JP2003337933A - Knowledge data processor - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To solve the problem that it is difficult to rearrange and use information as knowledge which is naturally accumulated. <P>SOLUTION: In a knowledge data processor 10, an indication acquisition part 12 acquires an operation instruction from a user, a management unit 20 manages a group of accumulated knowledge data, and a processing unit 30 extracts knowledge data of interest from the management unit 20 and displays them on a monitor 40. A relation detection part 34 detects latent relativity between pieces of knowledge data according to a viewpoint specified by the user and a data extraction part 36 extracts a chain of knowledge data having relativity. A display processing part 38 visualizes them in a desired form to make the latent relativity visual. <P>COPYRIGHT: (C)2004,JPO

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a knowledge data processing device. The present invention particularly relates to a technique for extracting desired information from a vast amount of knowledge data stored in a computer.

[0002]

2. Description of the Related Art Conventionally, when people write their thoughts and creations in the form of papers or novels, it is common to gradually complete the text by repeating drafts and making partial corrections. It was In that case, I had to finish writing the sentence in the length of a phrase in my head and then actually write it down. With the spread of personal computers, people gradually changed to a method of completing a sentence by typing the sentence without thinking about the final form of the completed sentence and then editing or correcting it while checking it on the screen. It was As a text creation tool, in addition to a word processor, an idea processor, which is an auxiliary tool for solving a problem or developing an idea, is used. The idea processor is equipped with an outline function that gives a hierarchical association according to the user's intention to ideas and memos that the user types as he or she thinks. These tools are used not only for academic purposes such as writing papers, but also for situations such as business document creation.

[0003]

In the idea processor, it is the user himself who devises individual ideas that are elements. Further, the work of organizing and associating these ideas in a bottom-up manner is also performed manually at the user's discretion. And the final purpose is to make up and correct these organized ideas,
It is to complete one document by fleshing.

The inventor of the present invention has come to recognize that a method such as "idea organization" among the functions of the idea processor is effective for applications other than the document creation support. In other words, adopting "structured" as a means of organizing ideas is effective in the sense that it directly appeals to the human visual sense, and if a variation is provided in the visualizing means, it will further stimulate the human senses. The inventor has paid attention to the fact that this can be done.

The present inventor has made the present invention based on the above recognition, and an object of the present invention is to provide a technique for automatically organizing complex and diversified data. Another object of the present invention is to provide a technique for efficiently managing a huge amount of data. Still another object is to provide a technology that supports the construction of a knowledge database. Still another object is to provide a technology for efficiently utilizing the accumulated information.

[0006]

One aspect of the present invention relates to a knowledge data processing device. This apparatus is a data storage unit that stores a plurality of knowledge data and a viewpoint determination that acquires a designation of the viewpoint from a user in order to detect a potential mutual relation of the plurality of knowledge data from a desired viewpoint. Section, a relationship detection section that detects the relationship from the acquired viewpoint, a data extraction section that extracts a series of knowledge data chained by the detected relationship, and the state of the chain is visualized in a space of a desired dimension. A display processing unit for displaying the series of knowledge data in the selected format.

The "knowledge data" corresponds to a document file generated by application software such as a word processor, presentation software, an electronic mail client, and a scheduler. A data file mainly containing a text is assumed in the sense that these are targets of the language analysis processing, but the file is not necessarily limited to the text data file, and may be an image file, a CAD data file, or the like. The “data storage unit” mainly indicates a storage device such as a hard disk. The apparatus may be configured such that the data storage unit is remotely connected via the network. A data group including a plurality of pieces of knowledge data may be centrally stored in one place or may be distributed and stored in a plurality of places. It is not necessary to perform a special process such as indexing on these knowledge data, and in that sense, the storage position and storage state of the knowledge data can assume various forms. However, the knowledge data may be indexed.

