JP2013003626A - Network visualization system, network visualization method, and network visualization program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To visualize a network constituted by nodes and links so as to overview mutual relationships including a node with low relationship.SOLUTION: A network visualization system includes: layer link generation means for extracting a link not included in link information from a correlation between a node and a link included in cluster information to generate the link as a layer link on the basis of the cluster information produced by grouping link information indicating a relation between pre-set nodes related to a network constituted by nodes and links; graph arrangement means for arranging individual nodes in a space by reflecting the generated link; and layer visualization means for displaying the arranged nodes and individual links by belonging to each layer without changing display position of nodes in matching with the layer.

Description

  The present invention relates to a network visualization system of a network configured by nodes and links (relationships), and more particularly to a network visualization system capable of visualizing a network structure suitable for requirements management in system and software development.

  As a network visualization technique, there is a technique in which nodes and links are automatically arranged by a computer in a two-dimensional (three-dimensional) space. As an example, Patent Document 1 is cited.

  Patent Document 1 is provided with a self-organizing map unit that obtains an arrangement coordinate position of each node from input link data. Each node is appropriately formed as a two-dimensional network using a matrix indicating the distance between the nodes. It is described that can be arranged.

  In the spring model of the node arrangement, the link is regarded as a spring, and the node arrangement is determined so that the spring force applied to the link is minimized. In this method, nodes are arranged using the attractive force between adjacent nodes (connected by links) and the repulsive force between non-adjacent nodes (not connected by links). As an improvement, in Non-Patent Document 1, there is a method for arranging nodes on the assumption that there is a spring having a graph theoretical distance (shortest path length) as a natural length between all pairs of nodes without using repulsive force. It is disclosed.

  However, with such a technique, it is difficult to display in an easy-to-understand manner how much relations there are between nodes, and to find a node that has little relation to others at a glance.

  On the other hand, for example, in Patent Document 2, the nodes are arranged so that the nodes related to each other are close to each other on the two-dimensional plane, and the altitude coordinates are calculated based on the importance and relevance of the nodes, A technique for displaying a structural relationship between nodes in an easy-to-understand manner by generating a topographic map is disclosed. However, with the disclosed technology, it is difficult to investigate a node that has a main node that is conspicuous and has low relevance to others.

By the way, in system and software development, associations (traceability links) between development products within and between processes are generated to manage requirements and specifications. The traceability link is a link that indicates whether there is a relationship between requirements and specifications when the requirements and specifications are nodes.
The traceability link is used to understand the scope of influence when requirements and specifications are changed, to detect reflections in the subsequent processes of requirements and specifications, and to check consistency. In order to grasp the range of influence when a requirement / specification is changed, it is necessary to grasp how much the relationship between the requirement / specification is. In addition, in order to detect project reflection omissions and confirm consistency, it is also important to detect requirements and specifications that have low relevance (nodes with few links). From the viewpoint of project man-hours and progress management, it is also important to understand the amount of requirements and specifications and the status of omission of reflection for each development unit such as components.

JP 2004-362306 A JP 2009-28889A

Kamada and S. Kawai, “An algorithm for drawing general undirected graphs”, Journal information processing letter published in 1989. Vol.31, No.1, pp.7-15

  With existing technology, with regard to a network visualized based on the relationship between a node and that node, humans understand how many nodes are related to each other, or look at nodes that are less related to others. It was difficult to find out.

  If it demonstrates using an example of a previous requirement specification, it can be said that it is difficult to grasp | ascertain the spillover influence range accompanying correction of a predetermined requirement (node) at the time of a requirement change. It can also be said that it is difficult to grasp the relevance of a request that has a small number of relevance (number of links) with other requests. These tend to become more difficult as the network becomes larger.

  An object of the present invention is to provide a network visualization system that visualizes a network composed of nodes and links, including nodes with low relevance, so that the relevance of each other can be overlooked.

  The network visualization system of the present invention is based on link information indicating a relationship between nodes set in advance and cluster information in which individual nodes are grouped with respect to a network configured by nodes and links. A link that is not included in the link information is extracted from the relationship between the nodes and links included in the cluster information and is generated as a hierarchical link, and the generated hierarchical links are reflected to reflect the generated hierarchical links in space. Graph placement means arranged above, and hierarchy visualization means for displaying the arranged nodes and each link for each hierarchy to which the nodes are displayed without changing the display position of the nodes according to the hierarchy.

  ADVANTAGE OF THE INVENTION According to this invention, the network visualization system which visualizes the network comprised by the node and the link so that a mutual relationship can be overlooked also including a node with low relationship can be provided.

It is a block diagram which shows the structural example of 1st Embodiment of the network visualization system of this invention. 4 is an explanatory diagram illustrating an example of initial link information stored in an initial link storage unit 100. FIG. It is explanatory drawing which shows the example of the network visualized based on the initial link. 4 is an explanatory diagram illustrating an example of cluster information stored in a cluster storage unit 101. FIG. It is explanatory drawing which shows the example of the hierarchical link information memorize | stored in the hierarchical link memory | storage part. It is explanatory drawing which shows the example of the network which the graph arrangement | positioning part 104 has arrange | positioned. It is explanatory drawing which shows the example of the 1st layer visualized by the hierarchy visualization part. It is explanatory drawing which shows the example of the 2nd layer visualized by the hierarchy visualization part. It is explanatory drawing which shows the example of the 3rd layer visualized by the hierarchy visualization part. It is a flowchart which shows the process example of 1st Embodiment of the network visualization system of this invention. It is a flowchart which shows the process example of the hierarchical link production | generation part 103 which produces | generates the hierarchical link between the node i and the node j of 1st Embodiment of a network visualization system. It is a flowchart which shows the process example of the hierarchy visualization part 105 which sets the display node and display link of the 1st layer of 1st Embodiment of a network visualization system. It is a flowchart which shows the process example of the hierarchy visualization part 105 which sets the display node and display link of the 2nd layer of 1st Embodiment of a network visualization system. It is a flowchart which shows the process example of the hierarchy visualization part 105 which visualizes the 3rd layer of 1st Embodiment of a network visualization system. As an example, it is explanatory drawing which shows the example of the 1st layer at the time of applying this embodiment to a text analysis. It is explanatory drawing which shows the example of the 1st layer at the time of applying and applying this embodiment to the request | requirement / specification of sales management as an Example. It is explanatory drawing which shows the example of the 3rd layer at the time of applying and visualizing this embodiment in order to detect the omission of a requirement and a specification with respect to the requirement and specification of the same sales management.

