JP2011123743A - Information map creating apparatus, information map creating method and program - Google Patents

Abstract

To create an information map in which the concentration of relation lines to a specific node and the loss of important information are appropriately avoided.
An information map creation device includes upper limit value setting means for setting an upper limit value of a relational line connectable for each information element, and related information storage means for storing the strength of the relation for each relation between the information elements. Based on the above, in the descending order of the strength, the relation selection means for selecting the relation to be displayed as the relation line within the range of the upper limit value, and the relation that has not been selected as a display target because it exceeds the upper limit value range For each information element, and an upper limit value for increasing the upper limit value of the information element for which the number recorded in the upper limit excess number storage means is the maximum. Updating means, and the relation selection means re-executes the relation selection within the range of the updated upper limit value.
[Selection] Figure 5

Description

  The present invention relates to an information map creating apparatus, an information map creating method, and a program for creating an information map that expresses a relationship between information elements by a relationship line connecting the information elements.

  Text mining products, patent analysis systems, etc. are equipped with information map creation and display functions to support search and analysis. In the information map, the relationship between words and data items (patents and bibliographic information of patents) included in searched information or information to be analyzed is illustrated as a network diagram as shown in FIG. . The placement position of words and data items on the information map (hereinafter referred to as “information elements”) and the relationship lines (edges) between the information elements indicate the co-occurrence information between the information elements (the degree of appearance together in the same document). Information) or the like. If a word map is created as an information map, the main topics of the document group can be known. If an IPC (International Patent Classification) map is created as an information map, it is possible to know the dependency relationship of patent documents in the technical field. If an inventor's map is created as an information map, the network by joint application can be known. As described above, it is possible to easily grasp the summary information of the large-volume document group by the information map.

  In the information map, it is important to simplify the diagram in order to improve the readability of the diagram. In order to simplify the drawing, a technique for performing edge thinning has been developed. In thinning out edges, a method of deleting in order from weakly related edges is common. However, there is a possibility that edges are concentrated on a specific node in the simple thinning method. As a result, the information map may be meaningless (no information amount) as a network diagram. For example, FIG. 2 is a diagram illustrating an example of an information map in which edges are concentrated on a specific node. The information map as shown in FIG. 2 only shows that the information element X has a relationship with other information elements.

  Therefore, a technique for avoiding edge concentration on a specific node by devising the edge thinning method (for example, limiting the maximum number of edges to each node) has been proposed (for example, Patent Document 1). According to the technique described in Patent Document 1, as shown in FIG. 3, it is possible to avoid nodes from being concentrated on specific nodes.

Japanese Patent No. 4167855

  In the case of Patent Document 1, in order to avoid the concentration of edges on a specific node, it is based on the idea that even if a relatively strong edge is deleted, it is inevitable. Specifically, due to the limit on the maximum number of edges calculated in advance based on the statistical information of the node, there is a possibility that edges related to weak associations will be given priority for display rather than edges related to relatively strong associations. There is.

  However, the optimum display state of the information map varies depending on the purpose of use of the information map. In order to balance the concentration of edges on a specific node and prioritizing edges related to strong associations as display targets, an information map that is emphasized by the latter may be required.

  The present invention has been made in view of the above points, and is an information map creation that can create an information map that appropriately avoids the concentration of relation lines to specific nodes and the loss of important information. An object is to provide an apparatus, an information map creation method, and a program.

  Therefore, in order to solve the above-described problem, the information map creation device stores an upper limit value setting unit that sets an upper limit value of a relation line that can be connected for each information element, and stores the strength of the relation for each relation between the information elements. Based on the relation information storage means, the relation selection means for selecting the relation to be displayed as the relation line within the range of the upper limit value in descending order of the strength, and the selection target to be displayed because it exceeds the range of the upper limit value. A recording unit that records the number of associations that did not exist in the upper limit excess number storage unit for each information element; and a predetermined number of the upper limit value of the information element that has the maximum number recorded in the upper limit excess number storage unit An upper limit updating means for increasing the relation, and the relation selection means re-executes the selection within the range of the updated upper limit value.

  According to the disclosed technique, it is possible to create an information map in which the concentration of relation lines to a specific node and the lack of important information are appropriately avoided.

It is a figure which shows the example of an information map. It is a figure which shows the example of the information map which the edge concentrated on the specific node. It is a figure which shows the example of the information map by which thinning was performed by avoiding the concentration of the edge to a specific node. It is a figure which shows the hardware structural example of the information map creation apparatus in embodiment of this invention. It is a figure which shows the function structural example of the information map creation apparatus in embodiment of this invention. It is a flowchart for demonstrating the process sequence of an information map production apparatus. It is a figure which shows the example of the statistical information of an information element. It is a figure which shows the example of the relevance matrix table which shows the relevant information of an information element. It is a figure which shows the example of a relevance sort table. It is a flowchart for demonstrating the process sequence of a related thinning-out process. It is a figure which shows the example of a connection number management table. It is a figure which shows the example of a class value management table. It is a figure which shows the example of an effective related matrix table. It is a figure which shows the example of an update of a connection number management table. It is a figure which shows the example of an update of a class value management table. It is a figure which shows the example of an upper limit excess frequency management table. It is a figure for demonstrating the specific example of a related thinning-out process. It is a figure for demonstrating the specific example of a related thinning-out process. It is a figure for demonstrating the specific example of a related thinning-out process. It is a figure which shows the example of the effective related matrix table in a state (2). It is a figure which shows the example of the class value management table in a state (2). It is a figure which shows the example of the connection number management table in a state (2). It is a figure which shows the example of the upper limit excess frequency management table in a state (3). It is a figure which shows the example of the effective related matrix table in a state (5). It is a figure which shows the example of the class value management table in a state (5). It is a figure which shows the example of the connection number management table in a state (5). It is a figure which shows the example of the upper limit excess frequency management table in a state (7). It is a figure which shows the example of the effective related matrix table in a state (9). It is a figure which shows the example of the class value management table in a state (9). It is a figure which shows the example of the connection number management table in a state (9). It is a figure which shows the example of the upper limit excess frequency management table in a state (9). It is a figure which shows the example of the connection number management table after an upper limit connection number update. It is a figure which shows the example of the upper limit excess frequency management table in a state (10).