The "potential mutual relevance" is revealed only when a plurality of pieces of knowledge data that can be classified and ordered from a certain point of view are classified and ordered. It is the relationship between knowledge data. The “desired viewpoint” is a guideline or standard for extracting a series of knowledge data that the user wants to extract from a plurality of pieces of knowledge data. For example, the point of interest for finding the upper and lower relations of the concept, the nature of the contents common to a plurality of pieces of knowledge data, the items referred to in order determination such as the creation date and time, and the like are applicable. The “designation of viewpoint” may be a word designation from the user, or may be a form in which the user selects one of several viewpoint options prepared in advance.

The "chained series of knowledge data" refers to a data group composed of a plurality of knowledge data in which a chained hierarchical relationship is found due to relationships such as upper and lower levels when viewed from the above viewpoint. The “space of desired dimension” is, for example, 1
A three-dimensional space, a two-dimensional space, and a three-dimensional space are shown, and a designation of which format to visualize the chained state is acquired from the user. The “display processing unit” may change the dimension of the space for visualizing the state of the chain based on an instruction from the user.

Another aspect of the present invention also relates to a knowledge data processing device. This device is a viewpoint determination unit that acquires a designation of a viewpoint from a user in order to reveal a latent relationship among a plurality of accumulated knowledge data from a desired viewpoint, and the relationship from the acquired viewpoint. The relation detecting unit to detect, the data extracting unit to extract the knowledge data group focused on by the user based on the detected relation, and the structure of the comprehensive correlation in the knowledge data group to be noticed are constructed by a graph and And a display processing unit for visualizing.

"Knowledge data of interest" indicates a knowledge data group related to a desired theme that the user intends to extract. "Comprehensive correlation structure" is, for example, a relationship between data that emerges when a focused knowledge data group is viewed from a certain point of view, and a network-like structure that covers the entire knowledge data group. . "Graph"
Is a relational structure composed of a plurality of points and lines connecting them.

Any combination of the above components,
It is also effective as an aspect of the present invention that the components and expressions of the present invention are mutually replaced among a method, a device, a system, a computer program, a recording medium storing the computer program, a data structure, and the like.

[0013]

BEST MODE FOR CARRYING OUT THE INVENTION The knowledge data processing apparatus according to the present embodiment extracts a desired data group from a viewpoint (view: view) desired by a user from a vast amount of accumulated knowledge data, and finds out from that viewpoint. Visually represent the relationship between knowledge data As the number of pieces of knowledge data increases, their correlation changes from centralized to complex and diverse. From a different point of view, a completely different relationship may appear between the same knowledge data. Focusing on such a point, as a criterion for revealing a latent relationship between knowledge data, a user specifies a certain limited viewpoint necessary for data extraction, and the data extraction is processed from that viewpoint. This allows the user, that is, the owner of the knowledge data, to easily and quickly find out the necessary information without specially organizing them. However, a series of extracted knowledge data may be unnecessarily large depending on the concept of the specified viewpoint. Therefore, the user may specify the starting point (or viewpoint) of the information to be extracted. If only the knowledge data that is chained with this starting point as the root is extracted, the desired information will not be buried in unnecessary information.

When the user looks at the relationship between the organized data, the knowledge that should be inherent and the points that should be fleshed out between the data are highlighted, and it functions as a new knowledge creation auxiliary tool. A latent relationship between knowledge data is also a piece of knowledge, but it is information that belongs to so-called tacit knowledge.
By visualizing this, we change it into explicit knowledge, which is explicit knowledge. The change can also be called the creation of new knowledge.

Here, in the case of organizing ideas using the idea processor or the outline processor, the object of association is the content newly typed by the user, and the user himself also performs the association work. On the other hand, this knowledge data processing device can take in already accumulated various files as knowledge data, and their association is automatically processed by the language analysis process, and the relation is put in a format that is easy for the user to understand. To visualize. Note that desired information can be retrieved from the existing information group by a search engine based on keyword search. However, the search result is only an enumeration in score order according to the appearance frequency of keywords, etc., and it is not possible to obtain the relevance between the extracted information and the relation structure thereof.