  Next, an embodiment of a network visualization system according to the present invention will be described in detail with reference to the drawings.

  Referring to FIG. 1, the network visualization system according to the present embodiment includes a storage unit 11 that stores information and a calculation unit 12 that operates under program control.

  The storage unit 11 includes an initial link storage unit 100, a cluster storage unit 101, and a hierarchical link storage unit 102.

  The initial link storage unit 100 stores link information of a network to be visualized. The link information may be specified by the user of the system by inputting through a keyboard or the like, may be stored in the system as a default, or may be specified in other manners. In the present embodiment, since another hierarchical link is generated from information about a given link, the link given as an input will be referred to as an initial link hereinafter. FIG. 2 is an explanatory diagram illustrating an example of the initial link information stored in the initial link storage unit 100. In the figure, A0, A1,..., D0 indicate nodes, and each element of the matrix indicates the presence or absence of an initial link. For example, the value is 0 because there is no initial link between the node A0 and the node A0, and the value is 1 because there is an initial link between the node A0 and the node A1. Here, a case where the initial link has no direction is described as an example. Therefore, since the matrix is symmetric, only the upper triangular matrix or the lower triangular matrix may be stored. If there is a direction in the initial link, for example, the row may be determined as the start node of the initial link and the column as the end node of the initial link. FIG. 3 shows a network example in which nodes are arranged and visualized based on the initial link illustrated in FIG.

  The cluster storage unit 101 stores cluster information obtained by grouping nodes. The cluster information may be specified by the system user inputting it via a keyboard or the like, may be stored in the system as a default, or may be specified in any other manner. FIG. 4 is an explanatory diagram illustrating an example of cluster information stored in the cluster storage unit 101. In the figure, the first column shows the ID of the representative node of the cluster, and the second and subsequent columns show the IDs of the element nodes of the cluster. In this example, the elements of the cluster whose representative is A0 are A0, A1, and A2, the elements of the cluster whose representative is B0 are B0, B1, and B2, and the element of the cluster whose representative is D0 is only D0. is there. In the following description, a cluster whose representative is A0 is called cluster A, a cluster whose representative is B0 is called cluster B, a cluster whose representative is C0 is called cluster C, and a cluster whose representative is D0 is called cluster D.

  The hierarchical link storage unit 102 stores hierarchical link information. FIG. 5 is an explanatory diagram illustrating an example of hierarchical link information stored in the hierarchical link storage unit 102. In the figure, A0, A1,..., D0 indicate nodes, and each element of the matrix indicates the weight of the hierarchical link. Here, the value of the hierarchical link is indicated by a value larger than 0 when there is a link, and 0 when there is no link. The weight value is used to determine the position of the node. The numerical value of the weight will be described later.

  The calculation unit 12 includes a hierarchical link generation unit 103, a graph arrangement unit 104, and a hierarchy visualization unit 105.

  The hierarchical link generation unit 103 generates hierarchical link information based on the initial link information stored in the initial link storage unit 100 and the cluster information stored in the cluster storage unit 101. The generated hierarchical link information is stored in the hierarchical link storage unit 102. The initial link information includes the links included in the initial link information and the links that are added in a bird's-eye view by the hierarchical link generation unit 103 reflecting the relationship between individual nodes.

  Based on the hierarchical link information stored in the hierarchical link storage unit 102, the graph arrangement unit 104 calculates the arrangement of the graph by reflecting the added link. As the hierarchical link to be reflected in the graph, the initial link and the added hierarchical link are used. Further, a weight (emphasis) may be given to the added hierarchical link, and the graph may be arranged by reflecting the weight. FIG. 6 is an explanatory diagram showing the network shown in FIG. 5 arranged by the graph arrangement unit 104. In the figure, a hierarchical link when there is an initial link between two nodes is indicated by a solid line, and a hierarchical link added without an initial link is indicated by a broken line.

  The hierarchy visualization unit 105 hierarchizes into a plurality of hierarchies based on the arrangement calculated by the graph arrangement unit 104 and the hierarchy link, and visualizes the graph for each hierarchy as necessary. At this time, the visualization position of each node is fixedly handled in the graph, and is processed so as to be the same position even if the hierarchy is changed. FIG. 7 is an explanatory diagram illustrating an example of the first layer visualized by the hierarchy visualization unit 105. FIG. 8 is an explanatory diagram illustrating an example of the second layer visualized by the hierarchy visualization unit 105. FIG. 9 is an explanatory diagram showing an example of the third layer visualized by the hierarchy visualization unit 105. In FIG. 3 visualized based on the initial link, the relationship between the clusters A, B, C, and D was not easily visualized. However, FIG. 7 makes it easy to understand that there is a relationship between cluster A and cluster C, or between cluster B and cluster C, and there is no direct relationship between cluster A and cluster B. can do. Further, the number of nodes (quantity) included in the cluster A, the cluster B, and the cluster C can be easily visualized by FIG. Further, according to FIG. 9, the cluster D has a small number of nodes (quantity), and by comparing the position with FIG. 8, it is possible to easily visualize that the cluster D is related to the cluster D. it can.

  Further, the hierarchical visualization unit 105 acquires hierarchical link information generated by the hierarchical link generation unit 103 and recorded in the hierarchical link storage unit 102, and for links not included in the initial link information (extracted hierarchical links), Virtual nodes may be additionally arranged in the hierarchical link and visualized as a graph. At this time, it is preferable to extract similar word candidates from the nodes (hierarchical link source and hierarchical link destination) at both ends, and narrow down the words to be displayed on the graph for the nodes. This wording can narrow down candidates by referring to the wording of the node adjacent to the hierarchical link on the graph. For example, the graphs may be arranged so that the graph theoretical distance of each node is the shortest, and the words of the nodes adjacent to the virtual node in a planar and hierarchical manner may be referred to. In addition to the words of nodes that are close to each other in a planar or hierarchical manner, the cluster to which they belong may be referenced. Moreover, it is not necessary to handle these words uniformly, and weights may be given to the upper layer, the isotopic layer, the lower layer, the link source, and the link destination, respectively. The weighting may be determined as appropriate according to the relationship between nodes to be extracted and what kind of node is less related. Further, in parallel with the wording of the virtual node, the extracted hierarchical link may be given a direction by determining a link source and a link destination. At this time, the wording of the link source node may be weighted with respect to the text of the link destination.

