Consensus building assistance device, consensus building assistance method, and consensus building assistance program
The consensus building support system addresses the lack of minority consideration in existing methods by using network graphs with weighted edges to promote fair and efficient agreement among participants, ensuring all voices are heard and committed to the decision-making process.
Patent Information
- Application Number
- PCT/JP2023/018606
- Authority / Receiving Office
- WO · WO
- Patent Type
- Applications
- Current Assignee / Owner
- Filing Date
- 2023-05-18
- Publication Date
- 2024-11-21
AI Technical Summary
Existing consensus building methods fail to consider minorities and fairness, leading to potential social issues and lack of rationality in decision-making processes, as they often rely on majority voting or leader-driven decisions without addressing the relationship between minorities and the majority.
A consensus building support system that utilizes a processor and storage device to acquire human relationship information and dissimilarity data, generating network graphs with weighted edges based on disagreement levels, and outputs graph features and promotion information to encourage agreement among participants, prioritizing fairness and minority inclusion.
The system supports fair and efficient consensus building by visualizing opinion distributions, facilitating understanding of minority opinions and central opinions, and promoting agreement through human relationships, ensuring commitment from all participants and maintaining stability in agreements.
Smart Images

Figure JP2023018606_21112024_PF_FP_ABST
Abstract
Description
Consensus building support device, consensus building support method, and consensus building support program
[0001] The present invention relates to a consensus building support device, a consensus building support method, and a consensus building support program that support consensus building.
[0002] In solving various social problems in countries, local governments, communities, etc., it is important to build consensus while respecting the diverse opinions of residents. Decision-making by majority vote or by a specific leader can force minorities or those under their control to agree to the decision, which can lead to new social problems. In order to aim for an equality-oriented democracy, it is necessary to build consensus among all people so that no one feels so strongly about disagreement that they refuse to agree.
[0003] Consensus building requires fairness in the process and results, efficiency in terms of time and cost, rationality based on scientific evidence and future predictions, and stability in the content of the agreement and the relationships between the parties.
[0004] Japanese Patent Laid-Open No. 2006-124493 discloses a system that supports consensus building among participants. In this system, participants' opinions are associated based on links of preset key phrases, and these opinions are expanded into a hierarchical network. The system then aggregates the participants' preferred key phrases and the number of times they selected the links, thereby presenting the preference trends of all participants as an overall opinion network.
[0005] Patent Document 2 below discloses a system for presenting options to participants. In this system, the similarity of the participants' preferences for multiple options is calculated, and options are recommended to the participants based on the preference similarity network between the participants or information on its matrix.
[0006] JP 2004-13256 A JP 2011-53920 A International Publication No. 2020 / 032125
[0007] Yasuhiro Asa, Takeshi Kato and Ryuji Mine, “Composite Consensus-Building Process: Permissible Meeting Analysis and Compromise Choice Exploration,”arXiv:2211.08593, 16 Nov. 2022.
[0008] Although Patent Document 1 provides an overview of the overall opinion network of participants, it is classified as a majority consensus because it is based on the participants' aggregated opinions. Therefore, consideration is not given to minorities or to fairness or rationality. Furthermore, while the overall opinion network contributes to the discussions among participants, the process leading to a final consensus is not disclosed.
[0009] In Patent Document 2, similarities are found from the preference similarity network of participants and options preferred by the participants are presented, but no consideration is given to how to reach a consensus from there, for example, which participants should be encouraged to reach a consensus. Furthermore, no consideration is given to the relationship between the minority and the majority.
[0010] The present invention aims to support fair consensus building while taking minority groups into consideration.
[0011] A consensus building support device that is one aspect of the invention disclosed in this application is a consensus building support device having a processor that executes a program and a storage device that stores the program, wherein the processor executes an acquisition process that acquires human relationship information that indicates the human relationships between a plurality of participants and dissimilarity information that indicates the degree of difference in opinions between the plurality of participants, and a first generation process that generates a network graph having nodes that indicate the participants, edges that indicate that there is a human relationship between two of the plurality of participants, and weights of the edges that indicate the dissimilarity, based on the human relationship information and dissimilarity information acquired by the acquisition process.
[0012] A representative embodiment of the present invention enables support for fair consensus building while taking minority groups into consideration. Other issues, configurations, and effects will become clear from the following description of the examples. Furthermore, the designations "first," "second," "third," etc., used in this specification are used to identify components and do not necessarily limit the number or order of components.
[0013] FIG. 1 is a block diagram illustrating an example of a configuration of a consensus building support system that supports consensus building according to a first embodiment. FIG. 2 is a flowchart illustrating an example of a consensus building support process procedure executed by the consensus building support device. FIG. 3 is a flowchart illustrating an example of an internal process procedure for content output processing in step S209 in FIG. 2. FIG. 4 is an explanatory diagram illustrating an example of data output in the consensus building support device. FIG. 5 is an explanatory diagram illustrating an example of data output in a personal device. FIG. 6 is an explanatory diagram illustrating an example of a human relationship information table. FIG. 7 is an explanatory diagram illustrating an example of first human relationship matrix data. FIG. 8 is an explanatory diagram illustrating an example of second human relationship matrix data. FIG. 9 is an explanatory diagram illustrating an example of an opinion network. FIG. 10 is an explanatory diagram illustrating an example of human relationship information. FIG. 11 is an explanatory diagram illustrating an example of opinion information. FIG. 12 is an explanatory diagram illustrating an example of dissimilarity information. FIG. 13 is an explanatory diagram illustrating an example of a first network graph of participants. FIG. 14 is an explanatory diagram illustrating an example of a second network graph. FIG. 15 is an explanatory diagram illustrating a first example of a third network graph. FIG. 16 is an explanatory diagram illustrating a second example of a third network graph. FIG. 17 is an explanatory diagram illustrating a third network graph example 3. FIG. 18 is an explanatory diagram illustrating a third network graph example 4. FIG. 19 is an explanatory diagram illustrating an update example of opinion information. FIG. 20 is an explanatory diagram illustrating an update example of dissimilarity information. FIG. 21 is an explanatory diagram illustrating a first second network graph example according to the second embodiment. FIG. 22 is an explanatory diagram illustrating a second network graph example 2 according to the second embodiment. FIG. 23 is an explanatory diagram illustrating a third second network graph example according to the second embodiment. FIG. 24 is an explanatory diagram illustrating a fourth second network graph example according to the second embodiment. FIG. 25 is an explanatory diagram illustrating a fifth second network graph example according to the second embodiment. FIG. 26 is an explanatory diagram illustrating an update example of human relationship information according to the third embodiment. FIG. 27 is an explanatory diagram illustrating an update example of dissimilarity information according to the third embodiment. FIG. 28 is an explanatory diagram illustrating a first second network graph example according to the third embodiment. FIG. 29 is an explanatory diagram illustrating a second network graph example 2 according to the third embodiment. FIG. 30 is an explanatory diagram illustrating a third second network graph example according to the third embodiment. FIG. 31 is a diagram illustrating a fourth example of the second network graph according to the third embodiment.FIG. 32 is an explanatory diagram illustrating a fifth example of a second network graph according to the third embodiment. FIG. 33 is an explanatory diagram illustrating an example of opinion information according to the fourth embodiment. FIG. 34 is an explanatory diagram illustrating an example of dissimilarity information according to the fourth embodiment. FIG. 35 is an explanatory diagram illustrating a first example of a second network graph according to the fourth embodiment. FIG. 36 is an explanatory diagram illustrating a first example of a third network graph according to the fourth embodiment. FIG. 37 is an explanatory diagram illustrating a second example of a third network graph according to the fourth embodiment. FIG. 38 is an explanatory diagram illustrating a third example of a third network graph according to the fourth embodiment. FIG. 39 is an explanatory diagram illustrating a first example of a third network graph according to the fifth embodiment. FIG. 40 is an explanatory diagram illustrating an example of human relationship information according to the fifth embodiment. FIG. 41 is an explanatory diagram illustrating a second example of a third network graph according to the fifth embodiment. FIG. 42 is an explanatory diagram illustrating a third example of a third network graph according to the fifth embodiment. FIG. 43 is an explanatory diagram illustrating a fourth example of a third network graph according to the fifth embodiment. FIG. 44 is an explanatory diagram illustrating a fifth example of a third network graph according to the fifth embodiment. FIG. 45 is an explanatory diagram illustrating an example of a network graph according to the sixth embodiment. FIG. 46 is a table illustrating an example of the calculation results of the graph feature amounts according to the first to sixth embodiments.
[0014] 1 is a block diagram showing a configuration example of a consensus building support system that supports consensus building according to Example 1. The consensus building support system 100 includes a consensus building support apparatus 110, a database 140, and personal devices 150 carried by participants, such as personal computers, tablet terminals, and smartphones. These are connected to each other via a network 130.
[0015] The consensus building support device 110 includes a processor 111, a memory 120, a storage 112, a user interface 113, and a network interface 114. The consensus building support device 110 is connected to a network 130 via the network interface 114 and communicates with a personal device 150 and a database 140.