  Hereinafter, embodiments of the present invention will be described with reference to the drawings. FIG. 4 is a diagram illustrating a hardware configuration example of the information map creation device according to the embodiment of the present invention. 4 includes a drive device 100, an auxiliary storage device 102, a memory device 103, a CPU 104, a display device 105, and an input device 106 that are mutually connected by a bus B. .

  A program that realizes processing in the information map creation apparatus 10 is provided by a recording medium 101 such as a CD-ROM. When the recording medium 101 on which the program is recorded is set in the drive device 100, the program is installed from the recording medium 101 to the auxiliary storage device 102 via the drive device 100. However, the program need not be installed from the recording medium 101 and may be downloaded from another computer via a network. The auxiliary storage device 102 stores the installed program and also stores necessary files and data.

  The memory device 103 reads the program from the auxiliary storage device 102 and stores it when there is an instruction to start the program. The CPU 104 realizes functions related to the information map creation device 10 according to a program stored in the memory device 103. The display device 105 displays a GUI (Graphical User Interface) or the like by a program. The input device 106 includes a keyboard and a mouse, and is used for inputting various operation instructions.

  FIG. 5 is a diagram illustrating a functional configuration example of the information map creation device according to the embodiment of the present invention. In the figure, an information map creation apparatus 10 includes a document management DB 11, a search unit 12, an information extraction and totaling unit 13, an output element selection unit 14, an extended thinning unit 15, a visualization processing unit 16, and the like. Each of these units is software realized by a process executed by the CPU 104 by a program installed in the information map creation apparatus 10.

  The document management DB 11 is a database that systematically manages documents (document data) using the auxiliary storage device 102. In the present embodiment, the content of the document does not matter. A wide variety of documents can be targeted, such as patent documents, papers, books, or business materials.

  The search unit 12 searches the document management DB 11 for document data that matches the input search condition, and outputs a set of searched documents as a target document set 21. That is, the target document set 21 includes the contents (bibliographic information, text, etc.) of each document. In the present embodiment, the target document set 21 is an information set as an information map creation target. The target document set 21 may be given from outside the information map creation apparatus 10. Specifically, the target document set 21 may be input to the information map creation device 10 via a network or a portable recording medium. Therefore, the information map creation apparatus 10 does not necessarily have the document management DB 11 and the search unit 12.

  The information extraction / aggregation unit 13 executes general (known or publicly known) information analysis processing, extracts information elements from the target document set 21, outputs statistical information about the information elements, analyzes relationships between information elements, and the like. I do. The processing result of the information extraction / aggregation unit 13 is output as an extraction / aggregation result 22. Therefore, the extracted total result 22 includes the extracted information elements, statistical information, information related to the relationship (related link information), and the like. Note that general information analysis technology is a morphological analysis process that divides a sentence into words, a dependency analysis process that extracts subjects, predicates, objects, modification relationships, and the like, and the appearance frequency and importance of words are determined. Includes statistical processing, co-occurrence relationship counting processing for counting the number of times two words appear simultaneously, and the like. An information element is a constituent element of an information set such as a word extracted from an information set to be created in an information map, a value of each item of bibliographic information in a document, and can be a node in an information map. . In the present embodiment, the information element is an element extracted from the document (that is, a constituent element of the document), and thus can be referred to as a document element. The bibliographic information in the document is each item in the application, the name of the invention in the specification, etc., taking patent literature as an example.

  The output element selection unit 14 selects (selects) an information element to be displayed as a node on the information map from the information elements included in the extraction count result 22. The selected information element, statistical information related to the information element, related information, and the like are output as a selection result 23.

  The extended thinning unit 15 executes processing for thinning out the relationship between information elements. Association is information expressed as a relationship line (edge) on the information map. Therefore, thinning out the relationship is substantially synonymous with thinning out the relationship line (excluding it from the display target). The former is an expression based on the viewpoint of the processing contents of the computer, and the latter is an expression based on the visual viewpoint of the information map. Similarly, an information element and a node are substantially synonymous.

  The extended thinning unit 15 includes an upper limit connection number setting unit 151, a relation selection unit 152, an upper limit excess number recording unit 153, an upper limit connection number update unit 154, a connection number management table 155, an effective association matrix table 156, and an upper limit excess number management table 157. And a class value management table 158 and the like.

  The upper limit connection number setting unit 151 sets the maximum number of connection lines (upper limit connection number) for each information element included in the selection result 23. The setting result is recorded in the connection number management table 155. The connection number management table 155 is a table for recording the upper limit number of connections and the number of connected relation lines for each information element.

  The relation selection unit 152 selects a relation to be displayed as a relation line from the relations indicated by the relation information included in the selection result 23. Associations that can be displayed as relationship lines are limited by the upper limit number of connections set in each information element. That is, when there is a relationship that exceeds the maximum number of connections of the information element for the same information element, the number of relationships that can be displayed as a relationship line is limited to the maximum number of connections. A relation selection result to be displayed as a relation line is recorded in the valid relation matrix table 156. The valid relation matrix table 156 is a table for recording relations that are valid as display targets (that is, used for displaying relation lines). Further, the number of relation lines connected for each information element is recorded in the connection number management table 155 based on the relation selected as the relation line display target.