  Next, the overall operation of the first embodiment will be described in detail with reference to the flowchart of FIG. In this embodiment, a case where a specified layer is visualized by arranging a network in a three-layer structure will be described as an example. In addition, by preparing clusters in several layers in the cluster storage unit 101, it is possible to visualize even in the case of other numbers of layers by a similar method.

  First, the hierarchical link generation unit 103 determines whether or not there is a hierarchical link between all nodes based on the initial link information stored in the initial link storage unit 100 and the cluster information stored in the cluster storage unit 101. A value is calculated (step A1).

  Next, the graph arrangement unit 104 arranges nodes and hierarchical links on the space based on the generated hierarchical link information (step A2). Here, before setting display nodes and links for each layer, nodes and links are arranged. Therefore, nodes and links are arranged at the same position in all layers. The graph layout may be performed using graph layout software such as Graphviz.

Next, the hierarchy visualization unit 105 initializes N indicating the layer number to 1 (step A3).
Next, the hierarchy visualization unit 105 compares N with 3 corresponding to the maximum number of hierarchies (step A4). When N is 3 or less, the process proceeds to the next step A5. If N is greater than 3, the operation is terminated.

Next, the hierarchy visualization unit 105 sets a node to be displayed on the Nth layer (step A5). Details of the operation will be described later.
Next, the hierarchy visualization unit 105 sets a link to be displayed on the Nth layer (step A6). Details of the operation will be described later.
Next, the hierarchy visualization unit 105 sets the displayed link corresponding to the initial link to a solid line (step A7).
Next, the hierarchy visualization unit 105 sets a link to be displayed (added link) that was not included in the initial link as a broken line (step A8).

  Next, if the Nth layer is a layer designated to be displayed, the hierarchy visualization unit 105 displays only the nodes and links in this layer among the arranged nodes and links (step A9).

  Next, the hierarchy visualization unit 105 increments N indicating the layer number (step A10). Then, the process proceeds to step A4.

  By operating in this way, it is possible to create a graph of a network structure showing a predetermined hierarchy by generating hierarchical link information from information given in advance. As a result, the hierarchical link information shown in FIG. 5 is generated, and any of the designated hierarchical structures shown in FIGS. 7 to 9 can be visualized. For visualization, all levels may be prepared so as to be visible sequentially, and the levels may be presented each time based on designation from the operator.

  Next, the operation of the hierarchical link generation unit 103 that generates a hierarchical link between the node i and the node j according to the first embodiment will be described in detail with reference to the flowchart of FIG.

First, the hierarchical link generation unit 103 initializes the layered link value Vij to 0 (step A11).
Next, the hierarchical link generation unit 103 refers to the initial link storage unit 100 to check whether there is an initial link between the node i and the node j (step A12). If there is an initial link, the process proceeds to the next step A13. When there is no initial link, the process proceeds to the next step A14.

  Next, the hierarchical link generation unit 103 sets the layered link value Vij to 2 (step A13). Here, the case where the value of the hierarchical link is 2 will be described as an example if there is an initial link, but this value may be other than 2. When there is an initial link, the value of the hierarchical link may be specified by the user of the system through a keyboard or the like, or may be stored as a default in the system, or otherwise. May be specified.

  Next, the hierarchical link generation unit 103 checks whether the node i matches the node j, that is, whether i and j are the same (step A14). If they match, the operation of the hierarchical link generation unit 103 that generates a hierarchical link between the node i and the node j is terminated. If not, the process proceeds to the next step A15.

  Next, the hierarchical link generation unit 103 refers to the cluster storage unit 101 and checks whether the node i and the node j belong to the same cluster (step A15). If the nodes are in the same cluster, the process proceeds to step A16. If the nodes are not in the same cluster, the process proceeds to step A17.

  The hierarchical link generation unit 103 adds 3 to the layered link value Vij (step A16). Then, the operation of the hierarchical link generation unit 103 that generates a hierarchical link between the node i and the node j is ended. Here, a case where 3 is added to the value of the hierarchical link if belonging to the same cluster will be described as an example, but this value may be other than 3. The value to be added to the hierarchical link when belonging to the same cluster may be specified by the user of the system via a keyboard, or may be stored as a default in the system, or otherwise It may be specified in the following manner.

  The hierarchical link generation unit 103 checks whether the layered link value Vij is 0 (step A17). If Vij is 0, the process proceeds to the next step A18. When Vij is not 0, the operation of the hierarchical link generation unit 103 that generates a hierarchical link between the node i and the node j is terminated.

  Next, the hierarchical link generation unit 103 refers to the cluster storage unit 101 and checks whether both the node i and the node j are cluster representatives (step A18). If both are cluster representatives, the process proceeds to the next step A19. If neither of them is a cluster representative, the operation of the hierarchical link generation unit 103 that generates a hierarchical link between the node i and the node j is terminated.

  Next, the hierarchical link generation unit 103 initializes the node number p to 0 (step A19). Here, the following operation will be described by taking as an example a case where there is a hierarchical link between cluster representatives when there is an initial link between cluster elements. In addition to this, when the ratio of links between cluster elements is equal to or greater than a predetermined threshold, there may be hierarchical links between cluster representatives. Further, the weight assigned to the hierarchical link between the cluster representatives may be adjusted based on the average value or the total of the importance of the element nodes belonging to the cluster.

  Next, the hierarchical link generation unit 103 checks whether p is less than the number N of nodes (step A20). If p is less than N, the process proceeds to the next step A21. If p is greater than or equal to N, the operation of the hierarchical link generation unit 103 that generates a hierarchical link between the node i and the node j is terminated.