[0016] The processor 111 executes calculation processing in the consensus building support device 110. The storage 112 is a non-transitory or temporary recording medium that stores various programs and data. Examples of the storage 112 include a read-only memory (ROM), a random access memory (RAM), a solid state drive (SSD), and a hard disk drive (HDD).
[0017] The storage 112 stores a human relationship information acquisition program 121 , an opinion information acquisition program 122 , an opinion difference calculation program 123 , a graph generation program 124 , a graph feature calculation program 125 , a graph output program 126 , and a content output program 127 .
[0018] The storage 112 also stores human relationship information and opinion information. The human relationship information is acquired from the personal device 150 via the network 130 by a human relationship information acquisition program 121. The opinion information is acquired by an opinion information acquisition program 122.
[0019] The user interface 113 includes, for example, at least one of a display, a keyboard, a mouse, a touch panel, a speaker, a microphone, a camera, and a sensor.
[0020] The network interface 114 is connected to a network 130 consisting of the Internet or a communication line, and transmits and receives data such as human relationship information, opinion information, and content to and from the personal device 150 .
[0021] The personal device 150 is a personal computer, a tablet terminal, a smartphone, or other smart device.
[0022] The consensus building support device 110 causes the processor 111 to execute a human relationship information acquisition program 121, an opinion information acquisition program 122, an opinion difference calculation program 123, a graph generation program 124, a graph feature calculation program 125, a graph output program 126, and a content output program 127 stored in the storage 112.
[0023] The human relationship information acquisition program 121 is a program that causes the processor 111 to execute a process of acquiring human relationship information of participants from the personal device 150 and storing the information in the storage 112 or the database 140. The human relationship information is information about whether a participant is connected to other participants, and includes call history and chat history between participants via the personal device 150, information declared by the participants themselves, relationship information based on a preliminary survey, etc.
[0024] The opinion information acquisition program 122 is a program that causes the processor 111 to execute a process of acquiring participant opinion information from the personal device 150 and storing the information in the storage 112 or the database 140. The opinion information is information that indicates the opinions of participants, such as the degree of approval or disapproval of a plurality of options or the order of preference of the participants.
[0025] The dissimilarity calculation program 123 is a program that causes the processor 111 to execute a process of calculating the degree of dissimilarity of opinions between participants based on opinion information. The degree of dissimilarity of opinions is the distance between the opinions of one participant and another participant, and is calculated based on the opinion information.
[0026] For example, if the opinion information is the degree of approval or disapproval of a plurality of options, the consensus building support device 110 calculates the degree of disagreement by adding up the difference in approval or disapproval of each option among the participants.
[0027] Furthermore, in the case of a preference order of multiple options, the consensus building support device 110 calculates the degree of disagreement based on the range of tolerance in the preference order of the participants, i.e., the range of preference order that the participants are willing to accept and believe to be acceptable. Furthermore, the consensus building support device 110 calculates the degree of disagreement based on the number of rearrangements made in the preference order of the options until the order agrees with the opinion of the other party, using the compromise option search described in Non-Patent Document 1.
[0028] The graph generation program 124 is a program that causes the processor 111 to execute processing for generating the first to third network graphs. Specifically, the consensus building support device 110 first generates a first network graph based on human relationship information. The first network graph is a network graph in which participants are nodes and human relationships are edges connecting the nodes, but the weights of the edges are all configured to be equal because the degree of disagreement is not taken into consideration.
[0029] Next, the consensus building support device 110 generates a second network graph by weighting the edges of the first network graph based on the degree of disagreement. In the second network graph, the degree of disagreement is used as the weight of the edges, so when the degree of disagreement between participants is large, the distance between the edges becomes long.
[0030] The consensus building support apparatus 110 generates a third network graph based on the graph feature of either the first network graph or the second network graph. The graph feature is calculated by the graph feature calculation program 125.
[0031] The graph feature calculation program 125 is a program that causes the processor 111 to execute a process of calculating the graph feature of the first network graph or the second network graph.
[0032] The graph features are modularity, closeness centrality, node betweenness centrality, edge betweenness centrality, average distance, graph density, link efficiency, clustering coefficient, and assortative selectivity, which are well known in network analysis.
[0033] Modularity is an index that measures the quality of network cluster division (community detection). Specifically, for each divided cluster, modularity is calculated by subtracting the square of the ratio of the total number of edges of the nodes included in the cluster to the total number of edges of the entire graph from the total number of edges in the cluster, and then summing up the values for all clusters. In terms of how the clusters are divided, the more edges there are within the cluster and the fewer connections there are with the outside of the cluster, the greater the modularity.
[0034] Closeness centrality is the inverse of the average value of the distance (shortest path length) from a node to all other nodes.
[0035] The node betweenness centrality is an index value that indicates the percentage of the node of interest that is included in the shortest path between two other nodes.
[0036] Edge betweenness is an index value obtained by applying node betweenness centrality to an edge, and is an index value that indicates the degree to which an edge exists on the shortest path between nodes.
[0037] The average distance is an index value obtained by adding up the shortest distances between all pairs of nodes and dividing the sum by the total number of such pairs.
[0038] The graph density is an index value obtained by dividing the number of edges in a network graph by the total number of edges when all nodes in the network graph are connected by edges.
[0039] Link efficiency is an index value that evaluates the strength of connections in a network graph. Specifically, for example, link efficiency is a value obtained by subtracting the value obtained by dividing the average distance by the total number of edges from 1.
[0040] The clustering coefficient is an index value that evaluates the density of a network graph. Specifically, the clustering coefficient is calculated by dividing the actual number of links between neighboring nodes of a given node by the number of possible links, and averaging the values for all nodes.
[0041] Assortative selectivity is the correlation coefficient of the degrees between connected nodes; a positive value indicates that high-degree nodes tend to be adjacent to other high-degree nodes, and a negative value indicates that high-degree nodes tend to be adjacent to low-degree nodes.
[0042] The graph output program 126 is a program that causes the processor 111 to execute a process of outputting the first network graph, the second network graph, and the third network graph via the user interface 113 or the personal device 150 .
[0043] The content output program 127 is a program that causes the processor 111 to execute a process for outputting content to the personal devices 150 of the participants. The content includes the graph features described above and promotional information that encourages participants to reach consensus and build relationships. The promotional information is automatically generated by a known artificial intelligence agent such as that described in Patent Literature 3, or is input via the user interface 113 by the facilitator's operation.
[0044] In this way, the consensus building support system 100 can visualize the distribution of opinions among participants based on the interpersonal relationships between participants and the degree of disagreement between opinions. Furthermore, by visualizing the distribution of opinions, minority opinions, central opinions, and the interpersonal relationships surrounding them can be easily grasped. This makes it possible to support fair consensus building while taking minority opinions into consideration.
[0045] <Consensus Building Support Processing> Next, a specific description will be given of the consensus building support processing executed by the consensus building support device 110. Here, a description will be given of processing for supporting consensus building based on the second network graph.
[0046] 2 is a flowchart showing an example of the consensus building support processing procedure executed by the consensus building support device 110. First, the consensus building support device 110 executes human relationship information acquisition processing using the human relationship information acquisition program 121 (step S201). Next, the consensus building support device 110 executes opinion information acquisition processing using the opinion information acquisition program 122 (step S202). Next, the consensus building support device 110 executes opinion difference calculation processing using the opinion difference calculation program 123 (step S203).
[0047] Next, the consensus building support device 110 uses the opinion difference calculation program 123 to determine whether the opinion difference between all participants is 0, i.e., whether consensus has been reached among all participants (step S204). A difference of 0 indicates that there is no difference in opinion among all participants, i.e., consensus has been reached among all participants. If the opinion difference is 0 (step S204: Yes), the consensus building support process ends.
[0048] On the other hand, if the degree of disagreement is not 0 (step S204: Yes), the process proceeds to step S205.
[0049] The consensus building support device 110 executes a graph generation process to generate a first network graph and a second network graph using the graph generation program 124 (step S205).
[0050] Next, the consensus building support apparatus 110 executes a graph feature calculation process using the graph feature calculation program 125 to calculate the graph feature in the first network graph or the second network graph (step S206).
[0051] The consensus building support device 110 executes a graph generation process using the graph generation program 124 to generate a third network graph based on the graph feature amount (step S207).
[0052] The consensus building support device 110 executes a graph output process using the graph output program 126 to output the first network graph, the second network graph, or the third network graph via the user interface 113 or the personal device 150 (step S208).
[0053] The third network graph includes, for example, communities detected by modularity-based clustering, nodes with high closeness centrality or node betweenness centrality, or edges with high edge betweenness centrality, and these nodes, edges, or communities are highlighted in the output of the third network graph. Note that a community is a group of nodes within a cluster.
[0054] The consensus building support device 110 executes a content output process using the content output program 127 to output content including graph features and promotion information via the user interface 113 or the personal device 150 (step S209). For example, the consensus building support device 110 outputs a line graph showing the time-series changes in the average distance, graph density, link efficiency, cluster coefficient, or similar selectivity values throughout the consensus building process.
