JP5309780B2 - Display information generating apparatus and program - Google Patents

Description

  The present invention relates to a display information generation device and a program.

A network consisting of these elements and connection relations, with each web page existing on the Internet as an element, a hyperlink relation between web pages as a connection relation, a paper or patent document as an element, and a citation relation of each document as a connection relation May be visualized. For example, Patent Document 1 below discloses a technique for visualizing elements having a connection relationship by arranging them in a low-dimensional space such as two-dimensional or three-dimensional.
JP 2004-318739 A

  In addition to the connection relationship between elements, attribute information such as classification information may be assigned to the element.

  One of the objects of the present invention is to provide a display information generating apparatus and a program for visualizing attribute information other than the connection relationship in addition to the connection relationship between elements.

  In order to achieve the above object, the invention of the display information generating device according to claim 1 is characterized in that means for acquiring information on a plurality of elements and connection relation information between the plurality of elements, Based on the information, the means for classifying the plurality of elements into one or a plurality of element groups, the arrangement area determining means for determining the arrangement area for each of the classified element groups, and the element group for each element group Generating means for generating display information for arranging the classified elements in the determined arrangement area and displaying the arranged elements in a connected state based on the connection relation information, To do.

  Further, the invention according to claim 2 is the display information generating device according to claim 1, wherein the generating means generates the display information to be displayed in a manner indicating a boundary for each element group. And

  According to a third aspect of the present invention, in the display information generating device according to the first or second aspect, the weight of the connection relation between the element groups is used as the connection relation information between the elements included in each element group. The arrangement area determining means further determines an arrangement area for each element group based on the set connection relation weight.

  According to a fourth aspect of the present invention, in the display information generating device according to the third aspect, the arrangement area determining means sets the size of the arrangement area for each element group to the number of elements included in the element group. It is characterized by determining based on.

  In the display information generating device according to any one of claims 1 to 4, the generating unit may arrange the elements included in the element group for each element group. It is characterized in that it is determined so that the total sum of the connection distances becomes small.

  According to a sixth aspect of the present invention, there is provided the program according to claim 6, wherein the plurality of elements are acquired based on the information on the plurality of elements and the connection relation information between the plurality of elements, and the information on the plurality of elements. Are classified into one or a plurality of element groups, arrangement area determining means for determining an arrangement area for each of the classified element groups, and the elements classified into the element groups for each of the element groups are determined. The computer is caused to function as a generation unit that generates display information that is arranged in an arrangement area and that displays the arranged elements in a connected state based on the connection relation information.

  According to the first aspect of the present invention, display information for visualizing attribute information other than the connection relationship is generated in addition to the connection relationship between the elements.

  According to invention of Claim 2, it visualizes in the aspect which recognizes the group for every element group.

  According to the invention described in claim 3, the relationship between the element groups is reflected in the arrangement of the element groups.

  According to the fourth aspect of the present invention, each element group is arranged in a size corresponding to the number of elements.

  According to invention of Claim 5, the connection between elements becomes easy to visually recognize.

  According to the sixth aspect of the invention, the computer functions to generate display information for visualizing attribute information other than the connection relation in addition to the connection relation between the elements.

  DESCRIPTION OF EXEMPLARY EMBODIMENTS Hereinafter, preferred embodiments (hereinafter referred to as embodiments) for carrying out the invention will be described with reference to the drawings.

  FIG. 1 shows a functional block diagram of a display information generation apparatus 10 according to the present embodiment. The display information generation device 10 is configured by a computer including a central processing unit (CPU) and a storage device (memory). As shown in FIG. 1, the display information generation device 10 includes a network data storage unit 12, a classification unit 14, a cluster arrangement region determination unit 16, a node arrangement position determination unit 18, a graph display data generation unit 20, and a display. Part 22 is included. The function of each part is that a program stored in a computer-readable information storage medium is read into the display information generation device 10 which is a computer system using a medium reader (not shown), and instructions of the program are executed by the central processing unit. It may be realized by. Note that the program may be supplied to the display information generation device 10 by an information storage medium, or may be supplied via a data communication network such as the Internet.

