JP2005250701A - Information visualization system, method and program - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide an information visualization system that visualizes the position of each person to certain information according to the activities of each person to the information. <P>SOLUTION: The information visualization system 1 has a database 10 for databasing Web contents on a Web bulletin board, a member profile creation part 22 for creating profiles of members (subjects) by referring to the contents in the database 10, a content profile creation part 24 for creating profiles of the contents in the database 10, a member object data creation part 30 for creating data representing relevance between the members and the contents by comparing the member profiles and the content profiles, and a data display part 40 for displaying, about certain content, relative positional relations between the content and the members according to the created data. <P>COPYRIGHT: (C)2005,JPO&NCIPI

Description

  The present invention relates to an information visualization system, method, and program, and more particularly, to an information visualization system, method, and program for visualizing people related to certain information.

  In order to find people related to certain information in communities such as organizations and companies, indicators such as expertise and interest of each person are set in advance by hand, and the indicators and information contents are specified. A technique for listing people in the order of association by matching between search keys is common.

  Further, Patent Document 1 discloses an invention for displaying knowledge and information utilization status in an organization in an easy-to-understand manner. In the present invention, each category of “respondent” of the questionnaire, “knowledge / information” used by the person, and “source of acquisition” of the knowledge / information is assigned to each circle of three center of gravity concentric circles.・ A node representing the information and source is placed in the category to which the node belongs, and an arc connecting the respondent node and the knowledge / information node used by the respondent is displayed. Yes.

JP 2003-281320 A

  However, conventional methods for finding people related to information have the following problems. In other words, it takes time to manually input each person's index, the point that the determination of the index is arbitrary, the point that it takes time to determine the keyword and specify it as a search key, and the indexed search results There are problems such as the difficulty of intuitively grasping the relationship from the list of people. On the other hand, although patent document 1 displays the utilization status of information in an easy-to-understand manner, no solution is presented or suggested for these problems.

  The present invention solves the above-described conventional problems and provides an information visualization system, method, and program for visualizing each person's position with respect to certain information based on each person's activity (activity) with respect to the information. For the purpose.

  The information visualization system according to the present invention is a database creation means for creating a database of contents including information reflecting human activities, and refers to the contents created in the database, and analyzes the activities performed on the contents by the subject, Subject profile creation means for creating a profile of the subject based on the analysis result, content profile creation means for analyzing content stored in the database, and creating a content profile based on the analysis result, and subject by the subject profile creation means Data creation means that compares the profile with the content profile created by the content profile creation means and creates data indicating the relationship between the subject and the content, and the content is based on the data created by the data creation means And a display means for displaying the relative positional relationship between the content and subject to. By displaying the relative positional relationship between the content and the subject based on the content, it is possible to intuitively grasp the subject related to the content.

  Preferably, the information visualization system includes time span setting means for extracting content for a certain period from the database means, and the subject profile creation means and the content profile creation means refer to the extracted content and the profile of the subject and the content. Create Thereby, the relationship between the content and the subject within a period desired by the user can be visualized.

  Preferably, when each profile for a plurality of contents is created by the content profile creation means, the display means determines the relative positional relationship between the content and the subject with reference to the content selected from the plurality of contents. indicate. Thereby, the relationship between each of the plurality of contents and the subject can be visualized.

  Preferably, when a plurality of content profiles are synthesized by the content profile creation means, the data creation means compares the synthesized content profile with the subject profile, creates data indicating the relevance of both, and the display means The relative positional relationship between the content and the subject is displayed on the basis of the synthesized content. Thereby, the test subject relevant to this can be visualized from the synthesized viewpoint of a plurality of contents.

  For example, the database means includes an electronic document as content, and the subject profile includes a word appearance frequency vector in the electronic document viewed by the subject and a word appearance frequency vector in the electronic document written by the subject, The profile includes a word appearance frequency vector in the digitized document, and the data creation means calculates the relevance by calculating the inner product of the word appearance frequency vector for browsing and writing the subject profile and the word appearance frequency vector of the content profile. . The display means displays the content at the reference point of the coordinate axis, the vertical axis passing through the reference point is related to the content read or written by the subject, and the horizontal axis is related to the content written or read by the subject. , Display subject in coordinates.