[0055] When step S209 is executed, if the consensus building support device 110 outputs promotion information that promotes human relationships, the process returns to the human relationship information acquisition process (step S201), and if the consensus building support device 110 outputs promotion information that promotes consensus on opinions, the process returns to the opinion information acquisition process (step S202). In either case, the consensus building process from steps S201 to S209 is repeated until the degree of disagreement becomes 0, that is, until consensus is reached among all participants.
[0056] 3 is a flowchart showing an example of the internal processing procedure of the content output process (step S209) in FIG. 2. First, in step S301, nodes in the third network graph with high closeness centrality or node betweenness centrality are designated as hubs. Based on the calculation results of the graph feature calculation process (step S206), the consensus building support device 110 determines whether the average inter-hub distance in the second network graph is smaller than the average distance in the entire second network graph (step S301). The average inter-hub distance is an index value obtained by adding up the shortest distances between all pairs of hubs and dividing the sum by the total number of such pairs.
[0057] If the inter-hub average distance is smaller than the average distance of the entire second network graph (step S301: Yes), the process proceeds to step S302. If the inter-hub average distance is not smaller than the average distance of the entire second network graph (step S301: No), the process proceeds to step S303.
[0058] In step S302, the consensus building support device 110 executes an inter-hub agreement promotion information output process. That is, the consensus building support device 110 outputs promotion information to encourage agreement to participants corresponding to hubs with high closeness centrality or betweenness centrality.
[0059] Considering that the average distance between hubs is smaller than the overall average distance (step S301: Yes), it is easier to reach an agreement among hubs than among all participants, and it is believed that a hub with high centrality has a greater influence on the participants of the nodes around it. Specifically, the consensus building support device 110 outputs, via the user interface 113 or the personal device 150, promotion information to encourage discussion and compromise between participants corresponding to hubs with high closeness centrality or betweenness centrality, along with their opinions and the degree of difference between them. Then, the process returns to step S202.
[0060] Since discussions and compromises among participants take time, the consensus building support device 110 re-executes the opinion information acquisition process (step S202) after a predetermined time has elapsed. Alternatively, the consensus building support device 110 may re-execute the opinion information acquisition process (step S202) when receiving an instruction to re-execute the opinion information acquisition process (step S202) from the facilitator.
[0061] In step S303, the consensus building support device 110 determines, based on the calculation results of the graph feature calculation process (step S206), whether the average distance between communities, which are clusters calculated by modularity (average inter-community distance), is smaller than the average intra-community distance.
[0062] The inter-community average distance is an index value obtained by adding up the shortest distances between all pairs across communities and dividing the sum by the total number of such pairs. The intra-community average distance is an index value obtained by adding up the shortest distances between all pairs of nodes within a community, dividing the sum by the total number of such pairs, and then averaging the sum across all communities.
[0063] If the inter-community average distance is small (step S303: Yes), the process proceeds to step S304, and if the inter-community average distance is not small (step S303: No), the process proceeds to step S305.
[0064] In step S304, the consensus building support device 110 executes a process for outputting content for building fan relationships. That is, the consensus building support device 110 outputs promotion information for encouraging the building of fan relationships to participants who do not yet have human relationships but who share the same or similar opinions. Considering that the average distance between communities is smaller than the average distance within a community (step S303: Yes), the degree of difference of opinion within a community is large, and it is considered more advantageous to first build relationships among people with similar interests and then seek consensus among groups of people with similar interests, rather than seeking consensus among participants within the community.
[0065] Specifically, the consensus building support device 110 identifies pairs of participants who do not yet have a relationship with each other within the community and the opinions of each of the pairs of participants, and if the opinions are the same or similar, outputs to each personal device 150 of the pair of participants information about the other participant (unique ID or attribute data) as well as promotion information encouraging the start of an exchange of opinions. Then, the process returns to step S201.
[0066] Since it takes time for participants to exchange opinions, the consensus building support device 110 re-executes the human relationship information acquisition process (step S201) after a predetermined time has elapsed. Alternatively, the consensus building support device 110 may re-execute the human relationship information acquisition process (step S201) when receiving an instruction to re-execute the human relationship information acquisition process (step S201) from the facilitator.
[0067] In step S305, the consensus building support apparatus 110 executes a spoke conference formation promotion information output process. That is, the consensus building support apparatus 110 outputs promotion information to the personal devices 150 of all participants to promote the formation of relationships for forming a spoke conference.
[0068] A spoke conference is an equality-oriented meeting format in which representatives of communities, called spokes, reach consensus within their respective communities, and also reach consensus among themselves among themselves, aiming to reach consensus among all.
[0069] That is, in a spoke conference, relationships are built among all participants in the community and among all spokes. In this case, the average distance between hubs is greater than the overall average distance, which indicates a large degree of disagreement between hubs in the community (step S301: No).
[0070] Furthermore, considering a difference situation where the degree of difference of opinion within a community is small and the average distance within the community is small compared to the average distance between communities (step S303: No), it is considered desirable to form a spoke conference with the hub as the spoke.
[0071] Specifically, the consensus building support apparatus 110 outputs to the personal device 150 of the participant corresponding to the hub in each community (the node with the highest degree (number of edge connections) in the community) a message informing the participant that he or she will become a spoke, along with promotional information encouraging the participant to build relationships with all participants in the community and exchange opinions. The consensus building support apparatus 110 outputs to the personal devices 150 of the other participants promotional information encouraging the participant to build relationships with the spokes and all participants in the community and exchange opinions. After this, the process returns to step S201.
[0072] Since building relationships and exchanging opinions among participants takes time, the consensus building support device 110 re-executes the human relationship information acquisition process (step S201) after a predetermined time has elapsed. Alternatively, the consensus building support device 110 may re-execute the human relationship information acquisition process (step S201) when receiving an instruction to re-execute the human relationship information acquisition process (step S201) from the facilitator.
[0073] In this way, the consensus building support device 110 outputs the human relationships, the degree of disagreement, the first network graph, the second network graph, the third network graph, the graph features, and the promotion information, thereby enabling efficient facilitation and supporting the process of consensus building among all participants.
[0074] This process is not a consensus-building process based on majority vote, but rather on "human relationships" based on modularity, closeness centrality, betweenness centrality, and edge betweenness centrality. Furthermore, since all participants share the content, this process allows for consensus building that takes minorities into consideration. Furthermore, since this process requires everyone's commitment, it is believed that the implementation of the agreement after the fact will also be maintained.
[0075] Therefore, the consensus building support system 100 can support consensus building that is fair, efficient, and stable. Furthermore, by outputting multiple options when outputting content or acquiring opinion information, and at the same time outputting information based on scientific evidence and future predictions to participants, it enables rational consensus building.
[0076] In FIG. 2, steps S201 to S209 are executed sequentially, but for example, the order of steps S201 and S202 may be reversed, or steps S206 to S207, S208, and S209 may be executed in parallel.
[0077] It is also possible to transition from step S201 to steps S205 and S206, and have a user using the consensus building support device 110, such as a conference organizer or facilitator, analyze closeness centrality, node betweenness centrality, edge betweenness centrality, and community based solely on human relationships, identify important people involved in consensus building, and then return to step S202 to proceed with the consensus building process. Furthermore, steps S207 and S208 may be omitted, and only step S209 may be output.
[0078] Furthermore, in step S203, the consensus building support device 110 calculated the degree of difference in opinion from the degree of approval or disapproval, the acceptable range of preference order, and the number of changes. However, the degree of difference in opinion may also be obtained by a user using the consensus building support device 110, such as the organizer or facilitator of the meeting, directly asking participants about their subjective differences with other participants, and storing the results in the storage 112 or the database 140 via the user interface 113.
[0079] In addition, in step S205, the consensus building support device 110 generated a second network graph in which the degree of disagreement is used as the edge weight, but in order to take minority groups into greater consideration, edges or nodes may be weighted according to age, gender, occupation, position, income, or family structure, and edges may be changed from undirected to directed.
[0080] In addition, in step S206, the consensus building support device 110 used modularity, closeness centrality, betweenness centrality, edge betweenness centrality, average distance, graph density, link efficiency, cluster coefficient, and assortative selectivity as graph features, but it may also use graph diameter (maximum distance), graph radius (eccentricity), node connectivity, edge connectivity, reciprocity, correlation similarity, average degree, etc., and may output the graph in step S207 and the graph features in step S208 based on these.
[0081] Furthermore, in step S209, the consensus building support device 110 determined the content output using the average distance between hubs and the overall average distance, or the average distance between communities and the average distance within a community, but it is also possible to make the determination using other graph features described above, and the processing may be different from the processing shown in steps S301 to S305.
[0082] Examples of different operations include adding edges to other nodes for nodes with low closeness centrality, adding edges randomly throughout the network graph to increase graph density, adding edges between pairs of nodes with large distances to increase link efficiency, and adding edges between neighboring nodes of a node to increase the clustering coefficient.