  The upper limit excess count recording unit 153 records, for each information element, the number of associations (hereinafter referred to as “upper limit excess count”) that has not been displayed as a relationship line due to the limit on the upper limit connection count in the upper limit excess count management table 157. To do.

  The upper limit connection number updating unit 154 updates the upper limit connection number of some information elements based on the upper limit number of times. After updating the upper limit number of connections, the process by the relation selection unit 152 and the like is re-executed.

  The class value management table 158 is a table in which class values are recorded for each information element. The class value is an identifier for identifying the connection relationship between information elements based on the relationship selected as the display target of the relationship line. Specifically, initially, a unique value is assigned to each information element as the class value. Thereafter, when an association displayed as a relation line is selected, the class values of the information elements related to the association are set to the same value. Thus, information elements having the same class value indicate that they are connected directly or indirectly by a relationship line. Indirect connection means that two information elements are connected via one or more other information elements.

  Each table such as the connection number management table 155, the valid association matrix table 156, the upper limit excess number management table 157, and the class value management table 158 is realized using the memory device 103 or the auxiliary storage device 102.

  The visualization processing unit 16 visualizes the information map based on the result of the thinning process by the extended thinning unit 15.

  Hereinafter, the processing procedure of the information map creation device 10 will be described. FIG. 6 is a flowchart for explaining the processing procedure of the information map creating apparatus.

  For example, in response to input of a search condition by the user, the search unit 12 searches the document management DB 11 for a set of documents that match the search condition, and the searched target document set 21 is stored in the memory device 103 or the auxiliary storage device 102 (hereinafter referred to as “subject storage”). , Referred to as “storage means”) (S101). Depending on the search conditions, the target document set 21 may include only one document. Subsequently, the information extraction / aggregation unit 13 analyzes the target document set 21 and outputs an extraction / aggregation result 22 including statistical information of the information elements, related information, and the like to the storage unit (S102).

  FIG. 7 is a diagram illustrating an example of statistical information of information elements. In the figure, the statistical information 221 includes the appearance frequency, the number of appearance unit texts, the importance, etc. for each information element extracted from the target document set 21. That is, alphabets (X, P to Q, A to H, etc.) in the figure are abstract representations of information elements, and the same applies to the following description. The appearance frequency is the sum of the appearance frequencies in all unit texts of the information element. The unit text is also described in “Text Mining by Association of Words (Information Processing Society of Japan 55th Informatics Fundamental Research Group Material (1999), Isamu Watanabe, Kazuo Miki)” (hereinafter referred to as “Reference 1”). As it is, it means a set of meaningful sentences such as one paragraph or one article. The number of appearance unit texts is the total number of unit texts in which information elements appear. Importance refers to the importance of information elements in a set of unit texts, and is obtained as the sum of the importance of information elements in a unit text. The importance of the information element in the unit text is determined as a function of the statistical information in the narrow sense of the information element. The narrowly defined statistical information includes the appearance probability of information elements in a unit text and the appearance probability of information elements in a set of all unit texts, or the appearance frequency of information elements in a unit text and the number of appearance unit texts of information elements. The importance is also detailed in Reference Document 1.

  FIG. 8 is a diagram illustrating an example of a relevance matrix table indicating related information of information elements. In the drawing, the relevance matrix table 222 is a table in which the relevance is recorded for each combination of information elements having a relevance (that is, for each relevance between information elements). The relevance is, for example, the sum of products of importance in the same unit text, and is detailed in Reference Document 1. However, in the present embodiment, weighting is applied to the sum.

  The information extraction / aggregation unit 13 further sorts the associations whose association degrees are recorded in the association degree matrix table 222 in descending order based on the association degrees, and records the sorting result in the association degree sorting table 223.

  FIG. 9 is a diagram illustrating an example of the relevance sort table. As shown in the figure, associations are recorded in the association degree sort table 223 in descending order of association degree. The association is expressed by a combination (information element pair) of information elements related to the association (at both ends of the association). Specifically, a code (alphabet) indicating two information elements is represented by being connected by “-(hyphen)”.

  Note that the number of co-occurrence may be calculated in addition to the degree of association or instead of the degree of association. The number of times of co-occurrence is, for example, the number of times of appearing simultaneously (co-occurring) in the unit text. Similar to the degree of association, the number of co-occurrence is also an index value or statistical value indicating the strength of association.

  Subsequently, the output element selection unit 14 selects (extracts) information elements to be displayed on the information map based on the statistical information 221, and records the selection result 23 in the storage unit (S103). For example, information elements up to the 50th highest appearance frequency are extracted as constituent elements of the information map. Nodes that are not extracted are removed from the relevance matrix table 222. Therefore, the information element pair including the node that has not been extracted is removed from the relevance sort table 223. Note that the information element may be selected based on another index included in the statistical information 221 such as the number of appearance unit texts or the importance level.