  Here, the “content” may be information that reflects people's activities and can be made into a database, and includes, for example, a document, a file, a web page, and the like created using an electronic device such as a computer. The “subject” is a person related to the content and is visualized by the information visualization system, and corresponds to the “member” in the embodiment.

  The information visualization method according to the present invention refers to a database for content including information reflecting human activity, analyzes activities performed on the content by the subject, and creates a profile of the subject based on the analysis result. A first profile creation step, a second profile creation means for analyzing the contents of the database and creating a profile of the content based on the analysis result, a subject profile and a second profile creation step by the first profile creation step A data creation step for comparing the content profiles of the subject and the data to create the data indicating the relationship between the subject and the content, and based on the data created by the data creation step, Display positional relationship And a step shows.

  The information visualization program according to the present invention refers to a database for content including information reflecting human activities, analyzes activities performed by the subject on the content, and creates a profile of the subject based on the analysis result A first profile creation step, a second profile creation means for analyzing the contents of the database and creating a profile of the content based on the analysis result, a subject profile and a second profile creation step by the first profile creation step A data creation step for comparing the content profiles of the subject and the data to create the data indicating the relationship between the subject and the content, and based on the data created by the data creation step, Show positional relationship And a display step of.

  According to the information visualization system, method, and program according to the present invention, a subject profile and a content profile are created using a database, and their relevance is obtained and visualized. It is possible to save labor for specifying an index and a search key, and to provide an efficient visualization system.

  Furthermore, the visualization technology of the present invention makes it possible to find people intuitively with a simple operation when finding people related to certain information in the community, thereby providing an opportunity for highly productive communication. From the viewpoint of information acquisition and information provision, by visualizing the distance between information and people, it is possible to intuitively understand how each subject interacts with the information, and with simple procedures without being aware of search. Find related people across a variety of information.

  The information visualization system according to the present invention is preferably implemented in a network system using a personal computer or the like. Hereinafter, embodiments will be described with reference to the drawings.

  FIG. 1 is a block diagram illustrating the configuration of the information visualization system according to the present embodiment. In the figure, the information visualization system 1 creates a profile of members (subjects) and contents by referring to an activity database 10 that stores contents including information reflecting human activities, and contents stored in the database 10. The profile creation unit 20 that compares the profile created by the profile creation unit 20 and the member / object data creation unit 30 that visualizes the relationship between members and content, and the data created by the data creation unit 30 are displayed. The data display unit 40 includes a content setting unit 50 and a time span setting unit 60 that set a profile creation condition for the profile creation unit 20.

  The profile creation unit 20 analyzes the activity performed by the member on the content, analyzes the content stored in the database 10 and the member profile creation unit 22 that creates the profile of the observer based on the analysis result, Content profile creation means 24 for creating a content profile based on the analysis result.

  The member / object data creation unit 30 compares the member profile by the member creation profile unit 22 with the content profile by the content profile creation unit 24 and creates data indicating the relationship between the member and the content.

  Based on the data created by the member / object data creation unit 30, the data display unit 40 displays the relationship between the content and the member in a relative positional relationship based on the content.

  The content setting unit 50 is an input for setting the content created by the content profile creation unit 24. For example, the content setting unit 50 provides an input for selecting a desired content from the content extracted from the database 10. Also good. The time span setting unit 60 sets the period of the profile created by the profile creation unit 20. That is, content for a certain period is cut out from the database 10 and a profile is created in that range. The time span may preferably be set by the user.

  The database 10 of the present embodiment uses, as an example, a web bulletin board that can be browsed and written by a network such as the Internet as content. In the web bulletin board, a plurality of people access a web page, browse or write, and the activities are accumulated to create a specific web page. It is assumed that the web bulletin board includes “Java (registered trademark)” and “XML” pages. The database 10 records an activity log (history) for each page of the web bulletin board, and constitutes a database.