[0083] In step S209, the consensus building support device 110 outputs content that encourages consensus and relationship building. However, the consensus building support device 110 may output new options as content that reduce the degree of disagreement based on at least one of the first network graph, the second network graph, and the third network graph and its graph features, and then return to step S202 to acquire opinion information again. For example, the storage 112 may store advantages and disadvantages as features corresponding to each option. The consensus building support device 100 may extract options preferred by nodes in the second network graph whose degree of disagreement is equal to or greater than a threshold, or options preferred by nodes in the third network graph whose closeness centrality or node betweenness centrality is equal to or greater than a threshold, read the advantages and disadvantages of the extracted options from the storage 112, and output them as content. Furthermore, the consensus building support device 100 may output options that have common advantages and exclude disadvantages based on the advantages and disadvantages of the extracted options as content to be edited. The content to be edited is edited by the facilitator's operation.
[0084] Furthermore, in the consensus building support system 100, data such as interpersonal relationship information, opinion information, and content is transmitted and received between the consensus building support apparatus 110 and the participants' personal devices 150 via the network 130, but it is also possible for information exchange and discussion to take place directly between participants or between a facilitator and participants without using the personal devices 150, or for information exchange via the personal devices 150 to be performed in combination with direct discussion. Interpersonal relationship information and opinion information based on the direct information exchange and discussion in the consensus building support apparatus 110 can be updated by participants again inputting the information via their personal devices 150 or by the facilitator inputting the information via the user interface 113.
[0085] 4 is an explanatory diagram showing an example of data output in the consensus building support device 110. A display 400 is connected to the user interface 113 of the consensus building support device 110, and outputs screens 401 to 403.
[0086] A network graph 411 is displayed on the screen 401 by the graph generation process (steps S205 and S207). The network graph 411 is at least one of the first to third network graphs. Buttons 412 are GUIs (Graphical User Interfaces) for selecting a network graph based on the first, second, or third modularity, closeness centrality, betweenness centrality, or edge betweenness centrality.
[0087] In the network graph 411, numeric numbers indicating the unique IDs of participants are displayed at the nodes, but if there are no problems from the viewpoint of protecting personal information, names or nicknames of participants may be displayed instead of the unique IDs. For example, attribute information such as age and gender may be displayed in a pop-up.
[0088] On the screen 402, a line graph 421 showing the time series changes of the graph feature amount is displayed as a result of step S208. A button 422 is a GUI for selecting the average distance, graph density, link efficiency, cluster coefficient, or similar selectivity from the graph feature amounts.
[0089] Screen 403 displays the contents of operations such as multiple options, discussions among participants regarding those options, chat conversations between participants and between participants and the facilitator, and the generation and input of content that encourages participants to reach an agreement or build relationships through tolerance or compromise, which is used by the content output program 127 in step S209. The output of these contents is selected from the menu on the left side of screen 403.
[0090] According to Figure 4, the facilitator can view the network graph 411, the line graph 421, and the content on screen 403, which includes promotional information that encourages discussions, conversations, agreements, and relationship building among participants, as a dashboard, allowing for even more efficient facilitation.
[0091] 5 is an explanatory diagram showing an example of data output in the personal device 150. The personal device 150 is an example of the personal device 150, and is connected to the consensus building support apparatus 110 via the network interface 114 and the network 130, and content is displayed by content output processing (step S209).
[0092] A network graph 411 is displayed on the screen 501 of the personal device 150 by the graph output process (step S208). In the network graph 411, the unique IDs, names, or nicknames of the participants are displayed at the nodes, and the participants' own nodes are highlighted. Buttons 503 are used to select the network graph 411 based on first, second, or third modularity, closeness centrality, betweenness centrality, or edge betweenness centrality from a pull-down menu.
[0093] The screen 504 displays a preference order 540, which is composed of an inequality between multiple options and a selection order symbol, as the opinion of the participant with the personal device 150. a, b, c, . . . are the options on the screen 505.
[0094] The selection order symbols include the greater than symbol ">", the greater than or equal to symbol ">", the approximation symbol for greater than or equal to, and the approximation symbol.
[0095] The greater than symbol ">" is a symbol that indicates that the option on the right side is preferred over the option on the left side. The greater than or equal symbol ">" is a symbol that indicates that the option on the right side is preferred over or equal to the option on the left side. The greater than or equal approximation symbol is a symbol that indicates that the option on the right side is preferred over or roughly equal to the option on the left side. The approximation symbol is a symbol that indicates that the option on the left side and the option on the right side are roughly equal.
[0096] If a participant selects a greater-than-or-equal approximation symbol or an approximation symbol, the left and right sides are determined to be in an indiscriminate relationship, i.e., both sides are in an allowable relationship. If a participant changes their opinion toward a consensus, they can change the options or preference order symbols using buttons, and the changes are sent from the personal device 150 to the consensus building support device 110.
[0097] Screen 505 displays the contents of the multiple options a, b, c, ..., discussions among participants regarding the multiple options a, b, c, ..., chat conversations between participants and between participants and the facilitator, and promotion information resulting from the content output process (step S209). Selection of the display of these contents is performed from the menu on the left side of screen 505. Input regarding discussions and conversations, building relationships with other participants, and declarations regarding human relationships including acquaintances is also performed from screen 505.
[0098] Screen 506 is an example of a screen on which a participant declares interpersonal relationship information. When an acquaintance is selected from the menu on screen 505, the personal device 150 displays screen 506. From screen 506, the participant can enter the name or nickname of the acquaintance, or can press a button to enter a name from a list of names. This input information is sent to the consensus building support device 110 and stored in the storage 112 or database 140 as interpersonal relationship information.
[0099] According to Figure 5, participants can confirm their own position in the first network graph, the second network graph, or the third network graph, reconsider the order of their preferences for multiple options in response to prompting information that encourages agreement, and form relationships with other participants in response to prompting information that encourages relationship building, thereby promoting the process toward total agreement.
[0100] Note that this effect does not depend solely on the configuration of screens 501, 504, and 505 and button 503 shown in Fig. 5, and other configurations may also be used. Screen 504 prompts participants to select a preference order for multiple options and displays the preference order as is, but it is also possible to prompt participants to input the degree of approval or disapproval for each option and display, for example, a bar graph. Also, while Fig. 5 uses personal device 150, personal device 150 may also be a personal computer, tablet terminal, or other smart device.
[0101] <Database> Next, information stored in the storage 112 or the database 140 will be described.
[0102] 6 is an explanatory diagram showing an example of a human relationship information table. The human relationship information table has human relationship information. The human relationship information is information that incorporates human relationship and opinion history information, human relationship information, and opinions.
[0103] The human relationship information table 600 is history information of human relationships and opinions stored in the storage 112 or the database 140 in steps S201 and S202. Specifically, for example, the human relationship information table 600 includes fields such as a first ID 601, a first opinion 602, a second ID 603, a second opinion 604, the number of calls 605, the number of chats 606, whether a report has been made 607, and whether a relationship exists 608.
[0104] The first ID 601 is identification information that uniquely identifies the first participant.
[0105] The first opinion 602 is history information indicating the opinion of the first participant, for example, the first opinion 602 is the preference order 540 of the first participant regarding the content discussed with the second participant.
[0106] The second ID 603 is identification information that uniquely identifies the second participant with whom a personal relationship with the first participant is to be built.
[0107] The second opinion 604 is history information indicating the opinion of the second participant, for example, the second opinion 604 is the preference order 540 of the second participant regarding the content discussed with the first participant.
[0108] The second opinion 604 is character string data indicating a message that the personal device 150 of the second participant has sent to the personal device 150 of the first participant using a message application.
[0109] The number of calls 605 is the number of calls between the first participant and the second participant between the personal device 150 of the first participant and the personal device 150 of the second participant. The number of calls 605 is counted, for example, from call history data acquired by the consensus building support apparatus 110 from either the personal device 150 of the first participant or the personal device 150 of the second participant.
[0110] The number of chats 606 is the number of times the first participant and the second participant chatted (the number of times messages were sent) using a message application between the personal device 150 of the first participant and the personal device 150 of the second participant. The number of chats 606 is counted, for example, from message transmission history data acquired by the consensus building support apparatus 110 from the personal device 150 of the first participant and the personal device 150 of the second participant.
[0111] The declaration presence / absence 607 is a flag indicating whether the first participant has declared that they have a personal relationship with the second participant. If it is "1", it indicates that the first participant has declared that they have a personal relationship with the second participant, and if it is "0", it indicates that the first participant has not declared that they have a personal relationship with the second participant.
[0112] For example, the personal device 150 of a participant transmits, for example, input from the screen 505, declaration data including the unique ID of the other party with whom a human relationship has been established to the consensus building support apparatus 110. The consensus building support apparatus 110 identifies an entry in the human relationship information table 600 in which the unique ID of the participant of the personal device 150 that is the sender is the first ID 601 and the unique ID of the other party is the second ID 602, and if the declaration presence / absence 607 is "0", updates it to "1".
[0113] Although the declaration data includes the unique ID of the other party, it may also include attribute data of the other party. In this case, the consensus building support device 110 may read the unique ID of the other party using the attribute data of the other party as a key and set it as the second ID 602.
[0114] Relationship status 608 is a flag indicating whether or not there is a relationship in the preliminary survey. If it is "1", it indicates that the first participant has a relationship with the second participant as a result of the preliminary survey, and if it is "0", it indicates that the first participant has a relationship with the second participant.