  The network data storage unit 12 stores network data including information for each of a plurality of nodes (elements) and connection relationship information that defines a connection relationship between the plurality of nodes. A node is a component of a network, and may correspond to, for example, one piece of electronic data. The node information may include information on the electronic data itself and attribute information given to the electronic data. The connection relationship between the nodes may be expressed by indicating whether there is a relationship between the nodes corresponding to “1” and “0”, or may be expressed by weighting the relationship. For example, when a node is an electronic document, the connection relationship between the nodes may be set based on a citation relationship or a reference relationship between electronic documents.

  The classification unit 14 classifies the plurality of nodes stored in the network data storage unit 12 into one or a plurality of clusters based on the node information. In the present embodiment, the classification unit 14 extracts feature amounts for each node, performs clustering based on the extracted feature amounts, and classifies each node into clusters. For example, when the node is an electronic document, the classification unit 14 generates a feature vector of each document based on a predetermined keyword group for each electronic document, and based on the inter-vector distance of the generated feature vector. Clustering.

  The classification unit 14 is not limited to the clustering described above, and may classify each node. For example, the classification unit 14 may classify each node into a cluster based on attribute information given to the node. In this case, if the node is an electronic document, the classification unit 14 may classify each node into a cluster based on attribute information such as classification information, creator information, creation date, and the like given to the electronic document.

  The cluster arrangement area determination unit 16 determines the arrangement area of each cluster classified by the classification unit 14. The arrangement area of each cluster may be an area in a three-dimensional space or an area in a two-dimensional plane. The cluster placement area determination unit 16 regards each cluster as one node, for example, the paper “T. Fruchterman and E. Reingold.“ Graph Drawing by Force-directed Placement ”, Software-Practice and Experience, 21, 1129- 1164, 1991 "may be used to determine the arrangement of each cluster. Hereinafter, an example of a process for setting the spring length between clusters when the algorithm based on the spring model is used will be described. In this embodiment, for the sake of simplicity, it is assumed that the cluster arrangement region is formed in a circle on a two-dimensional plane.

  The cluster placement area determination unit 16 sets the size of the cluster placement area to a value proportional to the number of elements belonging to the cluster. For example, in the case where the cluster arrangement area is constituted by a circle, the radius of the circle as the arrangement area may be set to a value proportional to the number of cluster elements. In this way, the size of the cluster placement area is set according to the number of elements in the cluster, because the nodes included in some clusters do not become extremely small or conversely large. This is because the node fits in the cluster placement area.

  Next, the cluster arrangement area determination unit 16 calculates the length of the spring connecting the clusters so that the arrangement areas of the clusters having the set radius do not overlap.

FIG. 2 shows the relationship between the arrangement areas of the cluster P and the cluster Q. As shown in FIG. 2, the radius of the arrangement region of the cluster P is r _P , the radius of the arrangement region of the cluster Q is r _Q, and the center of each of the cluster P and the cluster Q is a spring having a length of l _PQ. It will be connected. The natural length l _PQ of the spring connecting the clusters P and Q is determined by the following equation (1).
l _PQ = r _P + r _Q + c / W _PQ (1)
Here, c is a constant, and _WPQ is the weight of the link between the clusters P and Q. A method for calculating the link weight _WPQ between clusters will be described below.

The link weight _WPQ between the clusters is calculated based on the connection relationship between the nodes belonging to the clusters P and Q. Hereinafter, a method of calculating the link weight _WPQ between the clusters P and Q will be described with reference to FIG.

As shown in FIG. 3, it is assumed that the cluster P includes p nodes, the cluster Q includes q nodes, and the number of links between the clusters P and Q is n. At this time, the link weight W _pq between the clusters P and Q may be calculated by the following equation (2).
W _PQ = n / p · q (2)
In the above equation (2), the number n of links existing between the clusters P and Q is divided by the maximum value p · q of the number of links generated between the clusters P and Q, and the cluster does not depend on the number of nodes in the cluster. The weight of the link between them is calculated.