  FIG. 2 is a diagram for explaining the data structure of the activity database 10. The data includes the following fields: Time of activity (access) to the web bulletin board, person who acted (actor), page acted, activity type (action), content of activity (web content), content written (additional text), and remarks Including the opponent. The type of activity on the web bulletin board is either browsing or writing. When text is written, additional text is generated. The speaking partner is a person who is the target of writing when writing to the web content. For example, in FIG. 2, when actor D writes, the content of the writing is extracted in the additional text, and the writing A is the speaking partner, and a conversation is established between D and A. It is regarded as a thing.

  With reference to such content data stored in the database 10, the profile creation unit 20 and the data creation unit 30 perform an operation for visualizing the relationship between members and content.

  Next, the operation of the member profile creation unit 22 will be described using the flowchart of FIG. First, an arbitrary time span is set from the time span setting unit 60 shown in FIG. 1, and content data for a period corresponding to the time span is cut out from the activity database 10 (step S101).

  Next, by searching for actors in the database, members related to the web bulletin board are set (step S102). In this case, all member lists may be displayed on the display, and any member may be set by the user.

  After setting the members, one member is selected from them (steps S103 and S104). For example, when member A is selected from the members, the text viewed by member A is extracted within the time span (step S105). That is, the action corresponding to the actor “A” is selected from the data of FIG. 2, “Browse” is identified from the action, and the text data of the web content is extracted.

  The extracted text is subjected to morphological analysis, a word is extracted therefrom, and a word appearance frequency vector included in the browsed web content is created (step S106). For example, as shown in FIG. 4, the appearance frequencies (n1, n2,... N) of the words W1, W2, W3,. Is generated by a predetermined algorithm.

  In parallel with this operation, the text written by member A who wrote in the time span is extracted (step S107). The action corresponding to the actor “A” is selected from the data of FIG. 2, “Write” is identified from the action, and the text data written by A in the additional text is extracted.

  The extracted text is subjected to morphological analysis, the word is extracted, and a word appearance frequency vector included in the additional text is created (step S108). This is done in the same manner as in the above browsing. The

  The word appearance frequency vector relating to the browsed text and the word appearance frequency vector relating to the written text are stored together with the name of the member A (step S109). By this work, the analysis of the content of the member A acting on the web content within the time span is performed.

  Similar processing is performed for the other members, and profiles for all the set members are created and stored.

  Next, the operation of the content profile creation unit 24 will be described with reference to the flowchart of FIG. Web content data included in the time span is cut out from the database 10 (step S201). A content list is created from the cut web content data (step S202). The web bulletin board shown in FIG. 2 includes a JAVA (registered trademark) page and an XML page, and both these pages serve as a content list. Of course, when the target content includes a plurality of documents and a plurality of files, each document or file constitutes a content list.

  Next, one content is selected from the content list (step S203, step S204). For example, if an XML page is selected as content, text included in the XML page is extracted from the database. Then, a morphological analysis is performed on the text to extract words, a word appearance frequency vector is created (step S205), and this is stored (step s206).

  The processing from step S204 to S206 is performed for all the contents included in the content list, and then the stored word appearance frequency vectors are added together to create a content profile (step S207).

  Next, the operation of the member / object data creation unit 30 will be described with reference to the flowchart of FIG. The member / object data creation unit 30 compares the member profile and the content profile from the member profile creation unit 22 and the content profile creation unit 24, and calculates the relevance or difference between the two profiles.

  Specifically, the degree of profile difference is obtained by calculating the inner product of the word appearance frequency vector when browsing the member profile and the word appearance frequency vector included in the content profile (step S301). Similarly, the degree of profile difference is obtained by calculating the inner product of the word appearance frequency vector at the time of writing the member profile and the word appearance frequency vector included in the content profile (step S302).

  As shown in FIG. 4, the word appearance frequency vector includes a word and a component of the number of appearances. The smaller the inner product angle between the two, the more common both profiles, and the larger the angle. The greater the difference, the greater the difference between both profiles.