[0115] The preliminary survey is a process in which the consensus building support apparatus 110 inquires in advance about the presence or absence of a human relationship with the personal device 150 and acquires a response result from the personal device 150. The response result includes a combination of the unique ID of the participant of the personal device 150 that is the inquiry destination and the unique ID of the other party with whom the participant has a human relationship.
[0116] The consensus building support device 110 identifies an entry from the human relationship information table 600 in which the unique ID of the participant of the personal device 150 being inquired about is the first ID 601 and the unique ID of the other party is the second ID 602, and if the relationship existence 608 is "0", it updates it to "1".
[0117] Although the answer result includes the unique ID of the other party, it may also include attribute data of the other party. In this case, the consensus building support device 110 may read the unique ID of the other party using the attribute data of the other party as a key and set it as the second ID 602.
[0118] 7 is an explanatory diagram showing an example of first human relationship matrix data. The first human relationship matrix data 700 is data extracted from the human relationship information table 600 and stored in the storage 112 or the database 140. The first human relationship matrix data 700 indicates whether or not a human relationship exists, each consisting of a first ID 601 and a second ID 602. If the value of the matrix element specified by the first ID 601 and the second ID 602 is "1," this indicates that a human relationship exists between the first participant and the second participant, and if the value is "0," this indicates that no human relationship exists between the first participant and the second participant.
[0119] Specifically, for example, if at least one of the report / non-report 607 and the relationship / non-report 608 is "1", the value of the matrix element identified by the first ID 601 and the second ID 602 is set to "1", and if both the report / non-report 607 and the relationship / non-report 608 are "0", the value of the matrix element identified by the first ID 601 and the second ID 602 is set to "0".
[0120] Note that the diagonal elements where the first ID 601 and the second ID 602 are the same are left blank, in order to avoid drawing self-loop edges when generating the first network graph in step S205.
[0121] Using such first human relationship matrix data 700, the consensus building support device 110 can identify whether or not there is a human relationship between the first participant and the second participant.
[0122] 8 is an explanatory diagram showing an example of the second human relationship matrix data 800. The second human relationship matrix data 800 is data that incorporates the degree of disagreement extracted from the human relationship information table 600 and stored in the storage 112 or the database 140.
[0123] The second human relationship matrix data 800 indicates the degree of disagreement between a first participant and a second participant, each of which is composed of a first ID 601 and a second ID 602. The larger the value of the matrix element identified by the first ID 601 and the second ID 602, the greater the difference in opinion between the first and second participants, with "1" indicating a difference in opinion between the first and second participants. "0" indicating no difference in opinion between the first and second participants. "∞" indicates no disagreement between the first and second participants, i.e., at least one of the first opinion 602 and the second opinion 604 does not exist.
[0124] Note that the diagonal elements where the first ID 601 and the second ID 602 are the same are left blank, in order to avoid drawing self-loop edges when generating the second network graph in step S205.
[0125] The degree of disagreement is the difference between a first opinion 602 of a first participant identified by a first ID 601 and a second opinion 604 of a second participant identified by a second ID 603. Specifically, for example, the degree of disagreement is data that quantifies the difference between the preference order 540 identified by the first opinion 602 and the preference order 540 identified by the second opinion 604. An example of calculating the degree of disagreement will be described later.
[0126] Attribute data (not shown) of the participants, such as their names, nicknames, ages, and genders, are associated with their unique IDs and stored in the storage 112 or the database 140, and can be read out as needed. The unique IDs of the participants are applied to the first ID 601 and the second ID 603.
[0127] The opinions of the participants acquired in step S202, i.e., the preference order 540, are associated with symbols such as opinions α, β, γ, etc. and stored in the storage 112 or the database 140, and can be read out as needed. The degree of difference in opinions is found by referring to the argument table based on the opinion symbols of the two participants.
[0128] In this way, the consensus building support device 110 can efficiently execute the human relationship information acquisition process (step S201), the opinion information acquisition process (step S202), and the opinion difference calculation process (step S203), and can then proceed to the generation of the first network graph and the second network graph in the graph generation process (step S205).
[0129] 6 shows an example of the human relationship information table 600, but the contents and order of the columns and the order of the fields can be changed as needed. While the first human relationship matrix data 700 and the second human relationship matrix data 800 are shown in the adjacency matrix format in Figures 7 and 8, a list format or a table format can also be used.
[0130] <Degree of Difference of Opinion> Next, the degree of difference of opinion will be explained.
[0131] 9 is an explanatory diagram showing an example of an opinion network. Here, simply, the opinion network 900 shows the connections of six opinions α (a>b>c), β (b>a>c), γ (b>c>a), δ (c>b>a), ε (c>a>b), and ζ (a>c>b) for three options a, b, and c.
[0132] The opinion network 900 is a network in which opinions are nodes and line segments connecting preference orders 540 are edges. As shown in the opinion network 900, six opinions α, β, γ, δ, ε, and ζ are related to each other based on the number of times two adjacent options in the preference order 540 are swapped. Specifically, if two of the three options a, b, and c in the preference order 540 are swapped once, the resulting preference order 540 is adjacent and connected by an edge.
[0133] For example, the preference order 540 indicated by opinion α is "a>b>c," and the preference order 540 indicated by opinion β is "b>a>c." If the adjacent "a" and "b" in the preference order 540 "a>b>c" indicated by opinion α are swapped, the preference order 540 indicated by opinion β becomes "b>a>c."
[0134] Therefore, if opinion α's options are swapped twice, it is converted into opinion γ, and if opinion α's options are swapped twice, it is converted into opinion δ. The number of swaps is the degree of disagreement between the opinions. Note that there are two directions from one opinion to the destination opinion: clockwise and counterclockwise, and the lesser of the number of swaps in the clockwise direction and the number of swaps in the counterclockwise direction is used as the degree of disagreement between the opinions. In other words, the shortest number of hops from one node to another node is the degree of disagreement between the opinions.
[0135] For example, the number of times opinions α and γ are swapped is "2" in the clockwise direction and "4" in the counterclockwise direction, but the number of times "2" with fewer swaps is adopted as the degree of dissimilarity between opinions α and γ. Therefore, in the example of Figure 9, the degree of dissimilarity between opinions is an integer between 0 and 3.
[0136] If the dissimilarity between two participants is 0, it means that the opinions of both participants are the same, and if the dissimilarity is equal to or less than a threshold value, it means that the opinions of both participants are similar. The threshold value is set in advance, but in the example of Figure 9, since the maximum value of the dissimilarity of opinion is "3", the threshold value may be set to "1", which is an integer equal to or less than half of that.
[0137] FIG. 10 is an explanatory diagram showing an example of human relationship information. Human relationship information 1000 shows the human relationships of 20 participants. The human relationship information 1000 is a collection of human relationship data for first IDs 601 and second IDs 603. The human relationship data is made up of the first ID 601 on the left end, the second ID 603 on the right end, and a line segment connecting the first ID 601 and the second ID 603. In the example of FIG. 10, the human relationship information 1000 includes 37 sets of human relationship data.
[0138] In the following explanation, the human relationship between x and y in FIG. 10 (x is the first ID 601, and y is the second ID 603) will be referred to as human relationship h(x-y). When x and y are not specified, it will be simply referred to as human relationship h. Similarly, the degree of difference of opinion between participants specified by human relationship h(x-y) will be referred to as d(x-y). When x and y are not specified, it will be simply referred to as difference of opinion d.
[0139] 11 is an explanatory diagram showing an example of opinion information. The opinion information 1100 is a collection of opinions for each participant. The leftmost opinion is the participant's unique ID: 1, and the opinions (one of the six opinions α, β, γ, δ, ε, and ζ shown in FIG. 6 ) are arranged in ascending order of unique ID up to the rightmost unique ID: 20.
[0140] Fig. 12 is an explanatory diagram showing an example of dissimilarity information. The dissimilarity information 1200 shows the degree of dissimilarity between opinions in the human relationship information 1000 of Fig. 10. For example, in the case of human relationship h(1-2), referring to the opinion information 1100, the first opinion 602 of the first participant whose first ID 601 is "1" is "α", and the second opinion 604 of the second participant whose second ID 603 is "2" is "β".
[0141] The degree of difference between opinion α and opinion β is "1" when referring to the opinion network 900. Therefore, the degree of difference d(1-2) of opinion in human relationship h(1-2) is "1".
[0142] Similarly, in the case of human relationship h(4-9), when the opinion information 1100 is referred to, the first opinion 602 of the first participant whose first ID 601 is "4" is "ε", and the second opinion 604 of the second participant whose second ID 603 is "9" is "β".
[0143] The degree of difference between opinion ε and opinion β is "1" when referring to the opinion network 900. Therefore, the degree of difference d(4-9) of opinion in human relationship h(4-9) is "3".
[0144] 10 to 12 show the human relationship information 1000, opinion information 1100, and dissimilarity information 1200 in list format, but they may also be in table format or adjacency matrix format as shown in FIGS.
[0145] <Graph Generation Process (Step S205)> Next, a specific example of the graph generation process (Step S205) will be described.
[0146] 13 is an explanatory diagram showing an example of a first network graph of participants. The first network graph 1300 is a graph in which the unique IDs of 20 participants are nodes and 37 human relationships are edges, based on the human relationship information 1000 in FIG. 10. Hereinafter, a node whose unique ID is # (# is a number) will be referred to as node #.
