (registered trademark) OS, UNIX (registered trademark), LINUX (registered trademark), etc. The search server 1270 is accessed via the browser software using the HTTP protocol, a document registration and a reviewer search request are sent, and the search result is received as an HTTP response.

  FIG. 13 is a sequence diagram for a transaction between the client 1210 and the search server 1270. Note that the following sequence diagram is a process in which a database object implemented as each functional unit of the search server 1270 is called and executed by the knowing from the reviewer ranking unit 130. As shown in FIG. 13, the client 1210 issues a reviewer search request to the search server 1270 in step S1300. In the reviewer search request, as described with reference to FIG. 9, values such as a document name, a user ID, and a skill name as a search string are described in the data area. In step S1310, the search server 1270 receives the HTTP request, calls up a similar search process for documents using the search string, and executes a similar search using the database 1280.

  In step S1320, the search server 1270 first calculates a document similarity using a search string, an optionally added keyword, a comment, and the like, extracts documents that are equal to or higher than the similarity threshold, and stores them in the document. Related users are extracted, identification information is passed to the reviewer ranking unit 130 of the search server 1270, and registered in the reviewer set list 480. In step S1330, ranking is performed by sorting users by document similarity, and a recommended rank is generated. In step S1340, a skill level matching score calculation process is instructed, and score calculation is executed using the database 1280. In step S1350, a skill level matching score is acquired and registered in the reviewer ranking unit 130. In step S1360, the users are sorted and reranked by including the skill level matching score.

  Thereafter, in step S 1370, the schedule management unit 128, personal information management unit 128, etc. are called for the user to request calculation of matching scores for schedules, roles, departments, etc., and each score value calculated with reference to the database 1280 is obtained. Is sent to the reviewer ranking section 130. Thereafter, each score is acquired in step S1380, and in step S1390, the users are sequentially ranked including each score, and the processes from step S1370 to step S1380 are recursively performed as many times as there are attributes to be included in the evaluation. Repeat to generate a final user recommendation ranking. When the processing in step S1380 is completed, a search result is created as a structured document such as HTMLXML, and in step S1400, the reviewer search result is returned as an HTTP response to the client, and the series of transactions is terminated.

  As described above, in the present embodiment, it is possible to search for a prior expert as a reviewer in association with the document similarity, and it is possible to improve the document completion efficiency and the document completion workflow.

  The functions of the present embodiment are implemented by a device executable program described in an object-oriented programming language such as C ++, Java (registered trademark), Java (registered trademark) Applet, Java (registered trademark) Script, Perl, or Ruby. The program can be stored and distributed in a device-readable recording medium such as a hard disk device, a CD-ROM, an MO, a flexible disk, an EEPROM, or an EPROM. Can be transmitted.

  Although the present embodiment has been described so far, the present invention is not limited to the above-described embodiment, and other embodiments, additions, changes, deletions, and the like can be conceived by those skilled in the art. It can be changed, and any aspect is within the scope of the present invention as long as the effects and effects of the present invention are exhibited.

The figure which showed the functional block of the search server 110 which the personal search system of this embodiment contains. The flowchart of the process of the personal search method of this embodiment. The figure which showed the detailed flowchart of the process of step S205 among the processes shown in FIG. The figure which showed the data structure and search schema 400 which are accessed by the personal search process which the search server 110 of this embodiment performs. The figure which showed the data structure of the document data which the search server 110 of this embodiment uses. The figure which showed the data structure of the user data which the search server 110 of this embodiment uses. The figure which showed embodiment of the skill data 700 for providing a skill level. The figure which showed embodiment of the reviewer search form 800 which the search server 110 of this embodiment sends from a client. The figure which showed embodiment of the search result which the search server 110 of this embodiment produces | generates. The figure which showed embodiment of the other search result which the search server 110 of this embodiment produces | generates. The figure which showed embodiment of the further another search result which the search server 110 of this embodiment produces | generates. The figure which showed embodiment of the personal search system 1200 of this embodiment. FIG. 11 is a sequence diagram for a transaction between a client 1210 and a search server 1270.

Explanation of symbols

DESCRIPTION OF SYMBOLS 100 ... Individual search system, 110 ... Search server, 112 Network interface part, 114 ... Input interface part, 116 ... Output interface part, 120 ... Information integration management module, 122 ... Document management part, 124 ... Context management part, 126 ... Individual Information management section, 128 ... schedule management section, 130 ... reviewer ranking section, 132 ... data management section, 140 ... database, 150 ... network

Claims (10)

A personal search system for reviewing produced documents, the search system comprising:
A search server connected to the network and receiving a search request for requesting said review from a client computer;
The search server
A database for registering ready-made documents created in the past and index data for indexing the contents of the ready-made documents;
Document management means for calculating the similarity of the ready-made document stored in the database with respect to the search string included in the search request and extracting individuals related to similar similar documents with respect to the search string;
Personal information management means for calculating a matching score for the attribute value of the individual's schedule, skill level, role, department,
The personal identification information related to the similar document having a high similarity to the search string among the ready-made documents is registered in the reviewer set list, and is set in the context management means from each matching score of the related individual A personal ranking system for calculating a recommended rank as a reviewer for the individual according to a context.   The personal search system according to claim 1, wherein the context management unit manages a score function including the matching score as a variable.   The personal ranking means extracts and ranks the identification information by equivalently processing an attribute value of a similar document creation history or a past review history of the similar document with respect to the ready-made document. The personal search system described in.   The personal search system according to any one of claims 1 to 3, wherein the personal search system is implemented in a search server as reviewer determination means for creating a document. In order to search for an individual to review a produced document, an information processing method executed by an information processing device, the information processing device comprising:
Receiving a search request of an individual connected to the network and requesting said review;
Extracting a search string included in the search request from the search request;
For a ready-made document stored in a database, calculating a similarity to the search string;
A reviewer set list by extracting identification information of individuals who have created the similar document or have reviewed in the past, with respect to the similar document in which the similarity exceeds the set similarity threshold among the ready-made documents Registering with
For the individual registered in the reviewer set list, calculating a matching score sequentially from each attribute value of a schedule, skill level, role, and department;
Recursively calculating a recommended ranking as a reviewer for the individual according to the context set from the matching score of the attribute value.   The step of recursively generating the recommended ranking includes setting a total score determined by setting all the matching scores of the attribute values used for score calculation as a recommended score, and using the value of the recommended score. The personal search method according to claim 5, further comprising: ranking the individual.   The personal search according to claim 5 or 6, wherein the registering step includes a step of extracting the identification information of the individual by using an attribute value of a creation history or a review history of a similar document with respect to the ready-made document. Method.   The personal search method according to claim 5, wherein the personal search method is executed by the information processing apparatus to determine a reviewer for document creation.   The program which can perform information processing for an information processor to perform each step of any one of Claims 5-8.   An information processing apparatus-readable recording medium on which the program according to claim 9 is recorded.

Images