  Member object data is created based on the profile dissimilarity obtained in steps S301 and S302, and provided to the data display unit 40 (step S303).

  FIG. 7 shows a basic display function by the data display unit 40. This figure is an example in which the degree of association between content and members involved in the content is visualized. The content is displayed at the origin of coordinates, the vertical axis indicates the difference between the content and the content viewed by the member (calculated in step S301 in FIG. 6), and the horizontal axis indicates the difference between the content and the content written by the member. (Calculated in step S302 of FIG. 6). In the field in the coordinates, members M1 to M5 are displayed by icons, and approach the origin if the degree of coincidence or relevance with the origin content is high. For example, the member M1 can see that the browsing content and the writing content are the closest to the origin content. The member M2 is close in writing content, but the browsing content is very low in coincidence. The member M5 can grasp at a glance that the browsing content is very close, but the writing content is low in coincidence.

  The data display unit 40 further displays a content list display unit 52 adjacent to the coordinates. By dragging the content included in the content list display unit 52 to the origin, the content at the origin can be changed. The content set as the origin may be a single content or a plurality of content. When a plurality of contents are arranged at the origin, the relevance with members for the plurality of contents is visualized. When a single content is set as the origin from a plurality of content lists, the member / object data creation unit 30 uses the word appearance frequency vector of the single content stored in step S206 of FIG. The degree of difference is calculated and the calculation result is displayed.

  The data display unit 40 further displays a time span setting unit 62. For example, an arbitrary period (time span) can be designated by operating the time span setting unit 62 and sliding from the past to the present. In this case, the member profile creation unit 22 and the content profile creation unit 24 create a profile that responds to the period set by the time span setting unit 62.

  FIG. 8A shows an example of visualizing the degree of relevance of each member to the content when “JAVA (registered trademark) page” is used as the content. FIG. 8B is an example in which the relevance level of each member is visualized when “XML page” is further synthesized as content. When content is synthesized, the profile creation unit 20 and the data creation unit 30 operate accordingly.

  The preferred embodiments of the present invention have been described in detail above. However, the present invention is not limited to the specific embodiments, and various modifications can be made within the scope of the gist of the present invention described in the claims. Deformation / change is possible.

  In the above-described embodiment, the web bulletin board is used as content including information reflecting human activities. However, this is an example, and content such as other document files or web files may be used. Furthermore, the display of the icon by the data display unit is also an example, and other than this, the display may be able to recognize the relationship between the member and the content. Further, such display may include text information, audio information, and video information.

  The information visualization system, method, and program according to the present invention can be used to search for people related to certain information in the community such as an organization or a company, and can be used for analysis, evaluation, management, etc. of information and people, or It can be used as a system for communication with people.

It is a block diagram which shows the structure of the information visualization system which concerns on the Example of this invention. It is a figure explaining the data structure of an activity database. It is a flowchart explaining operation | movement of a member profile preparation part. It is a figure explaining a word appearance frequency vector. It is a flowchart explaining operation | movement of a content profile creation part. It is a flowchart explaining operation | movement of a member object data preparation part. It is the example of a display which visualized the relationship between a content and a member. FIG. 8A is an example of visualizing the relationship with a member when “JAVA (registered trademark) page” is set as the origin as the content, and FIG. 8 (b) is further synthesized using “XML page” as the origin. This is an example of visualizing the relationship with the member at the time.