  Subsequently, the extended thinning unit 15 performs related thinning processing (S104). Details of the related thinning process will be described later. Subsequently, the visualization processing unit 16 visualizes the information map based on the information generated by the extended thinning unit 15 (S105). For example, the visualization processing unit 16 causes the display device 106 to display an information map. Alternatively, the information map may be printed by a printer (not shown). In the information map, the arrangement position of each node is determined by the degree of relevance of the relationship line connecting the nodes. That is, each relationship line is regarded as a spring, and the length and strength of the spring are determined according to the degree of relevance of each relationship line. By making the repulsive force act on each node, the arrangement position of each node is determined at a position where the relationship between the tension and the initial length of the relation line formed as a spring and the repulsive force between the nodes is stable. The method for determining the location of the node is detailed in “Visualization technology of associative relations for text mining” (Information Processing Society of Japan, 55th Informatics Fundamental Research Group Material (1999), Kazuo Mitsue, Isamu Watanabe).

  Next, details of step S104 will be described. FIG. 10 is a flowchart for explaining the processing procedure of the related thinning-out processing.

  In step S <b> 201, the upper limit connection number setting unit 151 sets the upper limit connection number for each information element included in the selection result 23 and records the setting result in the connection number management table 155.

  FIG. 11 is a diagram illustrating an example of the connection number management table. As shown in the figure, the connection number management table 155 records the number of connections and the maximum number of connections for each information element. The upper limit number of connections is as described above, and the value is recorded in step S201. When the relation displayed as the relation line in the subsequent process is selected, a value is added to the information element related to the relation. That is, the number of connections is the number of relation lines connected to the information element.

  The upper limit number of connections (“3”) shown in the figure is merely an example. Other values may be used as the upper limit number of connections within a reasonable range for the visibility of the information map. Further, the upper limit number of connections for all information elements may not be the same value. For example, like the method described in Japanese Patent No. 4167855, the upper limit number of connections of each information element may be individually determined based on the statistical information 221 and the like.

  In the figure, for convenience, all information elements shown in FIG. 7 are targeted. That is, an example in which all the information elements shown in FIG. 7 are included in the selection result 23 is shown. In the following description, the information element shown in FIG. 11 is a processing target.

  Subsequently, the relation selecting unit 152 initializes the valid relation matrix table 156, the connection number management table 155, and the upper limit excess number management table 157 (S202). Initialization means emptying the recorded contents of the table.

  Subsequently, the association selecting unit 152 sets (assigns) a class value for each information element, and records the setting result in the class value management table 158 (S203).

  FIG. 12 is a diagram illustrating an example of a class value management table. In the figure, an example is shown in which numbers are set in order from 1 to each information element. The method of setting the class value is not limited to a predetermined one as long as there is no overlap between information elements. For example, the starting point may not be 1 or a random value may be set. Moreover, if the order relationship is clear, it is not limited to numerical values. For example, alphabets may be used as class values.

  Subsequently, the relationship selection unit 152 sets one record in order from the top of the relevance level sort table 223 (S204). In other words, records having a high degree of association (high association strength) are sequentially processed. Hereinafter, the relation to be processed is referred to as “current record”.

  Subsequently, the association selecting unit 152 compares the class values of the two information elements related to the current record (S205). That is, it is determined whether or not two information elements are already directly or indirectly connected by a relation line. The class value of each information element is acquired from the class value management table 158.

  When the class values of the two information elements are different (that is, when the two information elements are not directly or indirectly connected by the relationship line) (Yes in S205), the relation selection unit 152 determines that the number of connections of the two information elements is It is determined with reference to the connection number management table 155 whether each is less than the upper limit connection number (S206). When the number of connections of the two information elements is less than the upper limit number of connections (Yes in S206), the relationship selection unit 152 selects the relationship related to the current record as a display target (S207). Specifically, the relation selection unit 152 updates the valid relation matrix table 156.

  FIG. 13 is a diagram illustrating an example of an effective association matrix table. As shown in the figure, the valid association matrix table 156 can record the presence / absence of association displayed as a relation line for each combination of information elements. Specifically, “1” is recorded for a cell related to a combination having a relation displayed as a relation line. In the same figure, an example is shown in which the relation relating to the highest level record (X-P) in the relation degree sort table 223 is validated as a display target. There are two cells where the information element X and the information element P intersect. Therefore, “1” is recorded for two cells.

  Note that the relation being a display target means that a relation line indicating the relation is connected to two information elements related to the relation. Therefore, the relationship selection unit 152 increases the number of connections in the connection number management table 155 by 1 for the two information elements.

  FIG. 14 is a diagram illustrating an example of updating the connection number management table. In the figure, 1 is added to the number of connections of each of the information element X and the information element P.

  Further, in order to indicate that the two information elements are connected by a relationship line, the association selecting unit 152 updates the class value of one of the two information elements with the other value. In the present embodiment, the larger class value is updated by the smaller class value.

  FIG. 15 is a diagram illustrating an update example of the class value management table. In the figure, the class value of the information element P is updated with the class value “1” of the information element X. If there is another information element having the same class value as the updated class value, the class value is also updated for the other information element. That is, the update of the class value is propagated to information elements having the same class value. Thereby, the commonality of the class value between the indirectly connected information elements is ensured.

  As described above, when one association is selected as a display target, the effective association matrix table 156, the connection number management table 155, and the class value management table 158 are updated.

  On the other hand, when the class values of the two information elements related to the current record are the same (No in S205), or when the number of connections of at least one of the two information elements has reached the upper limit number of connections (No in S206). ), The relationship selection unit 152 does not display the relationship related to the current record. When the class values of the two information elements are the same, the relation is not included in the display target in order to prevent complication of the information map and deterioration of readability accompanying complication. That is, two information elements having a common class value are connected at least indirectly by a relation line. Then, even if the direct relationship between the two information elements is not to be displayed, it can be expressed on the information map that the two information elements are related. Therefore, there is a possibility that a better result can be obtained in the balance between the amount of information and the visibility of the information map, by reducing the display elements while hiding the relations, than by displaying the relations and increasing the display elements. Because it is expensive.