[0147] In addition, in FIG. 13, the first network graph 1300 is drawn using a layout method that determines a stable arrangement of nodes by regarding the nodes as charges and the edges as springs.
[0148] Although the weights of all edges are "1", the lengths of the edges change due to the layout method described above, and the placement of nodes changes depending on the relationships, i.e., how the edges are connected. From Figure 13, we can see that the participants at nodes 2, 3, and 4 have many relationships, i.e., they are hubs of relationships.
[0149] The first network graph 1300 has an average distance of 2.508, a graph density of 0.195, a link efficiency of 0.944, a clustering coefficient of 0.258, and an assortative selectivity of −0.459 (see table 4600 in FIG. 46).
[0150] Fig. 14 is an explanatory diagram showing an example of a second network graph. The second network graph 1400 is a graph in which the unique IDs of 20 participants are nodes, 37 human relationships are edges, and the weights of the edges represent the degrees of disagreement, based on the human relationship information 1000 in Fig. 10 and the dissimilarity information 1200 in Fig. 12.
[0151] 13, in Fig. 14, it can be inferred that nodes 3 and 4 are relatively close, meaning that it is easy to reach an agreement, and nodes 17 and 19 are relatively far, meaning that it is difficult to reach an agreement. Note that the average distance of the second network graph 1400 is 3.721, the graph density is 0.195, the link efficiency is 0.944, the clustering coefficient is 0.258, and the assortative selectivity is -0.459 (see table 4600 in Fig. 46).
[0152] Compared to the first network graph 1300, the second network graph 1400 uses the degree of disagreement as the edge weight. As a result, the second network graph 1400 has a larger average distance and smaller assortative selectivity than the first network graph 1300. Incidentally, since the graph density, link efficiency, and clustering coefficient are calculated based on the number of edges, the values do not change unless the graph structure changes.
[0153] 15 is an explanatory diagram showing Example 1 of the third network graph. The third network graph 1500 is a network graph based on the modularity (clustering) of graph features. When generating the third network graph 1500, the consensus building support device 110 calculates the modularity based on the human relationship information 1000 in FIG. 10 and the dissimilarity information 1200 in FIG. 12. Clustering is performed such that many edges connect to nodes within a community and a small number of edges connect to nodes in other communities.
[0154] In FIG. 15, it can be seen that three communities 1501 to 1503 have been formed, with node 3, node 4, and node 7 as hubs, respectively.
[0155] 16 is an explanatory diagram showing a second example of a third network graph. The third network graph 1600 is a network graph based on the closeness centrality of graph features. When generating the third network graph 1600, the consensus building support device 110 calculates the closeness centrality, which indicates that a node has a short average distance from other nodes, based on the human relationship information 1000 in FIG. 10 and the dissimilarity information 1200 in FIG. 12. The nodes are displayed larger depending on the magnitude of the closeness centrality.
[0156] In FIG. 16, it can be seen that nodes 3 and 4 have particularly high closeness centrality, that is, these participants are at the center of the consensus of the other participants.
[0157] 17 is an explanatory diagram showing a third network graph example 3. The third network graph 1700 is a network graph based on the betweenness centrality of graph features. When generating the third network graph 1700, the consensus building support device 110 calculates the betweenness centrality, which indicates that a certain node is on the shortest path between many other node pairs, based on the human relationship information 1000 in FIG. 10 and the dissimilarity information 1200 in FIG. 12. The node is displayed larger depending on the magnitude of the betweenness centrality.
[0158] In FIG. 17, it can be seen that the betweenness centrality of node 2, node 3, and node 4 is large, that is, they are the centers that connect the relationships between other participants.
[0159] Figure 18 is an explanatory diagram showing Example 4 of the third network graph. The third network graph 1800 is a network graph based on the edge betweenness centrality of the graph feature. When generating the third network graph 1800, the consensus building support device 110 calculates the edge betweenness centrality, which indicates the number of shortest paths between a node pair that pass through a certain edge, based on the human relationship information 1000 in Figure 10 and the opinion dissimilarity information 1200 in Figure 12. Edges with a high edge betweenness centrality are displayed as thick lines, and edges with a low edge betweenness are displayed as dotted lines.
[0160] In Figure 18, it can be seen that the edge betweenness centrality between node 3 and node 4 and the edge betweenness centrality between node 3 and node 16 are large, meaning that these relationships are central to the relationships between other participants.
[0161] 13 to 18, nodes 3, 4, and 7 can be considered to be hubs. The opinions of nodes 3, 4, and 7 are δ, ε, and ζ, respectively, and the average inter-hub distance is 1.333. Because the average inter-hub distance is smaller than the average distance of 3.721 in the second network graph 1400, the consensus building support device 110 can output promotion information to each participant at nodes 3, 4, and 7 to encourage consensus, in accordance with step S302 in FIG.
[0162] 13 to 18, a network graph layout method in which nodes are regarded as electric charges and edges as springs is used, but this is not limited to this, and other layout methods such as a circular layout method, a polygonal layout method, a lattice layout method, a stacked layout method, a linear layout method, or a random layout method may be used depending on the number of participants, their relationships, the degree of disagreement, etc.
[0163] Next, Example 2 will be described. Example 2 shows an example of generating a network graph when the hub of the participants is proactively prompted to agree in step S302 in Example 1. Note that, in Example 2, differences from Example 1 will be mainly described, and therefore, parts common to Example 1 will be assigned the same reference numerals and description thereof will be omitted.
[0164] 19 is an explanatory diagram showing an example of updating opinion information. When the inter-hub agreement promotion information output process (step S302) is executed, the consensus building support apparatus 110 outputs, for example, to the personal device 150, promotion information that encourages agreement to participants corresponding to hubs with high closeness centrality or betweenness centrality. Considering that the average distance between hubs is smaller than the overall average distance (step S301: Yes), agreement is more easily reached between hubs than among all participants, and it is thought that hubs with high centrality have a greater influence on participants in the nodes around them.
[0165] As a result, it is assumed that, prior to re-execution of the opinion information acquisition process (step S202), for example, a participant (node 3) having opinion δ and a participant (node 7) having opinion ζ agree with the opinion ε of the participant (node 4).
[0166] In this case, the personal device 150 of the participant (node 3) transmits a change notification from opinion δ to opinion ε to the consensus building support apparatus 110. Similarly, the personal device 150 of the participant (node 7) transmits a change notification from opinion ζ to opinion ε to the consensus building support apparatus 110.
[0167] Upon receiving the change notification, the consensus building support device 110 updates the opinion information 1100 to reflect the contents of the change notification. Specifically, for example, in FIG. 19 , a participant (node 3) having opinion δ and a participant (node 7) having opinion ζ agree with the opinion ε of a participant (node 4) in the opinion information 1100, so the consensus building support device 110 updates the opinion δ of node 3 and the opinion ζ of node 7 to the opinion ε of node 4. The updated opinion information 1100 is referred to as opinion information 1900.
[0168] 20 is an explanatory diagram showing an example of updating the dissimilarity information 1200. The consensus building support apparatus 110 updates the dissimilarity information 1200 in accordance with the update of the opinion information 1900 shown in FIG. The updated dissimilarity information 1200 is referred to as dissimilarity information 2000. In the dissimilarity information 1200, the dissimilarity d(3-4) of the human relationship h(3-4) is "1", and the dissimilarity d(3-7) of the human relationship h(3-7) is "2".
[0169] In the opinion information 1900, the opinion δ of node 3 has been converted to the opinion ε of node 4, so the dissimilarity d(3-4) of the human relationship h(3-4) is converted to "0." Similarly, in the opinion information 1900, the opinion ζ of node 7 has been converted to the opinion ε of node 4, so the dissimilarity d(3-7) of the human relationship h(3-7) is also converted to "0."
[0170] 21 is a diagram illustrating a first example of a second network graph according to the second embodiment. A second network graph 2100 is a network graph in which the degree of disagreement between participants is used as the weight of an edge.
[0171] The second network graph 2100 is a graph in which the human relationship information 1000 in Fig. 10 is used as the edges and the converted dissimilarity information 2000 in Fig. 20 is used as the edge weights. In the second network graph 2100, a loop-shaped edge indicates that the degree of disagreement is "0", and the hubs Node 3, Node 4, and Node 7 overlap at the same position (only "4" is shown in Fig. 21).
[0172] Also, for example, because nodes 3 and 4 overlap, the two edges from node 5 to nodes 3 and 4 are drawn curved so as not to overlap each other. In Figure 21, the opinions of the participants identified by the hubs nodes 3, 4, and 7 have agreed at ε, suggesting that nodes 8, 9, 10, 19, and 20 are now farther away from the hubs (nodes 3, 4, and 7), that is, consensus building is needed between the participants at nodes 8, 9, 10, 19, and 20 and the participants at the hubs (nodes 3, 4, and 7).