The knowledge data processing device is realized by combining an application program having a function of extracting and displaying knowledge data and a computer on which the application program is executed. The function of this device is as illustrated below. In other words, the function of importing multiple existing files as knowledge data, the function of accumulating knowledge data, the function of acquiring an instruction from the user to specify the viewpoint for extracting a desired information group from the accumulated knowledge data, the user specifying Function to detect the relationship between knowledge data from the viewpoint of the above, function to linguistically analyze the contents of knowledge data to detect the relationship, function to retain rules for linguistic analysis, edit the rule Functions, such as a function to extract a series of knowledge data groups in which detected relationships are linked, a function to visualize and display the extracted knowledge data groups and their relationships, and a function to edit knowledge data. is there. The present knowledge data processing device is configured by organizing all or any of these functions and other functions in an arbitrary structure. The embodiment described below is an example of a combination of the above functions.

FIG. 1 is a functional block diagram showing the basic configuration of the knowledge data processing device. Knowledge data processing device 10
Includes an instruction acquisition unit 12, a management unit 20, and a processing unit 30. A monitor 40 is connected to the knowledge data processing device 10. Each block included in the knowledge data processing device 10 can be realized by an element such as a CPU of a computer in terms of hardware, and can be realized by a program having a data extracting function and a data displaying function in terms of software. However, FIG. 1 depicts the functional blocks realized by the cooperation thereof. Therefore, these functional blocks can be realized in various forms by combining hardware and software. The entire system may be constructed in such a manner that each of these blocks is distributed via a network.

The instruction acquisition unit 12 acquires an instruction regarding the processing content from the user and notifies each block. The management unit 20 has a function of managing a plurality of pieces of knowledge data to be processed, and also has an input processing unit 22 and a data storage unit 24 to realize the function. The input processing unit 22 inputs knowledge data from outside the knowledge data processing device 10. The input processing unit 22 may have, for example, a function of capturing text data included in a web page on the Internet, a function of capturing knowledge data stored in a file server on a LAN (Local Area Network), and the like. The instruction to select the knowledge data to be fetched is acquired from the user via the instruction acquisition unit 12. The data storage unit 24 stores a plurality of pieces of knowledge data to be processed.

The processing unit 30 has a function of performing a predetermined process on the knowledge data, and in order to realize the function, the viewpoint determining unit 32, the relation detecting unit 34, the data extracting unit 36, the display processing unit 38, And an edit processing unit 39.

The viewpoint determining unit 32 acquires the designation of the viewpoint from the user via the instruction acquiring unit 12 in order to detect the potential mutual relations possessed by the plurality of pieces of knowledge data from the desired viewpoint. The latent relation between multiple pieces of knowledge data is revealed from the specified viewpoint. Viewpoint determination unit 32
The user may be presented with options of viewpoints registered in advance, and the user may be instructed to select one of the options. The viewpoint determination unit 32 may receive the designation of the viewpoint from the user in the form of natural sentence input.

The relation detecting section 34 detects the relation between a plurality of pieces of knowledge data from the viewpoint specified by the user. The relation detection unit 34 may have a function of performing a language analysis process on the content of knowledge data. The functions of language analysis processing include, for example, a morpheme analysis function, a phrase extraction function, a vocabulary statistical processing function, a related word extraction function, a concept search function, a document summarization function, and a document classification function.

The relation detecting section 34 may hold a rule defined for deriving the relation from the viewpoint specified by the user. This rule is defined in order to detect an organic and appropriate connection between knowledge data. For example, when detecting relevance from the perspective of the subject such as the sender and the recipient of an email, relevance with a common sender name, relevance with a common sender address, and reciprocal designation as multiple recipients. A rule for detecting sex or the like may be set.
For example, even if rules are set to detect relationships such as cause and effect, relationship between master-slave relationship and hierarchical relationship based on proper name, relationship between template document and document created based on the template, etc. Good. When the viewpoint determination unit 32 presents the viewpoint option to the user, the option and the rule may be associated with each other in advance.