  In addition, when either number of connections has reached the upper limit number of connections, the relation is not displayed. This is in order to avoid the concentration of relation lines for information elements whose number of connections has reached the upper limit number of connections. is there.

  On the other hand, when the reason why the relationship is not selected as a display target is that the number of connections of at least one of the two information elements has not reached the upper limit connection number (No in S206), the upper limit excess number recording unit 153 The upper limit excess count management table 157 is updated (S208).

  FIG. 16 is a diagram illustrating an example of the upper limit excess count management table. As shown in the figure, the upper limit excess count management table 157 is a table for recording the upper limit excess count for each information element. In step S208, the upper limit excess count recording unit 153 adds 1 to the upper limit excess count of the information element in which the number of connections reaches the upper limit connection count among the information elements related to the current record.

  Subsequent to No in step S207, S208, or S205, the relationship selection unit 152 determines whether at least one relation line is connected to all information elements (condition a), or is the current record the last record? (Condition b) is determined (S209). Whether or not at least one relation line is connected to all information elements is determined based on the number of connections in the connection number management table 155. That is, if the number of connections for all information elements is 1 or more, it is determined that at least one relational line is connected to all information elements. When both the condition a and the condition b are not satisfied (No in S209), Step S204 and subsequent steps are repeatedly executed. That is, in the relevance sort table 223, the record with the next highest relevance is set as the current record.

  When either one of the condition a or the condition b is satisfied (Yes in S209), the upper limit connection number updating unit 154 refers to the upper limit excess number management table 157 and determines whether there is an information element whose upper limit connection number should be updated. (S211). Specifically, the information element having the maximum number of times of exceeding the upper limit is selected as the update target of the upper limit number of connections. Alternatively, an information element having the maximum sum of related degrees related to the upper limit excess number may be selected as an update target. In this case, in step S208, the relevance level related to the current record may be added instead of the upper limit excess count. When a plurality of information elements are selected, some (for example, one) of them may be selected as update targets, or all may be selected. In the upper limit excess count management table 157, when the upper limit excess count for all information elements is “0”, it is determined that there is no information element that increases the upper limit connection count.

  If there is no information element for increasing the upper limit number of connections (No in S210), the process in FIG. 10 ends. When there is an information element that increases the upper limit connection number (Yes in S210), the upper limit connection number update unit 154 determines whether the information element selected as the update target is the same as the information element that was the last update target. Is determined (S211). That is, it is determined whether or not the upper limit number of connections of the same information element has been updated twice. When the upper limit number of connections of the same information element is an update target twice (Yes in S211), the processing in FIG. 10 ends. This is for avoiding the concentration of relation lines with respect to the information element. However, if there are multiple information elements selected as the update target for the upper limit number of connections, if there is an information element that is different from the information element selected as the previous update target, the process continues with the different information element as the update target. May be.

  On the other hand, when the information element selected as the last update target and the information element selected as the current update target are different (No in S211), the upper limit connection number updating unit 154 sets the upper limit of the information elements selected as the current update target. The number of connections is updated (S212). Specifically, in the connection number management table 155, the upper limit connection number of the information element is increased. The increment may be 1 or a predetermined value of 2 or more. Further, the upper limit connection number updating unit 154 records the history of the information element that is the update target in the memory device 103. This is for the purpose of comparison in step S210 to be executed next time. Subsequently, step S202 and subsequent steps are repeatedly executed. However, in step S202, the upper limit number of connections in the connection number management table 155 is not subject to initialization.

  Note that the end condition in step S211 is not limited to the same information element being selected twice in succession. Other conditions may be used as long as it is possible to determine that there is a high possibility that the relationship lines are concentrated on one information element (node). For example, the condition may be that the same information element is selected at a predetermined number of times of three or more. Alternatively, the condition may be that the same information element is selected M times or more (N ≧ M) in N times. In the case of this condition, the processing can be terminated even if the same information elements are not necessarily continuously updated. When N = M, it means that the same information element has been selected N times consecutively.

  Subsequently, the processing content described in FIG. 10 will be described by applying a specific example. 17, FIG. 18, and FIG. 19 are diagrams for explaining a specific example of the related thinning process. In each figure, information elements and relationships are expressed in the form of an information map. In this specific example, it is assumed that the upper limit number of connections of each information element is “3” as shown in FIG. In this specific example, the relevance sort table 223 shown in FIG.

  In FIG. 17, (0) indicates a state in which all the associations shown in the association degree matrix table 222 or the association degree sort table 223 are activated (displayed as a relation line). That is, the related thinning is not performed.

  (1) shows the state after execution of step S202 in FIG. That is, all the relations are cleared.

  (2) shows a state in which the top three records of the relevance sort table 223 have been processed. First, processing is performed for XP. At this stage, the information element X (hereinafter simply referred to as “X”; the same applies to other information elements) and P have different class values (Yes in S205). The number of connections of X and P is “0” and less than “3” (Yes in S206). Therefore, XP is selected as a display target (S207). At this stage, the relationship lines are not connected to all information elements. Further, XP is not the last record (No in S209). Accordingly, XQ is subsequently processed. X-R processed subsequent to X-Q and X-Q is also processed in the same manner as XP, and each association is selected as a display target. As a result, in the state (2), the valid association matrix table 156, the class value management table 158, and the connection number management table 155 are as shown below.

  FIG. 20 is a diagram illustrating an example of an effective association matrix table in the state (2). In the figure, “1” is recorded in each cell related to XP, XQ, or XR.