  Next, the hierarchical link generation unit 103 refers to the cluster storage unit 101 and checks whether the node p and the node i are nodes of the same cluster (step A21). If the nodes are in the same cluster, the process proceeds to step A23. If the nodes are not in the same cluster, the process proceeds to the next step A22.

  Next, the hierarchical link generation unit 103 increments p and proceeds to Step A20 (Step A22).

  The hierarchical link generation unit 103 initializes the node number q to 0 (step A23).

  Next, the hierarchical link generation unit 103 checks whether q is less than N (step A24). If q is less than N, the process proceeds to step A25. When q is N or more, the process proceeds to step A22.

  If q is less than N, the hierarchical link generation unit 103 refers to the cluster storage unit 101 and checks whether the node q and the node j are nodes of the same cluster (step A25). If the nodes are in the same cluster, the process proceeds to step A27. If the nodes are not in the same cluster, the process proceeds to the next step A26.

  The hierarchical link generation unit 103 increments q and proceeds to Step A24 (Step A26).

  The hierarchical link generation unit 103 refers to the initial link storage unit 100 and checks whether there is an initial link in the link Lpq between the node p and the node q (step A27). If there is an initial link, the process proceeds to step A28. If there is no initial link, the process proceeds to step A26.

  The hierarchical link generation unit 103 sets the layered link value Vij to 1 (step A28). Then, the operation of the hierarchical link generation unit 103 that generates a hierarchical link between the node i and the node j is ended.

  By operating in this way, it is possible to appropriately weight the hierarchical link information from information given in advance.

  Next, with reference to the flowchart of FIG. 12, the operation of the hierarchy visualization unit 105 for setting the display node and the display link of the first layer according to the first embodiment will be described in detail.

  First, the hierarchy visualization unit 105 initializes a number i indicating a node to 0 (step A29).

  Next, the hierarchy visualization unit 105 compares i with the number N of nodes (step A30). If i is less than N, the process proceeds to the next step A31. If i is greater than or equal to N, the process proceeds to step A36.

  Next, the hierarchy visualization unit 105 refers to the cluster storage unit 101 and checks whether i is a cluster representative (step A31). In the case of a cluster representative, the process proceeds to the next step A32. If it is not a cluster representative, the process proceeds to step A34.

  Next, the hierarchy visualization unit 105 refers to the cluster storage unit 101 and checks whether the number of elements in the cluster including i is equal to or greater than a predetermined threshold (step A32). The threshold value may be designated by the user of the system through a keyboard or the like, may be stored in the system as a default, or may be designated in other manners. For example, if the threshold is 2, only the representative nodes of the cluster including a plurality of elements are displayed. Note that here, a case where a representative node of a cluster including the number of elements equal to or greater than a predetermined threshold is displayed as an example. For example, display / non-display is given in advance for each cluster, and the display node is determined. May be. If it is equal to or greater than the threshold, the process proceeds to the next step A33. If it is less than the threshold, the process proceeds to step A34.

  The hierarchy visualization unit 105 sets the node i to display (step A33).

  The hierarchy visualization unit 105 sets the node i to non-display (Step A34).

  Next, the hierarchy visualization unit 105 increments i (step A35). Then, the process proceeds to step A30.

  The hierarchy visualization unit 105 initializes i to 0 again (step A36).

  Next, the hierarchy visualization unit 105 compares i with the number N of nodes (step A37). If i is less than N, the process proceeds to the next step A38. If i is greater than or equal to N, the operation of the hierarchy visualization unit 105 that sets the display node and display link of the first layer is terminated.

  Next, the hierarchy visualization unit 105 initializes the number j indicating another node to 0 (step A38).

  Next, the hierarchy visualization unit 105 compares j with the number N of nodes (step A39). If j is less than N, the process proceeds to the next step A41. If j is greater than or equal to N, the process proceeds to step A40.

  The hierarchy visualization unit 105 increments i (step A40). Then, the process proceeds to step A37.

  Next, the hierarchy visualization unit 105 checks whether both the node i and the node j are set to display (step A41). If both are set to display, the process proceeds to the next step A42. If one or both are set to non-display, the process proceeds to step A43.

  Next, the hierarchy visualization unit 105 sets the link between the node i and the node j to display (step A42).

  Next, the hierarchy visualization unit 105 sets the link between the node i and the node j to non-display (step A43).

  Next, the hierarchy visualization unit 105 increments j (step A44). Then, the process proceeds to step A39.

  Next, with reference to the flowchart of FIG. 13, the operation of the hierarchy visualization unit 105 for setting the display nodes and display links in the second layer according to the first embodiment will be described in detail.

  First, the hierarchy visualization unit 105 initializes a number i indicating a node to 0 (step A45).

  Next, the hierarchy visualization unit 105 compares i with the number N of nodes (step A46). If i is less than N, the process proceeds to the next step A47. If i is greater than or equal to N, the process proceeds to step A51.

  Next, the hierarchy visualization unit 105 refers to the cluster storage unit 101 and checks whether the number of elements in the cluster included in the node i is equal to or greater than a predetermined threshold (step A47). The threshold value may be designated by the user of the system through a keyboard or the like, may be stored in the system as a default, or may be designated in other manners. For example, if the threshold is 2, only the nodes of the cluster including a plurality of elements are displayed. In addition, here, a case where a cluster node including the number of elements equal to or greater than a predetermined threshold is displayed will be described as an example. For example, display / non-display is given in advance for each cluster, and a display node is determined. Also good. If it is equal to or greater than the threshold, the process proceeds to the next step A48. If it is less than the threshold, the process proceeds to step A49.

  The hierarchy visualization unit 105 sets the node i to display (step A48). When the node i is a cluster representative, it may be set to be highlighted such as a double circle.

  The hierarchy visualization unit 105 sets node i to non-display (step A49).

  Next, the hierarchy visualization unit 105 increments i (step A50). Then, control goes to a step A46.

  Next, the hierarchy visualization unit 105 initializes i to 0 again (step A51).

  Next, the hierarchy visualization unit 105 compares i with the number N of nodes (step A52). If i is less than N, the process proceeds to the next step A53. If i is greater than or equal to N, the operation of the hierarchy visualization unit 105 for setting the second layer display node and the display link is terminated.

  Next, the hierarchy visualization unit 105 initializes the number j indicating another node to 0 (step A53).