ここで、０≦ρ≦１である。上記式（３）に示されるように、クラスタ間のリンクの重みを設定することで、クラスタ間の関係のうち関連性の小さいものについては除外される。 The link weight W _pq between the clusters P and Q is not limited to the above formula (2), and for example, W _pq calculated by the above formula (2) may be further compared with a threshold value. For example, as shown in the following formula (3), when the calculated W _pq is below the threshold, the link weight may be set to “0” (that is, no link exists).

Here, 0 ≦ ρ ≦ 1. As shown in the above equation (3), by setting the link weight between the clusters, those having less relevance among the relationships between the clusters are excluded.

なお、ノード間のリンクが有向リンクの場合には、ｎ_ｐ→Ｑ、ｎ_Ｑ→Ｐの各向きのリンクに対してリンクの重みを算出することとしてよい。 Further, the link weight W _pq between the clusters P and Q is 1 based on the following equation (4), and is 0 when the link weight calculated by the above equation (2) exceeds the threshold, and 0 otherwise. It is good also as setting the weight of a link as follows. By doing so, the relationship between clusters is simplified and expressed.

In addition, when the link between nodes is a directional link, it is good also as calculating a link weight with respect to the link of each direction of _{np-> Q} and _{nQ-> P.}

  The cluster arrangement region determination unit 16 determines the arrangement region of each cluster based on the spring length set for each cluster by the above-described spring model algorithm.

  Next, the node arrangement position determination unit 18 determines the arrangement positions of the nodes included in each cluster whose arrangement has been determined by the cluster arrangement region determination unit 16. For each cluster, the node arrangement position determination unit 18 calculates the coordinate position of each node belonging to the cluster on the work coordinate space, and calculates the relative positional relationship between the nodes calculated on the work coordinate space. While maintaining the node group, the node group is translated, and if necessary, enlargement / reduction is performed to determine the arrangement coordinates in the cluster arrangement area of each node.

  Specific processing by the node arrangement position determination unit 18 will be described with reference to FIG. 4A shows a node group belonging to the cluster P, and FIG. 4B shows a node group belonging to the cluster Q. First, the node arrangement position determination unit 18 obtains the relative positional relationship of each node based on the connection relationship between the nodes included in the cluster for each cluster. 4A and 4B show examples of the relative positional relationship between the nodes of the cluster P and the cluster Q based on the obtained arrangement coordinates. Note that the spring model described above may be used when determining the relative positional relationship of the nodes. Then, as shown in FIG. 4C, the node arrangement position determining unit 18 keeps each of the node groups within the corresponding cluster arrangement area while maintaining the relative positional relationship of the obtained node groups. Deploy.

  The graph display data generation unit 20 is based on the arrangement area for each cluster determined by the cluster arrangement area determination unit 16, the node arrangement position determined by the node arrangement position determination unit 18, and the connection relationship between the nodes. Generate graph display data that visualizes the relationship between nodes.

  Based on the graph display data generated by the graph display data generation unit 20, the display unit 22 displays a graph related to the processing target node on a display connected to the display information generation device 10. The graph shown in FIG. 5A is an example in which graph display data generated based on only the node connection relation is displayed, and the graph shown in FIG. 5B is related to the present embodiment. It is the example which displayed the graph display data which visualized also about the classification | category relation of the node produced | generated by the display information generation process. In FIG. 5, nodes having a common pattern represent nodes to which the same classification information is assigned.

  As shown in FIGS. 5A and 5B, in FIG. 5A, the classification of each node is not recognized based on the graph display, but in FIG. 5B, each node is not displayed in the graph display. It is arranged according to the classification to which it belongs, and the classification to which each node belongs is clearly shown together with the connection relationship of each node.

  Next, the flow of graph display data generation processing by the display information generation apparatus 10 will be described with reference to the flowchart shown in FIG.