  FIG. 21 is a diagram illustrating an example of the class value management table in the state (2). In the figure, “1” is recorded in the class values of X, P, Q, and R. This is a result in which the class value of X is given to P, Q, and R because XP, XQ, and XR are the display targets. As is clear from FIG. 17 (2), these information elements are directly or indirectly connected by a relationship line.

  FIG. 22 is a diagram illustrating an example of the connection number management table in the state (2). Since the relation between XP, XQ, and XR is the display target, the number of connections of X, P, Q, and R is “3”, “1”, “1”, “1”. Is recorded.

  Subsequently, (3) shows a state where XS has been processed. As is apparent from FIG. 22, the number of X connections has already reached the upper limit number of connections (No in S206). Therefore, XS is not a display target. However, since the reason why XS is not displayed is that the number of X connections has reached the upper limit number of connections (No in S206), 1 is added to the upper limit number of times of X.

  FIG. 23 is a diagram illustrating an example of the upper limit excess count management table in the state (3). In the figure, the upper limit excess number for X is updated from “0” to “1”.

  In (3), XS is indicated by a dotted line. The dotted line indicates that the relation related to the broken line has not been selected as a display target. The same applies to the subsequent drawings.

  Subsequently, FIG. 18 (4) shows a state where PQ, PS, QR, and RS are processed in order. For PQ, P and Q have the same class value (No in S205). Therefore, PQ is not displayed. For PS, P and S have different class values (Yes in S205). Further, the numbers of connections of P and S are “1” and “0”, respectively, and less than “3” (Yes in S206). Therefore, PS is selected as a display target (S207). For QR, Q and R have the same class value (No in S205). Therefore, QR is not a display target. For RS, R and S have the same class value (No in S205). Therefore, RS is not a display target. As a result, the state shown in (5) is obtained. In the state (5), the valid association matrix table 156, the class value management table 158, and the connection number management table 155 are as shown below.

  FIG. 24 is a diagram illustrating an example of an effective association matrix table in the state (5). In the drawing, “1” is further recorded in each cell related to PS.

  FIG. 25 is a diagram illustrating an example of the class value management table in the state (5). The class value “1” of P is given to S because PS is the display target.

  FIG. 26 is a diagram illustrating an example of the connection number management table in the state (5). Since PS is displayed, 1 is added to the number of connections of P and S.

  In (4) and (5), the upper limit excess count management table 157 is not updated.

  Subsequently, (6) shows a state in which X-A, X-B, X-C, X-D, X-E, X-F, X-G, and X-H are processed. As is apparent from FIG. 26, the number of X connections has already reached the upper limit number of connections (No in S207). Accordingly, none of these associations are displayed. As a result, the state shown in (7) is obtained. That is, the same state as (5) is maintained. However, the reason why each relation is not displayed is that the number of connections of X has reached the upper limit number of connections (No in S207), and therefore the upper limit number of times of X is updated.

  FIG. 27 is a diagram illustrating an example of the upper limit excess count management table in the state (7). In the figure, the upper limit number of times of X is updated to “9”. That is, “8” is added as compared with the state of FIG. “8” is the number of associations that are not targeted for display in (6) due to the number of X connections.

  Subsequently, FIG. 19 (8) shows a state where PA, PB, QC, QD, RE, RF, SG, and SH are processed. Regarding PA, P and A have different class values (Yes in S205). Further, the numbers of connections of P and A are “2” and “0”, respectively, and both are less than “3” (Yes in S206). Therefore, PA is selected as a display target (S207). Regarding P-B, although the class values of P and B are different, the number of connections of P has reached the upper limit number of connections due to the display target of P-A (No in S206). Therefore, P-B is not a display target. However, 1 is added to the upper limit excess count for P. Regarding Q-C, Q and C have different class values (Yes in S205). The numbers of connections of Q and C are “1” and “0”, respectively, and both are less than “3” (Yes in S206). Therefore, QC is selected as a display target (S207). Thereafter, Q-D, R-E, R-F, S-G, and S-H are also selected as display targets. As a result, the state shown in (9) is obtained. In the state of (9), the valid association matrix table 156, the class value management table 158, the connection number management table 155, and the upper limit excess number management table 157 are as shown below.

  FIG. 28 is a diagram illustrating an example of an effective association matrix table in the state (9). In the figure, “1” is recorded in each cell related to PA, QC, QD, RE, RF, SG, and SH.

  FIG. 29 is a diagram showing an example of the class value management table in the state (9). In the figure, the class value “1” of P is given to A because PA is a display target. Further, Q class value “1” is given to C and D because QC and QD are displayed. Further, the class value “1” of R is given to E and F by making R-E and R-F display targets. Furthermore, the class value “1” of S is given to G and H because S-G and S-H are the display targets.

  FIG. 30 is a diagram illustrating an example of the connection number management table in the state (9). In comparison with FIG. 26, in FIG. 26, “1” is added to the number of connections of P and A because PA is a display target. Further, since QC and Q-D are displayed, “2” is added to the number of connections of Q, and “1” is added to the number of connections of C and D. Further, since R-E and R-F are displayed, “2” is added to the number of connections of R, and “1” is added to the number of connections of E and F. Furthermore, since S-G and S-H are displayed, “2” is added to the number of connections of S, and “1” is added to the number of connections of G and H.

  Further, FIG. 31 is a diagram showing an example of the upper limit excess count management table in the state (9). In the figure, “1” is added to the upper limit number of Ps based on the fact that PS is not selected as a display target because the number of connections of P has reached the upper limit number of connections.