Explanation of symbols

1: Information visualization system 10: Activity database 20: Profile creation unit 22: Member (subject) profile creation unit 24: Content profile creation unit 30: Member / object data creation unit 40: Data display unit 50: Observer setting unit 60: Time span setting section

Claims (11)

A database creation means for creating a database of contents including information reflecting human activities;
Subject profile creation means for referring to the databased content, analyzing the activities performed by the subject on the content, and creating a profile of the subject based on the analysis result;
A content profile creating means for analyzing the databased content and creating a profile of the content based on the analysis result;
A data creation means for comparing the subject profile by the subject profile creation means and the content profile by the content profile creation means, and creating data indicating the relationship between the subject and the content;
Display means for displaying the relative positional relationship between the content and the subject based on the data created by the data creation means;
An information visualization system. The information visualization system includes time span setting means for extracting content for a certain period from the database means, and the subject profile creation means and the content profile creation means create a subject and content profile with reference to the extracted content. The information visualization system according to claim 1. When each profile for a plurality of contents is created by the content profile creation means, the display means displays the relative positional relationship between the content and the subject with reference to the content selected from the plurality of contents. The information visualization system according to claim 1 or 2. When a plurality of content profiles are synthesized by the content profile creation means, the data creation means compares the synthesized content profile with the subject profile, creates data indicating the relationship between both profiles, and the display means is synthesized. The information visualization system according to any one of claims 1 to 3, wherein a relative positional relationship between the content and the subject is displayed with reference to the content. The database means includes an electronic document as content, the subject profile includes a word appearance frequency vector in the electronic document viewed by the subject and a word appearance frequency vector in the electronic document written by the subject, and the content profile is Including the word appearance frequency vector in the digitized document, the data creation means calculates and displays the relevance by calculating the inner product of the word appearance frequency vector of the browsing and writing of the subject profile and the word appearance frequency vector of the content profile The means displays the content at the reference point of the coordinate axis, the vertical axis passing through the reference point is related to the content viewed or written by the subject, the horizontal axis is related to the content written or viewed by the subject, The subject is displayed in the coordinates, and there is no claim 1 4 Visualization System according to any one. A first profile creating step of referring to a database including content reflecting information of a person's activity, analyzing an activity performed by the subject on the content, and creating a profile of the subject based on the analysis result;
A second profile creating means for analyzing the contents of the database and creating a profile of the contents based on the analysis result;
A data creation step of comparing the subject profile by the first profile creation step with the content profile by the second profile creation step and creating data indicating the relationship between the subject and the content;
A display step of displaying a relative positional relationship between the content and the subject based on the content based on the data created by the data creation step;
An information visualization method comprising: The information visualization method further includes a step of cutting out content for a certain period from the database, and the first and second profile creation steps create a profile of the subject and the content with reference to the extracted content. 6. The information visualization method according to 6. When each profile for a plurality of contents is created by the second profile creation step, the display step displays the relative positional relationship between the contents and the subject with reference to the content selected from the plurality of contents. The information visualization method according to claim 6 or 7. When a plurality of content profiles are synthesized by the second profile creation step, the data creation step compares the synthesized content profile with the subject profile, creates data indicating the relationship between both profiles, and the display step includes The information visualization method according to claim 6, wherein a relative positional relationship between the content and the subject is displayed on the basis of the synthesized content. The database includes an electronic document as content, the subject profile includes a word appearance frequency vector in the electronic document viewed by the subject and a word appearance frequency vector in the electronic document written by the subject, and the content profile is an electronic document. The word creation frequency vector in the document, and the data creation step calculates the relevance by calculating the inner product of the word appearance frequency vector of the browsing and writing of the subject profile and the word appearance frequency vector of the content profile, and the display step Displays the content at the reference point of the coordinate axis, the vertical axis passing through the reference point is the relevance with the content viewed or written by the subject, the horizontal axis is the relevance with the content written or viewed by the subject, and the coordinates The subject is displayed in the inside. Visualization method according to one or stone 9. A first profile creating step of referring to a database including content reflecting information of a person's activity, analyzing an activity performed by the subject on the content, and creating a profile of the subject based on the analysis result;
A second profile creating means for analyzing the contents of the database and creating a profile of the contents based on the analysis result;
A data creation step of comparing the subject profile by the first profile creation step with the content profile by the second profile creation step and creating data indicating the relationship between the subject and the content;
A display step of displaying a relative positional relationship between the content and the subject based on the content based on the data created by the data creation step;
An information visualization program having

Images