  Next, the hierarchy visualization unit 105 compares j with the number N of nodes (step A54). If j is less than N, the process proceeds to step A61. If j is greater than or equal to N, the process proceeds to step A55.

  The hierarchy visualization unit 105 increments i (step A55). And it transfers to step A52.

  Next, the hierarchy visualization unit 105 checks whether both the node i and the node j are set to display (step A56). If both are set to display, the process proceeds to the next step A57. If one or both are set to non-display, the process proceeds to step A61.

  Next, the hierarchy visualization unit 105 refers to the cluster storage unit 101 and checks whether the node i and the node j are included in the same cluster (step A57). In the case of the same cluster, the process proceeds to step A59. If they are not the same cluster, the process proceeds to step A58.

  The hierarchy visualization unit 105 refers to the hierarchy link storage unit 102 and checks whether there is an initial link between the node i and the node j (step A58). Here, in the example described in the present embodiment, the hierarchical link generation unit 103 adds Vij = 2 when there is an initial link and further adds 3 when there is the same cluster. The value of the hierarchical link is 5 or 2. If it is not possible to determine whether there is an initial link based on the value of the hierarchical link, the initial link storage unit 100 may be referred to. If there is an initial link, the process proceeds to step A60. If there is no initial link, the process proceeds to step A61.

  The hierarchy visualization unit 105 sets the link between the node i and the node j to a solid line display (step A59).

  The hierarchy visualization unit 105 sets the link between the node i and the node j to a broken line display (step A60). Here, a case in which links in the same cluster are indicated by solid lines and other initial links are indicated by broken lines will be described as an example. However, both cases may be indicated by solid lines. Also, both node i and node j are cluster representatives, there is no initial link, and there is a hierarchical link. That is, in the example described in this embodiment, the value of the hierarchical link may be 1. Alternatively, it may be set as a broken line display.

  The hierarchy visualization unit 105 sets the link between the node i and the node j to non-display (step A61).

  Next, the hierarchy visualization unit 105 increments j (step A62). Then, control goes to a step A54.

  Next, the operation of the hierarchy visualization unit 105 for visualizing the third layer of the first embodiment will be described in detail with reference to the flowchart of FIG.

  First, the hierarchy visualization unit 105 initializes a number i indicating a node to 0 (step A63).

  Next, the hierarchy visualization unit 105 compares i with the number N of nodes (step A64). If i is less than N, the process proceeds to the next step A65. If i is greater than or equal to N, the process proceeds to step A69.

  Next, the hierarchy visualization unit 105 refers to the cluster storage unit 101 and checks whether the number of elements in the cluster included in the node i is less than a predetermined threshold (step A65). The threshold value may be designated by the user of the system through a keyboard or the like, may be stored in the system as a default, or may be designated in other manners. For example, if the threshold is 2, only the nodes of a single element cluster are displayed. Here, a case where a representative node of a cluster including the number of elements equal to or greater than a predetermined threshold is displayed as an example. For example, display / non-display is given in advance for each cluster, and the second layer You may display the node which was not displayed. If it is less than the threshold, the process proceeds to the next step A66. If it is equal to or greater than the threshold, the process proceeds to step A67.

  The hierarchy visualization unit 105 sets the node i to display (step A66). When the node i is a cluster representative, it may be set to be highlighted such as a double circle.

  The hierarchy visualization unit 105 sets node i to non-display (step A67).

  Next, the hierarchy visualization unit 105 increments i (step A68). Then, the process proceeds to step A64.

  Next, the hierarchy visualization unit 105 initializes i to 0 again (step A69). Note that the steps after this step may be omitted, and only the node may be displayed and the link may not be displayed.

  Next, the hierarchy visualization unit 105 compares i with the number N of nodes (step A70). If i is less than N, the process proceeds to the next step A71. If i is greater than or equal to N, the operation of the hierarchy visualization unit 105 for setting the second layer display node and the display link is terminated.

  Next, the hierarchy visualization unit 105 initializes the number j indicating another node to 0 (step A71).

  Next, the hierarchy visualization unit 105 compares j with the number N of nodes (step A72). If j is less than N, the process proceeds to step A74. If j is greater than or equal to N, the process proceeds to step A73.

  The hierarchy visualization unit 105 increments i (step A73). And it transfers to step A70.

  Next, the hierarchy visualization unit 105 checks whether both the node i and the node j are set to display (step A74). If both are set to display, the process proceeds to step A75. If one or both are set to non-display, the process proceeds to step A77.

  The hierarchical visualization unit 105 refers to the hierarchical link storage unit 102 and checks whether there is a hierarchical link between the node i and the node j (step A75). If there is a hierarchical link, the process proceeds to step A76. If there is no hierarchical link, the process proceeds to step A77.

  The hierarchy visualization unit 105 sets the link between the node i and the node j to display (step A76).

  The hierarchy visualization unit 105 sets the link between the node i and the node j to non-display (step A77).

  Next, the hierarchy visualization unit 105 increments j (step A78). Then, control goes to a step A72.

Thus, according to this embodiment, it is possible to visualize a network composed of nodes and links so that the relevance of each other can be overlooked, including nodes with low relevance. In the first layer, only representatives of clusters having the number of elements equal to or greater than the threshold value are displayed, so that the overall schematic structure can be visualized. Since the second layer displays the elements of the cluster whose number of elements is equal to or greater than the threshold value, the amount of partial nodes relative to the whole can be easily visualized. Since the third layer displays the elements of the cluster whose number of elements is less than the threshold value, it is possible to easily visualize a node having a high possibility of leakage or inconsistency. In addition, since the positional relationship between the first layer, second layer, and third layer nodes is the same, for example, it is possible to visualize where the number of nodes with a high possibility of leakage or mismatching is high. can do.
For this reason, it is possible to extract a hidden relevance other than those that can directly acquire data, such as other links and groups, with respect to nodes that are less relevant to other nodes in the information given in advance. In addition, each layer can be switched, and even in a hierarchy where a specific node is not displayed, the position of the node is made into a graph, so that it is possible to overlook the recognition of hidden reciprocal relationships that are identified.