The data extraction unit 36 extracts from the data storage unit 24 a series of knowledge data linked by the relationships detected by the relationship detection unit 34. This allows the user to selectively extract the knowledge data group of interest. Display processing unit 38
Causes the monitor 40 to display a chained state of a series of knowledge data in a format visualized in a space of a desired dimension. For example,
Build and visualize a comprehensive correlation in the knowledge data group of interest with a graph structure.

The edit processing unit 39 processes addition of knowledge data to the data storage unit 24 based on an instruction from the user. At this time, the relation detection unit 34 detects the relation between the added knowledge data and the existing knowledge data in the data storage unit 24. The added knowledge data is displayed in a format in which the detected relevance is visualized.

FIG. 2 shows a plurality of pieces of knowledge data and their relations in the form of a directed graph. Each of the plurality of pieces of knowledge data is displayed as a node, and the relationship between the pieces of knowledge data is displayed as a branch. As shown in the figure, the line type of the branch, the thickness of the branch, the direction of the arrow, and the like may be changed according to the type and strength of the relationship. In this way, a plurality of pieces of knowledge data form a knowledge set as a whole in a state of having various relationships with each other. This set may be referred to as the “Knowledge Sea”.

Node 50 has two children, node 5
2, 54 are connected by branches 62, 64. Node 52
Has two children nodes 56 and 58 in branch 6
It is tied at 6,68. A node 60, which is a child, is connected to the node 54 by a branch 70. For example, node 56
Has a node 90 in addition to the node 52 as a parent, and each knowledge data has a complicated relationship of having a plurality of parents and children. These networks are represented by a two-dimensional graph as shown in the figure, but when the number of knowledge data to be handled further increases, for example, they may be represented by a three-dimensional graph with virtual depth added. .

Here, when the user specifies the viewpoint of "company", the nodes 50, 52, 54, 56, 58, 6 are selected.
0 is extracted as the attention knowledge data group. If the user specifies the node 52 as the starting point, the node 52,
Only 56 and 58 are extracted as the attention knowledge data group.
Similarly, when the user specifies the viewpoint of “project” and the start point of the node 80, the node 80 and its child nodes 82, 84, 88, 56, 60, 90.
Then, a node 86, which is a child of the node 84, is extracted as a focused knowledge data group. In this way, the knowledge data to be extracted also changes if the viewpoint is changed. Not only is it possible to connect a large number of pieces of knowledge data in a mesh pattern, but also to make it easy to retrieve the knowledge of interest, which facilitates editing by the user.

FIG. 3 shows a plurality of pieces of knowledge data extracted as a focused knowledge data group and their relationships in a tree structure. Here, the viewpoint of “project” and the attention knowledge data group extracted when the starting point of the node 80 in FIG. 2 is designated by the user are shown. The knowledge data 100 having the content of “meeting 1” becomes the parent, and the knowledge data 102, 104, 108, 11 that are children in the hierarchy one level below the parent.
0, 112, 114 are associated. Knowledge data 10
Knowledge data 106, which is a child of the hierarchy, is associated with the hierarchy below 4. The correlation in these knowledge data groups is visualized in a one-dimensional or two-dimensional space with a hierarchical tree structure. In general, many people who use personal computers on a daily basis are familiar with such a layered file structure. Therefore, many users can easily understand the relationship between the knowledge data by the tree structure shown in FIG.

FIG. 4 shows a screen configuration when the correlation between a plurality of pieces of knowledge data and the contents of each piece of knowledge data are displayed on the monitor 40. The window 120 mainly has a structure display area 122 and a document display area 126. Structure display area 122
Displays the relational structure of a plurality of knowledge data as shown in FIG. The content of the extracted knowledge data is displayed in the document display area 126.