[0173] The second network graph 2100 has an average distance of 2.916, a graph density of 0.195, a link efficiency of 0.944, a clustering coefficient of 0.258, and a similarity selectivity of -0.462 (see table 4600 in FIG. 46). Compared to the network graph 1400, the second network graph 2100 has a smaller average distance and a larger similarity selectivity. The former is because the participants at the hubs (nodes 3, 4, and 7) agreed, and the latter is because the participants at the hubs (nodes 3, 4, and 7) agreed, which increased the similarity of other participants from the perspective of the participants at the hubs (nodes 3, 4, and 7).
[0174] 22 is an explanatory diagram illustrating a second network graph example 2 according to the second embodiment. The second network graph 2200 is a network graph based on the modularity (clustering) of graph features. In the second network graph 2200, it can be seen that a community 2201 with node 3, node 4, and node 7 as hubs, a community 2202 with node 9 as hub, and a community 2203 with node 10 as hub have been formed.
[0175] 23 is an explanatory diagram illustrating Example 3 of the second network graph according to Example 2. The second network graph 2300 is a network graph based on the closeness centrality of the graph feature. In the second network graph 2300, it can be seen that the closeness centrality of nodes 11, 13, 17, and 18 is large after the hubs, nodes 3, 4, and 7.
[0176] 24 is an explanatory diagram illustrating Example 4 of the second network graph according to Example 2. The second network graph 2400 is a network graph based on the betweenness centrality of the graph feature. In the second network graph 2400, it can be seen that the betweenness centrality of the hub nodes 3, 4, and 7 is significantly higher than that of the other nodes.
[0177] 25 is an explanatory diagram illustrating a fifth example of a second network graph according to the second embodiment. The second network graph 2500 is a network graph based on the edge betweenness centrality of the graph feature. In the second network graph 2500, it can be seen that the edge betweenness centrality between the hubs Node 3, Node 4, and Node 7 (in FIG. 25, Nodes 3 and 7 overlap with Node 4) and Node 17 or Node 18 is large.
[0178] 21 to 25, it can be seen that it is necessary to proceed with consensus building mainly between node 9 or node 17, which is the hub of the new community, and the hubs of node 3, node 4, and node 7. Similarly, it can be seen that it is necessary to proceed with consensus building between node 10 or node 19, which is the hub of the new community, and the hubs of node 3, node 4, and node 7. It can also be seen that it is necessary to proceed with consensus building between node 18, which is in the same community as the hubs of node 3, node 4, and node 7, and the hubs of node 3, node 4, and node 7.
[0179] Therefore, the consensus building support apparatus 110 outputs prompting information to encourage consensus to the personal devices 150 of the participants identified by the nodes that are the target of consensus building in accordance with step S209. Then, the process returns to step S202. Note that, for simplicity, in the second embodiment, a case where there are only three options is shown, but by presenting an intermediate option between options a, b, and c as prompting information, it is possible to move closer to consensus, i.e., reduce the degree of disagreement.
[0180] Next, Example 3 will be described. Example 3 shows an example of generating a network graph when step S304 in Example 1 encourages participants with similar opinions to form relationships. Note that, since Example 3 will mainly describe differences from Example 1, parts in common with Example 1 will be assigned the same reference numerals and descriptions thereof will be omitted.
[0181] 26 is an explanatory diagram showing an example of updating the human relationship information 1000 according to Example 3. When the hobbyist relationship building promotion information output process (step S304) is executed, the consensus building support apparatus 110 outputs, for example, promotion information to the personal device 150, which promotes the building of hobbyist relationships among participants who do not yet have human relationships but who share the same or similar opinions. Considering that the average distance between communities is smaller than the average distance within a community (step S303: Yes), the degree of difference of opinion within a community is large, and it is considered more advantageous to first build relationships among hobbyists and then seek consensus among groups of hobbyists, rather than seeking consensus among participants within the community.
[0182] As a result, it is assumed that a new human relationship is established, for example, by creating a hobbyist relationship, prior to re-execution of the human relationship information acquisition process (step S201). In this case, the participant's personal device 150 transmits declaration data including the unique ID of the other party with whom the human relationship has been established to the consensus building support apparatus 110. The consensus building support apparatus 110 identifies an entry in the human relationship information table 600 in which the unique ID of the participant of the personal device 150 that sent the declaration data is the first ID 601 and the unique ID of the other party is the second ID 602, and if the declaration presence / absence 607 is "0," it updates it to "1."
[0183] The declaration data includes the unique ID of the other party, but may also include attribute data of the other party. In this case, the consensus building support device 110 reads the unique ID of the other party using the attribute data of the other party as a key and sets the ID as the second ID 602.
[0184] Furthermore, prior to re-executing the human relationship information acquisition process (step S201), the consensus building support apparatus 110 sends an inquiry about the presence or absence of a human relationship to the personal device 150 of the participant, and acquires a response result from the personal device 150 of the participant. The response result includes a combination of the unique ID of the participant of the personal device 150 that is the inquiry destination and the unique ID of the other party with whom the participant has a human relationship.
[0185] The consensus building support device 110 identifies an entry from the human relationship information table 600 in which the unique ID of the participant of the personal device 150 being inquired about is the first ID 601 and the unique ID of the other party is the second ID 602, and if the relationship existence 608 is "0", it updates it to "1".
[0186] Although the answer result includes the unique ID of the other party, it may also include attribute data of the other party. In this case, the consensus building support device 110 may read the unique ID of the other party using the attribute data of the other party as a key and set it as the second ID 602.
[0187] In this way, the human relationship information 1000 is updated to human relationship information 2600. In the example of Fig. 26, the human relationship information 2600 is information in which the human relationship information 1000 has been updated to include the following new relationships: h(1-6), h(1-8), h(1-12), h(1-14), h(1-19), h(2-9), h(2-15), h(3-13), h(3-17), h(5-10), h(5-16), h(5-18), h(7-11), and h(7-20).
[0188] 27 is an explanatory diagram illustrating an example of updating the dissimilarity information 1200 according to the third embodiment. The consensus building support apparatus 110 updates the dissimilarity information 1200 based on the updated human relationship information 2600 and the opinion information 1100 shown in FIG. 11. In the example of FIG. 27, the dissimilarity information 2700 is information in which the dissimilarity information 1200 is newly added with the following dissimilarity values: d(1-6), d(1-8), d(1-12), d(1-14), d(1-19), d(2-9), d(2-15), d(3-13), d(3-17), d(5-10), d(5-16), d(5-18), d(7-11), and d(7-20).
[0189] In this example, the opinions of like-minded people are consistent based on the opinion information 1100, so the degrees of difference d(1-6), d(1-8), d(1-12), d(1-14), d(1-19), d(2-9), d(2-15), d(3-13), d(3-17), d(5-10), d(5-16), d(5-18), d(7-11), and d(7-20) are all "0".
[0190] 28 is an explanatory diagram illustrating a first example of a second network graph according to the third embodiment. A second network graph 2800 is a network graph in which the degree of difference d between the opinions of the participants is the weight of the edge.
[0191] The second network graph 2800 is a graph in which the human relationship information 1000 in Fig. 10 is the edge and the dissimilarity information 2700 in Fig. 27 is the edge weight. A loop-shaped edge indicates that the dissimilarity d between like-minded people is "0".
[0192] Node 1 (opinion α) overlaps with nodes 6, 8, 12, 14, and 19. Node 2 (opinion β) overlaps with nodes 9 and 15. Node 3 (opinion δ) overlaps with nodes 13 and 17.
[0193] Node 5 (opinion γ) overlaps with nodes 16 and 18. Node 7 (opinion ζ) overlaps with nodes 11 and 20. Node 4 (opinion ε) is a single node that does not overlap with any other nodes.
[0194] The degree of difference of opinion d(1-2), d(3-4), and d(3-5) between node 1 and node 2, node 3 and node 4, and node 3 and node 5 is "1", and the degree of difference of opinion d(1-5), d(3-7) between node 1 and node 5, and node 3 and node 7 is "2".
[0195] The second network graph 2800 has an average distance of 1.542, a graph density of 0.268, a link efficiency of 0.963, a clustering coefficient of 0.326, and an assortment selectivity of -0.495 (see table 4600 in FIG. 46). Compared to the network graph 1400, the second network graph 2800 has a smaller average distance and a larger graph density, link efficiency, and clustering coefficient. This is because the number of like-minded relationships, i.e., edges, has increased.
[0196] 29 is an explanatory diagram illustrating a second example of a second network graph according to a third embodiment. The second network graph 2900 is a network graph based on the modularity (clustering) of graph features. In the second network graph 2900, it can be seen that the community is divided into two: a community 2901 consisting of participants at nodes 1, 4, and 5, and a community 2902 consisting of participants at nodes 2, 3, and 7.
[0197] When these are associated with the opinion network 900 in FIG. 9, they are the community of opinions α, γ, and ε, and the community of opinions β, δ, and ζ, whose degree of difference d between them is "2".
[0198] 30 is an explanatory diagram illustrating Example 3 of the second network graph according to Example 3. The second network graph 3000 is a network graph based on the closeness centrality of the graph feature. In the second network graph 3000, it can be seen that the closeness centrality of node 2 is the largest.