  Next, the present invention will be described with a visualization example. FIG. 15 is an explanatory diagram showing a first layer when the present embodiment is applied to text analysis as an example. In this example, for the free-form questionnaire survey results related to the office environment, texts that are the individual response results are used as nodes, links are created between nodes with high similarity, and clusters are created between nodes with extremely high similarity. This is a part of the application of this embodiment after creation. In the figure, representative words (keywords) of the cluster are shown in individual nodes that are cluster representatives. As shown in the figure, it is easy to understand that the thin disker desktop is related to environmental commuting and environmental air conditioning, the temperature is related to environmental air conditioning, and the scheduler is related to PC. In addition, regarding the relationship between environmental communication and the desktop PC, a relationship that cannot be found at first glance is visualized.

  FIG. 16 is an explanatory diagram showing an example of the first layer when the present embodiment is applied to a request / specification of sales management and visualized as an example. FIG. 17 is an explanatory diagram showing an example of the third layer when the present embodiment is applied and visualized to detect the omission of requirements / specifications for the same sales management requirements / specifications as FIG.

  Here, in general, the business contents of “ordering / sales processing” are the development of new customers, formulation of sales plans, proposal / price negotiations, delivery date response to customer requests and delivery date adjustments, finalization of order contents, etc. The business contents in the “ordering / purchase processing” include finding a new supplier, formulating a purchase plan, negotiating a price, answering a delivery date to a site request and adjusting a delivery date, and confirming the order content. In this way, “order / sales processing” and “order / purchase processing” are just “orders” and “orders”, “sales” and “purchases”, etc., and there are many common tasks. . Therefore, for many requirements and specifications related to sales management, a request / specification is used as a node, and a link is created between requirements / specifications with high similarity. When the present embodiment is applied after generating a cluster, etc., as shown in FIG. 16, in the first layer, “order / sales processing” and “order / purchase processing” are visualized symmetrically, Can be found to be symmetric. In addition, as shown in the figure, the relationship between "inventory data" and "checkout process" is found and visualized from the relationship between the store inventory data used in the checkout process and the illustrated inventory data. Yes.

And, for example, the requirements and specifications related to “cash register processing” and “inventory inquiry function” required for both “order / sales processing” and “order / purchase processing” are specified only for “order / sales processing”. As shown in FIG. 17, each node is arranged so that “Inventory inquiry” is displayed in the third layer, and “Register processing” and “Inventory inquiry function” appear on the left side of the figure. It can be seen that the symmetry of the system is impaired, that is, it can be easily visualized that there is a possibility that requirements and specifications are leaked.
As shown in these examples, this embodiment makes it possible to visualize a network composed of nodes and links so that the relevance of each other, including nodes with low relevance, can be overlooked on the graph. it can.

  The extraction of requirements and specifications related to sales management may be obtained from a document by natural language analysis, may be specified by a system user through a keyboard or the like, or stored as a default in the system. It may be specified, or may be specified in other manners. When using a document, the network visualization system of this embodiment is used to arrange each node in a variety of different formats, and the relationship between nodes that is difficult for humans to find. It becomes possible to look down on the nature.

  In addition, what is necessary is just to implement | achieve each part of a network visualization system using the combination of hardware and software. In a form in which hardware and software are combined, a network visualization program is developed in the RAM, and each unit is realized as various means by operating hardware such as a control unit (CPU) based on the program. The network visualization program may be configured to cause each of the processes to be executed by an operating system or other general software.

  The program may be recorded on a recording medium and distributed. The program recorded on the recording medium is read into a memory via a wired, wireless, or recording medium itself, and operates a control unit or the like. Examples of the recording medium include an optical disk, a magnetic disk, a semiconductor memory device, and a hard disk.

  If the above embodiment is described in another expression, an information processing apparatus that operates as a network visualization system is based on a network visualization program developed in a RAM, and a control unit as a hierarchical link generation unit, a graph arrangement unit, and a hierarchy visualization unit It is possible to realize this by operating.

  As described above, according to the network visualization system according to the present invention, it is possible to visualize a network composed of nodes and links so that the relevance of each other can be overlooked, including nodes with low relevance.

  In addition, the specific configuration of the present invention is not limited to the above-described embodiment, and changes within a range not departing from the gist of the present invention are included in the present invention. Further, a desired effect can be obtained by appropriately combining a plurality of components. For example, some components of all the components shown in the embodiment may be integrated or deleted.

In addition, a part or all of the above-described embodiments can be described as follows. Note that the following supplementary notes do not limit the present invention.
[Appendix 1]
A link that is not included in the link information based on link information indicating a predetermined relationship between nodes and cluster information obtained by grouping individual nodes with respect to a network composed of nodes and links. A hierarchical link generation means for generating a hierarchical link by extracting from the relationship between nodes and links included in the cluster information;
Reflecting the generated hierarchical link, graph placement means for placing individual nodes in space,
A hierarchy visualization means for displaying the arranged nodes and each link for each hierarchy to which the nodes belong without changing the display position of the nodes according to the hierarchy;
A network visualization system comprising:

[Appendix 2]
The network visualization system according to the above supplementary note, wherein the hierarchy visualization means selects a hierarchy to which each node belongs based on the cluster information.

[Appendix 3]
The cluster information is given a representative node for each cluster,
The hierarchical link generator generates a hierarchical link between two nodes, link information between the two nodes, cluster information of the two nodes, and two nodes when the two nodes are representative nodes of the cluster. The network visualization system according to the above supplementary note, wherein the network visualization system is generated based on link information of elements of both clusters to which the cluster belongs.

[Appendix 4]
The hierarchical link generation means generates a hierarchical link between two nodes as a weight,
The network visualization system according to the above supplementary note, wherein the graph arrangement unit arranges nodes and links based on weights of hierarchical links.

[Appendix 5]
The hierarchical link generation means initializes the weight of a hierarchical link between two nodes, and when there is a link between two nodes, the hierarchical link weight is set to a first value, and the two nodes belong to the same cluster. In some cases, the second value is added to the weight of the hierarchical link, otherwise the hierarchical link if the two nodes are representatives of different clusters and there is a link between the elements of the two clusters. The network visualization system as described in the above supplementary note, wherein the weight of is set to a third value.