  In the state (9), since all records of the relevance sort table 223 have been processed (Yes in S209), the presence / absence of an information element that increases the upper limit number of connections is determined (S210). From FIG. 31, it is X that has the maximum number of times of exceeding the upper limit. Therefore, the upper limit number of connections of X is updated from “3” to “4” (S212). In the example of FIG. 31, the upper limit number of connections of X is also updated when compared by the sum of relevance levels related to the upper limit excess count. As a result, in the next loop process (the process after S202), the selection of the relation to be displayed, that is, the relation thinning out, is performed based on the connection number management table 155 shown in FIG.

  FIG. 32 is a diagram illustrating an example of the connection number management table after the upper limit connection number is updated. In the figure, the upper limit number of connections of X is updated from “3” to “4”. The figure shows a state after the number of connections of each information element is initialized.

  In the thinning process based on the connection number management table 155 after the update, since the upper limit number of connections for X has been increased, XS that was to be displayed last time is selected as the display target. When XS is selected, PS is not displayed. This is because the class values of P and S are the same during processing of PS (that is, P and S are indirectly connected via X). Since PS is not a display target, P-B is selected as a display target. This is because the number of P connections is “2” during the P-B process, which is less than the upper limit number of connections. As a result, the state shown in (10) is obtained. In the state of (10), the upper limit excess count management table 157 is as shown below.

  FIG. 33 is a diagram illustrating an example of the upper limit excess count management table in the state (10). In the figure, X is the largest number of times of exceeding the upper limit. Therefore, X is selected as an update target for the upper limit number of connections. However, X is selected as the update target of the upper limit connection number in the previous time as well. Accordingly, when the end condition is that the same information element is selected twice as an update target of the upper limit number of connections, the information map of (10) is the final form of the specific example. Therefore, the visualization processing unit 16 visualizes the information map based on the validity relation matrix table 156 and the like in the state of (10).

  Comparing (9) where the upper limit number of connections is fixed and (10) where the upper limit number of connections is dynamically changed, in (10), X-S having a relatively high relevance is PS. Is given priority for display. Therefore, the lack of relatively important associations is properly avoided.

  Further, when comparing (0) in which the upper limit number of connections is not set in the information element with (10) in which the upper limit number of connections is set, in (10), the concentration of the relationship line to X is avoided. .

  As described above, the information map creating apparatus 10 according to the present embodiment dynamically updates the upper limit number of connections for each information element in the course of related thinning-out processing (related selection processing to be displayed). Specifically, in the course of the related thinning process, the upper limit number of connections for information elements that are highly likely to concentrate the relationship lines is increased. Therefore, it is possible to increase a possibility that a relation having a relatively high degree of relation is selected as a display target as compared with a case where the upper limit number of connections is fixed.

  In addition, since the number of connections of relation lines to each information element is limited within the range of the upper limit connection number that is dynamically changed, it is possible to appropriately avoid excessive concentration of relation lines to one information element. it can. As a result, the readability of the information map can be improved comprehensively.

  As mentioned above, although the Example of this invention was explained in full detail, this invention is not limited to such specific embodiment, In the range of the summary of this invention described in the claim, various deformation | transformation・ Change is possible.

Regarding the above description, the following items are further disclosed.
(Appendix 1)
An information map creation device that creates an information map that expresses a relationship between information elements by connecting nodes corresponding to the information elements by relationship lines,
An upper limit setting means for setting an upper limit value of the relation line connectable for each information element;
A relation selection means for selecting the relation to be displayed as the relation line within the range of the upper limit value in descending order of the strength based on the relation information storage means storing the strength of the relation for each relation;
Recording means for recording the number of associations not selected as a display target because it exceeds the range of the upper limit value in the upper limit excess number storage unit for each information element;
An upper limit update means for increasing the upper limit value of the information element whose number recorded in the upper limit excess number storage means is the maximum, by a predetermined number;
The association selection unit is an information map creation device that re-executes the association selection within the range of the updated upper limit value.
(Appendix 2)
The upper limit value update means records the history of the information element that is the update target of the upper limit value in a history storage means,
The relation selection unit re-executes the selection of the relation when it is determined that the same information element is equal to or more than the second predetermined number of times based on the history for the first predetermined number of times. The information map creation device according to supplementary note 1
(Appendix 3)
The information map creating apparatus according to supplementary note 2, wherein the related selection unit does not re-execute the related selection when it is determined that the same information element is continuously subjected to the update based on the history.
(Appendix 4)
The association selection unit assigns a unique identifier to each of the information elements, updates the identifier of one of the information elements related to the association selected as a display target with the identifier of the other information element, and The information map creation device according to any one of supplementary notes 1 to 3, wherein the association between the information elements having the same is not selected as a display target.
(Appendix 5)
A computer that creates an information map that expresses a relationship between information elements by connecting nodes corresponding to the information elements by relationship lines,
An upper limit value setting procedure for setting an upper limit value of the relation line connectable for each information element;
A relation selection procedure for selecting the relation to be displayed as the relation line within the range of the upper limit value in descending order of the strength based on the relation information storage means storing the strength of the relation for each relation;
A recording procedure for recording the number of associations not selected as a display target because it exceeds the range of the upper limit value in the upper limit excess number storage unit for each information element;
An upper limit update procedure for increasing the upper limit value of the information element whose number recorded in the upper limit excess number storage means is a maximum by a predetermined number;
An information map creation method for re-executing the related selection procedure based on the updated upper limit value.
(Appendix 6)
The upper limit update procedure records a history of the information element that is an update target of the upper limit in a history storage unit,
The information according to appendix 5, wherein the related selection procedure is not re-executed when it is determined that the information element that is the same as the second predetermined number of times or more is to be updated based on the history for the first predetermined number of times. How to create a map.
(Appendix 7)
The information map creation method according to supplementary note 6, wherein when it is determined that the same information element is continuously set as the update target based on the history, the related selection procedure is not re-executed.
(Appendix 8)
The association selection procedure assigns a unique identifier to each of the information elements, updates an identifier of one of the information elements related to the association selected as a display target with the identifier of the other information element, and The information map creation method according to any one of supplementary notes 5 to 7, wherein the association between the information elements having the same is not selected as a display target.
(Appendix 9)
A computer that creates an information map that expresses a relationship between information elements by connecting nodes corresponding to the information elements by relationship lines;
An upper limit value setting procedure for setting an upper limit value of the relation line connectable for each information element;
A relation selection procedure for selecting the relation to be displayed as the relation line within the range of the upper limit value in descending order of the strength based on the relation information storage means storing the strength of the relation for each relation;
A recording procedure for recording the number of associations not selected as a display target because it exceeds the range of the upper limit value in the upper limit excess number storage unit for each information element;
An upper limit update procedure for increasing the upper limit value of the information element whose number recorded in the upper limit excess number storage means is a maximum by a predetermined number;
A program for re-executing the related selection procedure based on the updated upper limit value.
(Appendix 10)
The upper limit update procedure records a history of the information element that is an update target of the upper limit in a history storage unit,
The program according to appendix 9, wherein the related selection procedure is not re-executed when it is determined that the information element that is the same as the second predetermined number of times or more is to be updated based on the history for the first predetermined number of times. .
(Appendix 11)
The program according to supplementary note 10, wherein when it is determined that the same information element is continuously subject to the update based on the history, the related selection procedure is not re-executed.
(Appendix 12)
The association selection procedure assigns a unique identifier to each of the information elements, updates an identifier of one of the information elements related to the association selected as a display target with the identifier of the other information element, and The program according to any one of appendices 9 to 11, wherein the association between the information elements having the same is not selected as a display target.