When the user clicks the net display button 130, the structure display area 122 displays the relational structure of the knowledge data visualized in the network format as shown in FIG. When the user clicks the tree display button 132, the relational structure of the knowledge data visualized in the tree structure as shown in FIG. 3 is displayed in the structure display area 122.

The date tab 134, company tab 136, and project tab 138 are switches for the user to specify a desired viewpoint. In Figure 4, the user has a company tab 1
The state after clicking 36, the structure display area 12
In the graph displayed in 2, the node 5 is the knowledge data group whose relevance is found from the viewpoint of "company".
0, 52, 54, 56, 58, 60 and their branches 62, 64, 66, 68, 70 are highlighted.
The user sets the pointer 12 to the starting point of the knowledge data group to be extracted.
It can also be specified by 4. In the figure, the node 50 is designated as the starting point.

In the document display area 126, the contents of knowledge data "company 1" corresponding to the node 50 are displayed. In the fields below it, a memo button 140, a mail button 142, a schedule button 144, and a link button 14
6. A drawing button 148 is displayed. By clicking these, information such as a memo can be inserted at the cursor position of the text displayed in the document display area 126. For example, the drawing 150, the memo 152, and the link 154 in the text of the document display area 126 are objects inserted by pressing the drawing button 148, the memo button 140, and the link button 146, respectively.

FIG. 5 shows a screen configuration when the correlation between a plurality of pieces of knowledge data and the contents of each piece of knowledge data are displayed on the monitor 40. This figure shows the tree display button 13 from the state of FIG.
2 shows the state after clicking 2, and the structure display area 122
Shows the relationship between a plurality of visualized knowledge data in the form of a tree structure.

FIG. 6 shows a screen configuration when the correlation between a plurality of pieces of knowledge data and the contents of each piece of knowledge data are displayed on the monitor 40. This figure shows the tree display button 13 from the state of FIG.
2 and the state after clicking the project tab 138, or the state after clicking the project tab 138 from the state of FIG. In the structure display area 122,
The relationship of a series of knowledge data from the viewpoint of "project" is visualized and displayed in the form of a tree structure.
In the document display area 126, the contents of the series of knowledge data displayed in the structure display area 122 are displayed in a one-dimensionally linked arrangement. The text is enlarged or reduced according to a predetermined operation by the user. Such a one-dimensional chained arrangement has a lot in common with the page structure of a word processor document and the link structure of a web page, and is a format familiar to users. Viewing, adding, and editing knowledge is also easy for the user.

FIG. 7 shows a screen configuration when the correlation between a plurality of pieces of knowledge data and the contents of each piece of knowledge data are displayed on the monitor 40. This figure shows the state after clicking the date tab 134 from the state of FIG. 6, and the structure display area 122 displays a series of knowledge data from the viewpoint of “date”. In the document display area 126, the contents of the series of knowledge data displayed in the structure display area 122 are displayed in a one-dimensionally linked state.

FIG. 8 schematically shows a process in which, when an editing process is performed on a series of knowledge data, the edited contents are reflected in the relational structure between the series of knowledge data. For example, when new knowledge data is inserted in a position between knowledge data in a chained document 160 in which the contents of a series of knowledge data are chained, the inserted knowledge data is placed in the tree 162 before and after. It is added to the same hierarchy as the knowledge data. For example, when the chained document 160 is clipped in any of the knowledge data, the knowledge data is added to the hierarchy below the knowledge data of the pasting destination in the tree 162. When new knowledge data is inserted into the tree 162, the knowledge set 164 shown in FIG.
The new node and the node related to this are connected in a branch.

The present invention has been described above based on the embodiments. This embodiment is an example, and various modifications can be made to the combination of each component and each processing process,
Those skilled in the art will understand that such modifications are also within the scope of the present invention. Hereinafter, modified examples will be described.