  First, the display information generation apparatus 10 reads network data including information on a plurality of nodes and connection relation information that defines connection relations between the nodes (S101), and selects one or more nodes based on the read network data. (S102). The display information generating apparatus 10 determines an arrangement area for each classified cluster (S103), and further arranges a node belonging to the cluster in the determined arrangement area of the cluster (S104). The display information generating apparatus 10 generates graph display data in a mode in which the nodes are connected based on the connection relation information between the nodes after arranging each node in the corresponding cluster arrangement area (S105). Based on the generated graph display data, the display information generation device 10 arranges nodes for each cluster and displays a graph configured in a manner in which the nodes are connected on the display (S106).

  In the display information generating apparatus 10 according to the present embodiment, the classification information of each node is shown along with the connection relationship between the nodes. Even when a node is added or deleted, it is sufficient to recalculate the arrangement of the cluster to which the changed node belongs, so the processing load is reduced. And, for example, network data composed of patent documents is processed, and the present invention is applied to patent documents and the like to which documents are added for each era. When visualizing a network configured based on the relationship, the position of a newly added document is specified without changing the overall structure of the graph, and the difference from the previous state in time is clarified.

  In the above embodiment, the node arrangement position determining unit 18 calculates the node position independently for each cluster. However, as described below, the node location is determined according to the connection relationship with the nodes belonging to other clusters. The arrangement coordinates may be corrected. In the following, referring to FIG. 7 and FIG. 8, the nodes are rotated so that the total sum of the link lengths between the nodes is reduced by rotating at least a part of the clusters without changing the relative positional relationship between the clusters. A process for correcting the arrangement coordinates will be described.

  FIG. 7A shows a graph in which each cluster and node are arranged based on the node arrangement coordinates determined for each cluster by the node arrangement position determination unit 18 and the link relationship between the nodes. It is. Then, as shown in FIG. 7B, with respect to one cluster, the other clusters are rotated around the cluster center of the cluster. At this time, the rotation angle of the cluster is determined so as to minimize the total length of links between nodes. In this method, the relative relationship between clusters and the relative relationship between nodes in the cluster are maintained.

ただし、Ｌは、他クラスタからのリンクの総数である。従って、クラスタｃの受ける回転モーメントの総和Ｍ_ｃは、次式（６）で表される。

ただし、Ｎはクラスタｃに属するノード数である。クラスタｃの一回の更新における回転角度Δφ_ｃは、（７）式に比例した値として定め、以下の（７）式で表される。

ただし、αは微小な正数である。そして、（７）式で定義された角度による回転をすべてのクラスタ(１，…，Ｃ)に適用する。さらに、この更新を繰り返して適用し、次式（８）を満たした場合に処理を終了する。

ここで、Ｃはクラスタの総数であり、Ｔｈは予め定められた正の閾値である。 Next, the calculation process of the rotation angle of the cluster will be described with reference to FIG. FIG. 8 shows the cluster center of the cluster to be processed and the nodes i included in the cluster. The cluster center is calculated as the center of the feature vector of the nodes belonging to the cluster. It is assumed that the node i has a link L _ij (j = 1, 2,..., N) from a node belonging to another cluster. Here, the distance between the node i belonging to the cluster c and the cluster center is l _i , and the angle between the straight line passing through the node i and the cluster center and the link L _ij from another cluster is θ _ij (counterclockwise is positive) , −π ≦ θ _ij ≦ π). At this time, if the force received from the other cluster is constant, the rotational moment M _i generated by the node i by the link from the other cluster is expressed by the following equation (5).

However, L is the total number of links from other clusters. Accordingly, the sum M _{c of} rotational moments received by the cluster c is expressed by the following equation (6).

Here, N is the number of nodes belonging to the cluster c. The rotation angle Δφ _c in one update of the cluster c is determined as a value proportional to the equation (7) and is expressed by the following equation (7).

However, α is a minute positive number. Then, the rotation by the angle defined by the equation (7) is applied to all the clusters (1,..., C). Further, this update is repeatedly applied, and the process is terminated when the following expression (8) is satisfied.