10 Information map creation device 11 Document management DB
DESCRIPTION OF SYMBOLS 12 Search part 13 Information extraction totaling part 14 Output element selection part 15 Extended thinning part 16 Visualization process part 100 Drive apparatus 101 Recording medium 102 Auxiliary storage apparatus 103 Memory apparatus 104 CPU
105 Display Device 106 Input Device 151 Upper Limit Connection Number Setting Unit 152 Related Selection Unit 153 Upper Limit Excess Number Recording Unit 154 Upper Limit Connection Number Update Unit 155 Connection Number Management Table 156 Valid Related Matrix Table 157 Upper Limit Excess Number Management Table 158 Class Value Management Table B bus

Claims (6)

An information map creation device that creates an information map that expresses a relationship between information elements by connecting nodes corresponding to the information elements by relationship lines,
An upper limit setting means for setting an upper limit value of the relation line connectable for each information element;
A relation selection means for selecting the relation to be displayed as the relation line within the range of the upper limit value in descending order of the strength based on the relation information storage means storing the strength of the relation for each relation;
Recording means for recording the number of associations not selected as a display target because it exceeds the range of the upper limit value in the upper limit excess number storage unit for each information element;
An upper limit update means for increasing the upper limit value of the information element whose number recorded in the upper limit excess number storage means is the maximum, by a predetermined number;
The association selection unit is an information map creation device that re-executes the association selection within the updated upper limit range. The upper limit value update means records the history of the information element that is the update target of the upper limit value in a history storage means,
The relation selection unit re-executes the selection of the relation when it is determined that the same information element is equal to or more than the second predetermined number of times based on the history for the first predetermined number of times. The information map creating apparatus according to claim 1, wherein:   3. The information map creation device according to claim 2, wherein the relation selection unit does not re-execute the relation selection when it is determined that the same information element is continuously subjected to the update based on the history. .   The association selection unit assigns a unique identifier to each of the information elements, updates the identifier of one of the information elements related to the association selected as a display target with the identifier of the other information element, and The information map creation device according to any one of claims 1 to 3, wherein the association between the information elements having the same is not selected as a display target. A computer that creates an information map that expresses a relationship between information elements by connecting nodes corresponding to the information elements by relationship lines,
An upper limit value setting procedure for setting an upper limit value of the relation line connectable for each information element;
A relation selection procedure for selecting the relation to be displayed as the relation line within the range of the upper limit value in descending order of the strength based on the relation information storage means storing the strength of the relation for each relation;
A recording procedure for recording the number of associations not selected as a display target because it exceeds the range of the upper limit value in the upper limit excess number storage unit for each information element;
An upper limit update procedure for increasing the upper limit value of the information element whose number recorded in the upper limit excess number storage means is a maximum by a predetermined number;
An information map creation method for re-executing the related selection procedure based on the updated upper limit value. A computer that creates an information map that expresses a relationship between information elements by connecting nodes corresponding to the information elements by relationship lines;
An upper limit value setting procedure for setting an upper limit value of the relation line connectable for each information element;
A relation selection procedure for selecting the relation to be displayed as the relation line within the range of the upper limit value in descending order of the strength based on the relation information storage means storing the strength of the relation for each relation;
A recording procedure for recording the number of associations not selected as a display target because it exceeds the range of the upper limit value in the upper limit excess number storage unit for each information element;
An upper limit update procedure for increasing the upper limit value of the information element whose number recorded in the upper limit excess number storage means is a maximum by a predetermined number;
A program for re-executing the related selection procedure based on the updated upper limit value.

Images