[Appendix 6]
There are three levels of relevance between individual nodes.
In the first layer, the cluster representative node and the hierarchical link between the cluster representatives are displayed,
In the second layer, the hierarchical link between the cluster element nodes having the number of elements equal to or greater than the predetermined threshold and the cluster representative and the original link having both ends of the element nodes of the cluster having the number of elements equal to or greater than the predetermined threshold are displayed. ,
The third layer displays the element nodes of the cluster having the number of elements less than the predetermined threshold and the original link having both ends of the element nodes of the cluster having the number of elements less than the predetermined threshold in the third layer. The network visualization system described.

[Appendix 7]
For the extracted hierarchical link, a virtual node is additionally arranged in the hierarchical link, and similar wording candidates are extracted from the nodes at both ends of the virtual link, and the wording of the node adjacent to the hierarchical link on the graph theoretical distance The network visualization system as described in the above supplementary note, wherein the display words of the additionally arranged virtual nodes are narrowed down from the extracted candidates with reference to the similarity.

[Appendix 8]
When narrowing the display wording by arranging additional virtual nodes,
With reference to the link direction included in the link information set in advance, for the hierarchical link where the virtual node is placed, the link source and link destination are determined and the direction is given, and the word of the link source node The network visualization system as described in the above supplementary note, wherein the display wording is narrowed down by assigning a weight to the wording of the link destination.

[Appendix 9]
A link that is not included in the link information based on link information indicating a predetermined relationship between nodes and cluster information obtained by grouping individual nodes with respect to a network composed of nodes and links. A hierarchical link generation step of generating a hierarchical link by extracting from the relationship between nodes and links included in the cluster information;
A graph placement step of reflecting individual generated hierarchical links and placing individual nodes in space;
A layer visualization step for displaying the arranged nodes and each link for each layer to which the nodes belong without changing the display position of the node according to the layer,
A network visualization method comprising:

[Appendix 10]
The network visualization method according to the above supplementary note, wherein, in the hierarchy visualization step, a hierarchy to which each node belongs is selected based on the cluster information.

[Appendix 11]
In the cluster information, a representative node is given for each cluster,
When generating a hierarchical link, the hierarchical link between two nodes is obtained by linking information between the two nodes, cluster information of the two nodes, and two nodes when both nodes are representative nodes of the cluster. The network visualization method according to the above supplementary note, wherein the network visualization method is generated based on link information included in elements of both clusters to which the data belongs.

[Appendix 12]
When generating a hierarchical link, generate a hierarchical link between two nodes as a weight,
The network visualization method according to the above supplementary note, wherein when arranging each node on a graph, the node and the link are arranged based on a weight of the hierarchical link.

[Appendix 13]
When creating a hierarchical link, initialize the weight of the hierarchical link between the two nodes, and if there is a link between the two nodes, the weight of the hierarchical link is the first value and the two nodes are in the same cluster. If it belongs, add the second value to the weight of the hierarchical link; otherwise, if the two nodes are representatives of different clusters and there is a link between the elements of the two clusters The network visualization method according to the above supplementary note, wherein the link weight is set to a third value.

[Appendix 14]
There are three levels of relevance between individual nodes.
In the first layer, the cluster representative node and the hierarchical link between the cluster representatives are displayed,
In the second layer, the hierarchical link between the cluster element nodes having the number of elements equal to or greater than the predetermined threshold and the cluster representative and the original link having both ends of the element nodes of the cluster having the number of elements equal to or greater than the predetermined threshold are displayed. ,
The third layer displays the element nodes of the cluster having the number of elements less than the predetermined threshold and the original link having both ends of the element nodes of the cluster having the number of elements less than the predetermined threshold in the third layer. The network visualization method described.

[Appendix 15]
For the extracted hierarchical link, a virtual node is additionally arranged in the hierarchical link, and similar wording candidates are extracted from the nodes at both ends of the virtual link, and the wording of the node adjacent to the hierarchical link on the graph theoretical distance The network visualization method according to the above supplementary note, wherein the display words of the additionally arranged virtual nodes are narrowed down from the extracted candidates with reference to the similarity.

[Appendix 16]
When narrowing the display wording by arranging additional virtual nodes,
With reference to the link direction included in the link information set in advance, for the hierarchical link where the virtual node is placed, the link source and link destination are determined and the direction is given, and the word of the link source node The network visualization method as set forth in the above supplementary note, wherein the display wording is narrowed down by assigning a weight to the wording of the link destination.

[Appendix 17]
In the information processing device,
A link that is not included in the link information based on link information indicating a predetermined relationship between nodes and cluster information obtained by grouping individual nodes with respect to a network composed of nodes and links. A hierarchical link generation process for generating a hierarchical link by extracting from the relationship between nodes and links included in the cluster information;
Reflecting the generated hierarchical link, graph placement processing to place individual nodes in space,
A hierarchy visualization process that displays the arranged nodes and each link for each hierarchy to which the nodes belong without changing the display position of the nodes according to the hierarchy,
A network visualization program characterized by having

[Appendix 18]
The network visualization program according to the above supplementary note, wherein in the hierarchy visualization process, a hierarchy to which each node belongs is selected based on the cluster information.

[Appendix 19]
In the cluster information, a representative node is given for each cluster,
When generating a hierarchical link, the hierarchical link between two nodes is obtained by linking information between the two nodes, cluster information of the two nodes, and two nodes when both nodes are representative nodes of the cluster. The network visualization program according to the above supplementary note, wherein the network visualization program is generated based on link information possessed by elements of both clusters to which.

[Appendix 20]
When generating a hierarchical link, generate a hierarchical link between two nodes as a weight,
The network visualization program as described in the above supplementary note, wherein when placing each node on a graph, the node and the link are arranged based on the weight of the hierarchical link.

[Appendix 21]
When creating a hierarchical link, initialize the weight of the hierarchical link between two nodes, and if there is a link between two nodes, the weight of the hierarchical link is the first value and the two nodes are in the same cluster If the two nodes are representatives of different clusters and there is a link between the elements of the two clusters, the second value is added to the weight of the hierarchical link. The network visualization program as described in the above supplementary note, wherein the weight of the hierarchical link is set to a third value.