Here, C is the total number of clusters, and Th is a predetermined positive threshold value.

FIG. 9 shows a flowchart of the rotation angle calculation process performed by the display information generation apparatus 10. As shown in FIG. 9, the display information generating apparatus 10 first initializes the relative position of the node set in the cluster (S201), and calculates the rotation angle Δφ _c for all clusters based on the above equation (7). (S202). Then, the display information generation device 10 rotates the cluster based on the calculated rotation angle (S203). Next, the display information generating apparatus 10 determines whether or not the arrangement of each node satisfies the termination condition based on the above equation (8) (S204). If it is determined that the end condition is not satisfied (S204: N), the process returns to S202 and the rotation process is performed. On the other hand, if it is determined that the end condition is satisfied (S204: Y), the arrangement coordinates of each node are determined (S205), and the process ends.

  In addition, this invention is not limited to said embodiment, For example, narrowing down based on the attribute information of a node may be performed, and the link relationship between nodes may be visualized. In other words, the links between nodes to which common attribute information is assigned may be narrowed down, and the link weight between clusters may be calculated based on the narrowed links between nodes, or conversely, common attribute information is given. The link weight between the nodes may be calculated by excluding the link between the nodes. For example, when a node is a patent gazette, citations with different applicants (other company citations) generally affect other companies than those with the same citation (in-house citation). In such cases, it is possible to exclude the company citations and extract and visualize the essential citation relationships between documents.

  Note that the attribute information given to the node may include a clustering result performed by another device, and the classification unit 14 may read the clustering result and classify each node.

It is a functional block diagram of a display information generation device. It is the figure which showed the relationship of the arrangement | positioning area | region of a cluster. It is the figure which showed the link relationship between clusters. It is the figure which showed the arrangement position of a node. It is the figure which showed an example of the graph display. It is a flowchart of the production | generation process of graph display data. It is a figure which shows an example of correction | amendment of the arrangement position of a node. It is a figure explaining the calculation process of the rotation angle of a cluster. It is a flowchart of the calculation process of a rotation angle.

Explanation of symbols

  DESCRIPTION OF SYMBOLS 10 Display information production | generation apparatus, 12 Network data memory | storage part, 14 Classification part, 16 Cluster arrangement | positioning area determination part, 18 Node arrangement position determination part, 20 Graph display data generation part, 22 Display part

Claims (5)

Means for acquiring information on a plurality of elements and connection relation information between the plurality of elements;
Means for classifying the plurality of elements into one or more element groups based on the information of the plurality of elements;
Means for setting a weight of connection relation between the element groups based on connection relation information between elements included in each element group;
An arrangement area determining means for determining an arrangement area for each of the classified element groups based on the set weight of the connection relation between the element groups;
Generating display information for displaying the elements arranged in the element group for each element group in the determined arrangement area and displaying the arranged elements connected based on the connection relation information Means,
A display information generating apparatus comprising: The display information generation apparatus according to claim 1, wherein the generation unit generates the display information to be displayed in a mode in which a boundary is indicated for each element group. The placement region determining means, the display information generation device according to the size of the placement area of each of the element groups to claim 1 or 2, wherein the determining based on the number of elements included in the element group. Said generating means according to any one of claims 1 to 3, characterized in that to determine the placement of the elements included in the element group for each of the element group so that the sum of the connection distance of each element is reduced Display information generation device. Means for acquiring information on a plurality of elements and connection relation information between the plurality of elements;
Means for classifying the plurality of elements into one or more element groups based on the information of the plurality of elements;
Means for setting a weight of connection relation between the element groups based on connection relation information between elements included in each element group;
An arrangement area determining means for determining an arrangement area for each of the classified element groups based on the set weight of the connection relation between the element groups;
Generating display information for displaying the elements arranged in the element group for each element group in the determined arrangement area and displaying the arranged elements connected based on the connection relation information A program for causing a computer to function as a means.

Images