4 to 7, the structure display area 122 is shown.
The display format of the related structure of the knowledge data displayed in is switched by switching the dimension of the space to be visualized with a button or by designating a tab corresponding to the viewpoint. In a modification, an interface for instructing switching by a so-called slider type switch may be provided. For example, a horizontal or vertical slider is provided, and the number of layers to be extracted as a tree structure, the genre of the extraction point, the starting node, etc. are gradually changed according to the slider position specified by the user. Good.

In the above embodiment, the case where the same user extracts desired knowledge from the knowledge data group possessed by a single user is mainly illustrated, but in the modified example, the knowledge data shared by a plurality of users. The configuration may be such that desired knowledge is extracted from each group.

In the above embodiment, a directed graph or a tree structure has been illustrated as a format for displaying the related structure of a plurality of pieces of knowledge data, but as a modification, the structure may be displayed in a table format.

The viewpoint or starting point designated by the user may be accumulated as one empirical rule or learning result. A function of presenting the viewpoint and the expected value of the starting point to the user based on the accumulated information may be provided.

The instructions obtained from the user may be stored as a history. Based on the history, a function may be provided to make the detection speed of the relationship or the extraction speed of the series of knowledge data groups efficient or speedy.

[0043]

According to the present invention, it is possible to improve the convenience of the user in utilizing the accumulated information.

[Brief description of drawings]

FIG. 1 is a functional block diagram showing a basic configuration of a knowledge data processing device.

FIG. 2 is a diagram showing a plurality of pieces of knowledge data and their relationships in the form of a directed graph.

FIG. 3 is a diagram showing a tree structure of a plurality of pieces of knowledge data extracted as a focused knowledge data group and their relationships.

FIG. 4 is a diagram showing a screen configuration when a correlation between a plurality of pieces of knowledge data and the contents of each piece of knowledge data is displayed on a monitor.

FIG. 5 is a diagram showing a screen configuration when a correlation between a plurality of pieces of knowledge data and the contents of each piece of knowledge data is displayed on a monitor.

FIG. 6 is a diagram showing a screen configuration when a correlation between a plurality of pieces of knowledge data and the contents of each piece of knowledge data is displayed on a monitor.

FIG. 7 is a diagram showing a screen configuration when a correlation between a plurality of pieces of knowledge data and the contents of each piece of knowledge data is displayed on a monitor.

FIG. 8 is a diagram schematically showing a process in which, when an edit process is performed on a series of knowledge data, the edited contents are reflected in a relational structure between the series of knowledge data.

[Explanation of symbols]

10 Knowledge Data Processing Device, 12 Instruction Acquisition Unit, 2
0 management unit, 22 input processing unit, 24 data storage unit, 30 processing unit, 32 viewpoint determining unit,
34 Relation Detection Unit, 36 Data Extraction Unit, 38
Display processing unit, 39 edit processing unit.

Claims (3)

[Claims] 1. A data storage unit for accumulating a plurality of pieces of knowledge data, and a viewpoint of acquiring a designation of the viewpoint from a user in order to detect a potential mutual relation of the plurality of knowledge data from a desired viewpoint. A determination unit, a relation detection unit that detects the relation from the acquired viewpoint, a data extraction unit that extracts a series of knowledge data linked by the detected relation, and a state of the chain in a space of a desired dimension. And a display processing unit for displaying the series of knowledge data in a form visualized in [4]. 2. A viewpoint determining unit that acquires a specification of the viewpoint from a user in order to reveal a latent relationship between a plurality of accumulated knowledge data from a desired viewpoint, and the relationship from the acquired viewpoint. A relation detecting unit that detects a sex, a data extracting unit that extracts a knowledge data group that the user pays attention to based on the detected relation, and a structure of a comprehensive correlation in the attention knowledge data group with a graph A knowledge data processing device comprising: a display processing unit which is constructed and visualized. 3. An edit processing unit that processes addition of knowledge data to the data storage unit based on an instruction from a user, wherein the association detection unit stores the added knowledge data and the data storage unit. The knowledge data processing device according to claim 1 or 2, wherein a relation with existing knowledge data is detected.