[Appendix 22]
There are three levels of relevance between individual nodes.
In the first layer, the representative node of the cluster and the hierarchical link between the cluster representatives are displayed,
In the second layer, the hierarchical link between the cluster element nodes having the number of elements equal to or greater than the predetermined threshold and the cluster representative and the original link having the element nodes of the cluster having the number of elements equal to or greater than the predetermined threshold as both ends are displayed. ,
In the third layer, the element node of the cluster having the number of elements less than the predetermined threshold and the original link having both ends of the element node of the cluster having the number of elements less than the predetermined threshold are displayed. The network visualization program described.

[Appendix 23]
For the extracted hierarchical link, a virtual node is additionally arranged in the hierarchical link, and similar wording candidates are extracted from the nodes at both ends of the virtual link, and the wording of the node adjacent to the hierarchical link on the graph theoretical distance The network visualization program as described in the above supplementary note, wherein the display wording of the additionally arranged virtual node is processed so as to be narrowed down from the extracted candidates with reference to the similarity.

[Appendix 24]
When narrowing the display wording by arranging additional virtual nodes,
With reference to the link direction included in the link information set in advance, for the hierarchical link where the virtual node is placed, the link source and link destination are determined and the direction is given, and the word of the link source node The network visualization program as described in the above supplementary note, wherein the display wording is narrowed by giving a weight to the wording of the link destination.

[Appendix 25]
A link that is not included in the link information based on link information indicating a predetermined relationship between nodes and cluster information obtained by grouping individual nodes with respect to a network composed of nodes and links. A hierarchical link generation process for generating a hierarchical link by extracting from the relationship between nodes and links included in the cluster information;
Reflecting the generated hierarchical link, graph placement processing to place individual nodes in space,
A hierarchy visualization process that displays the arranged nodes and each link for each hierarchy to which the nodes belong without changing the display position of the nodes according to the hierarchy,
A recording medium having recorded thereon a program characterized by causing an information processing apparatus to perform the operation.

11 Storage Unit 12 Calculation Unit 100 Initial Link Storage Unit (Initial Link Storage Unit)
101 Cluster storage unit (cluster storage means)
102 Hierarchical link storage unit (hierarchical link storage means)
103 Hierarchical link generator (hierarchical link generator)
104 Graph arrangement part (graph arrangement means)
105 Hierarchy visualization part (hierarchy visualization means)

Claims (10)

A link that is not included in the link information based on link information indicating a predetermined relationship between nodes and cluster information obtained by grouping individual nodes with respect to a network composed of nodes and links. A hierarchical link generation means for generating a hierarchical link by extracting from the relationship between nodes and links included in the cluster information;
Reflecting the generated hierarchical link, graph placement means for placing individual nodes in space,
A hierarchy visualization means for displaying the arranged nodes and each link for each hierarchy to which the nodes belong without changing the display position of the nodes according to the hierarchy;
A network visualization system comprising:   The network visualization system according to claim 1, wherein the hierarchy visualization unit selects a hierarchy to which each node belongs based on the cluster information. The cluster information is given a representative node for each cluster,
The hierarchical link generator generates a hierarchical link between two nodes, link information between the two nodes, cluster information of the two nodes, and two nodes when the two nodes are representative nodes of the cluster. The network visualization system according to claim 2, wherein the network visualization system is generated based on link information of elements of both clusters to which the cluster belongs. The hierarchical link generation means generates a hierarchical link between two nodes as a weight,
The network visualization system according to any one of claims 1 to 3, wherein the graph arrangement unit arranges nodes and links based on weights of hierarchical links.   The hierarchical link generation means initializes the weight of a hierarchical link between two nodes, and when there is a link between two nodes, the hierarchical link weight is set to a first value, and the two nodes belong to the same cluster. In some cases, the second value is added to the weight of the hierarchical link, otherwise the hierarchical link if the two nodes are representatives of different clusters and there is a link between the elements of the two clusters. The network visualization system according to claim 4, wherein the weight of is a third value. There are three levels of relevance between individual nodes.
In the first layer, the cluster representative node and the hierarchical link between the cluster representatives are displayed,
In the second layer, the hierarchical link between the cluster element nodes having the number of elements equal to or greater than the predetermined threshold and the cluster representative and the original link having both ends of the element nodes of the cluster having the number of elements equal to or greater than the predetermined threshold are displayed. ,
The third layer displays an element node of a cluster having an element number less than a predetermined threshold and an original link having both ends of the element node of the cluster having an element number less than a predetermined threshold. The network visualization system according to any one of claims 1 to 5. For the extracted hierarchical link, a virtual node is additionally arranged in the hierarchical link, and similar wording candidates are extracted from the nodes at both ends of the virtual link, and the wording of the node adjacent to the hierarchical link on the graph theoretical distance The network visualization system according to any one of claims 1 to 6, wherein the display words of the additionally arranged virtual nodes are narrowed down from the extracted candidates with reference to the similarity. When narrowing the display wording by arranging additional virtual nodes,
With reference to the link direction included in the link information set in advance, for the hierarchical link where the virtual node is placed, the link source and link destination are determined and the direction is given, and the word of the link source node The network visualization system according to claim 7, wherein the display wording is narrowed down by assigning a weight to the wording of the link destination. A link that is not included in the link information based on link information indicating a predetermined relationship between nodes and cluster information obtained by grouping individual nodes with respect to a network composed of nodes and links. A hierarchical link generation step of generating a hierarchical link by extracting from the relationship between nodes and links included in the cluster information;
A graph placement step of reflecting individual generated hierarchical links and placing individual nodes in space;
A layer visualization step for displaying the arranged nodes and each link for each layer to which the nodes belong without changing the display position of the node according to the layer,
A network visualization method comprising: In the information processing device,
A link that is not included in the link information based on link information indicating a predetermined relationship between nodes and cluster information obtained by grouping individual nodes with respect to a network composed of nodes and links. A hierarchical link generation process for generating a hierarchical link by extracting from the relationship between nodes and links included in the cluster information;
Reflecting the generated hierarchical link, graph placement processing to place individual nodes in space,
A hierarchy visualization process that displays the arranged nodes and each link for each hierarchy to which the nodes belong without changing the display position of the nodes according to the hierarchy,
A network visualization program characterized by having

Images