[0199] 31 is an explanatory diagram illustrating a fourth example of a second network graph according to a third embodiment. The second network graph 3100 is a network graph based on the betweenness centrality of graph features. In the second network graph 3100, it can be seen that nodes 1, 2, and 3 have a higher betweenness centrality than the other three nodes, nodes 4, 5, and 7.
[0200] 32 is an explanatory diagram illustrating Example 5 of a second network graph according to Example 3. The second network graph 3200 is based on the edge betweenness centrality of the graph feature. In the second network graph 3200, it can be seen that the edge betweenness centrality between node 2 and node 5 is particularly large.
[0201] 28 to 32, it can be seen that communities are formed based on the degree of disagreement between opinions, with each opinion becoming the center of mediation for interpersonal relationships. In Example 3, in order to provide a simple example, only six opinions indicating preference order are present, and these are aggregated into six nodes, dividing the community into two. Furthermore, according to Example 3, as described in the explanation of S209 in FIG. 3, by presenting intermediate opinions or options as promotion information, it is possible to promote the overlap of communities, i.e., the creation of relationships toward consensus.
[0202] In Example 2, which prioritizes agreement between hubs, the second network graph 2100 is radial, while in Example 5, which focuses on building relationships among like-minded people, the second network graph 2800 is spindle-shaped. The second network graph 2100 is suitable when the influence of the hub toward an agreement is relatively large, while the second network graph 2800 is suitable when the initial human relationships are sparse and it is necessary to build relationships toward an agreement.
[0203] Next, Example 4 will be described. Example 4 shows an example of generating a network graph when consensus is reached among all participants in Examples 1 to 3. Note that, in Example 4, differences from Example 1 will be mainly described, and therefore, parts common to Examples 1 to 3 will be assigned the same reference numerals and descriptions thereof will be omitted.
[0204] 33 is a diagram illustrating an example of opinion information according to Example 4. In opinion information 3300, the opinions of all participants with unique IDs 1 to 20 are η.
[0205] 34 is an explanatory diagram illustrating an example of dissimilarity information according to Example 4. Since all opinions in the opinion information 3300 are η, in the dissimilarity information 3400, the dissimilarity d of opinions in the human relationship information 1000 in FIG.
[0206] 35 is an explanatory diagram showing example 1 of a second network graph according to Example 4. A second network graph 3500 is a graph in which the human relationship information 1000 in FIG. 10 is used as an edge and the dissimilarity information 3400 in FIG. 34 is used as an edge weight. The loop-shaped edge indicates that the dissimilarity d between the opinions of all participants is 0, and all nodes other than node 1 overlap node 1.
[0207] In the second network graph 3500, all nodes are equivalent, so no community division by clustering is performed, and there is only one community. Note that the average distance of the second network graph 3500 is 0, the graph density is 0.195, the link efficiency is 0.944, the clustering coefficient is 0.258, and the assortative selectivity is 0 (see table 4600 in FIG. 46 ). Because the graph density, link efficiency, and clustering coefficient are calculated based on the number of edges, the values are the same as those of the network graph 1400 even if the edge weight is 0.
[0208] 36 is a diagram illustrating a first example of a third network graph according to Example 4. The third network graph 3600 is a network graph based on the closeness centrality of graph features.
[0209] 37 is an explanatory diagram illustrating a second example of a third network graph according to a fourth embodiment. The third network graph 3700 is a network graph based on the betweenness centrality of the graph feature. In the third network graphs 3600 and 3700, all nodes overlap at a single center, and therefore nodes with the same closeness centrality and betweenness centrality are gathered at that center.
[0210] 38 is an explanatory diagram illustrating a third example of a network graph according to a fourth embodiment. The third network graph 3800 is a network graph based on the edge betweenness centrality of the graph feature. In the third network graph 3800, the weights of all edges are 0, so the edges themselves no longer exist and are not drawn.
[0211] Example 4 shows, for example, starting from the first network graph 1300, second network graph 1400, and third network graphs 1500, 1600, 1700, and 1800 shown in Example 1, and going through the agreements and relationship building shown in Example 2 or Example 3 to finally reach the second network graph 3500 and third network graphs 3600, 3700, and 3800.
[0212] 2, this shows the case where the degree of disagreement d becomes 0 in step S204, and the process ends. In other words, consensus building involves degenerating the second network graph and the third network graph and overlapping the nodes of the participants.
[0213] Next, a fifth embodiment will be described. The fifth embodiment shows an example of generating a network graph for a spoke conference that takes into account human relationships. In the fifth embodiment, differences from the first embodiment will be mainly described, and therefore common parts to the first embodiment will be denoted by the same reference numerals and will not be described again.
[0214] 39 is an explanatory diagram illustrating a first example of a third network graph according to a fifth embodiment. The third network graph 3900 is a network graph based on the modularity (clustering) of graph features. The consensus building support device 110 calculates the modularity of the first network graph 1300 in which the human relationship information 1000 in FIG. 10 is used as edges and the edge weights are all set to 1 without using the degree of disagreement, and detects communities, thereby generating the third network graph 3900.
[0215] It can be seen that the third network graph 3900 has four communities 3901 to 3904 with node 2, node 3, node 4, and node 9 as hubs, respectively.
[0216] 40 is an explanatory diagram showing an example of human relationship information according to Example 5. Human relationship information 4000 shows human relationships when a spoke conference is formed based on communities 3901 to 3904, which are clusters in the third network graph 3900. The human relationship information 4000 shows the interrelationships of all participants in the communities 3901 to 3904 detected in the third network graph 3900, and the interrelationships of all hubs that serve as spokes.
[0217] The third network graph 4100 based on the human relationship information 4000 has an average distance of 2.189, a graph density of 0.279, a link efficiency of 0.959, a clustering coefficient of 0.812, and an assortative selectivity of 0.188 (see table 4600 in FIG. 46).
[0218] Although the average distance is larger than that of the first network graph 1300, the graph density and link efficiency are larger, and in particular, the clustering coefficient and assortative selectivity are significantly larger. This is because a fully connected network is formed within the community and between the spokes, and the equality of the participant nodes is improved.
[0219] 41 is an explanatory diagram illustrating a second example of a third network graph according to a fifth embodiment. The third network graph 4100 is a network graph based on the modularity (clustering) of graph features. In the third network graph 4100, the nodes in four communities 4101 to 4104 form a fully connected network, and node 2, node 3, node 4, and node 9 form the spokes of the fully connected network.
[0220] 42 is an explanatory diagram illustrating a third example of a network graph according to a fifth embodiment. The third network graph 4200 is a network graph based on the closeness centrality of graph features. In the third network graph 4100, it can be seen that the closeness centrality of the spokes, Node 2, Node 3, Node 4, and Node 9, is large.
[0221] 43 is an explanatory diagram illustrating Example 4 of the third network graph according to Example 5. The third network graph 4300 is a network graph based on the betweenness centrality of the graph feature. In the third network graph 4300, it can be seen that the spokes with a larger number of participants in the communities 4301 to 4304 have a larger betweenness centrality.
[0222] 44 is an explanatory diagram illustrating Example 5 of a third network graph according to Example 5. The third network graph 4400 is a network graph based on the edge betweenness centrality of the graph feature. In the third network graph 4400, it can be seen that the edge betweenness centrality between node 2 and node 3, which are spokes with a large number of participants in the community, and the edge betweenness centrality between node 2 and node 4 are particularly large, followed by the edge betweenness centrality between node 3 and node 4.
[0223] According to the fifth embodiment, a spoke conference is first formed, and then the consensus building process as shown in the first embodiment is carried out. This is suitable when priority is to be given to the interpersonal community. In the example of FIG. 2 , the consensus building support device 110 acquires interpersonal information in step S201, proceeds to step S205 to generate a first network graph, detects a community in step S206, proceeds to step S209 to output promotion information for forming a spoke conference based on the community, and then returns to step S201 to confirm the interpersonal information of the spoke conference and then executes the consensus building process from there.
[0224] First, by structuring a spoke meeting with consideration for human relationships, it becomes easier to build relationships through the output of promotional information, and consensus building can proceed more smoothly than if the spoke meeting were structured mechanically without consideration for human relationships.
[0225] Example 6 will be described. Example 6 shows an example of stepwise execution of generation of a network graph related to the spoke conference shown in Example 5 in the consensus building support system 100. Note that Example 6 will mainly describe differences from Example 5, and therefore common parts with Example 5 will be assigned the same reference numerals and description thereof will be omitted.
[0226] 45 is an explanatory diagram showing an example of a network graph related to a spoke conference according to Example 6. A network graph 4500 related to a spoke conference is generated by hierarchically operating the consensus building support system 100. The network graph 4500 related to a spoke conference is configured as a fully connected network including the insides of communities 4501 to 4507 and a spoke group 4508 that represents the communities.
[0227] By operating the consensus building support system 100 hierarchically within the communities 4501 to 4507 and the spoke group 4508, it is possible to promote equal and fair consensus building.
[0228] In the sixth embodiment, an example was shown in which the consensus building support system 100 is composed of communities 4501 to 4507 and spoke group 4508. While communities 4501 to 4507 are formed around spoke group 4508, each of communities 4501 to 4507 may be treated as a spoke group, and further communities may be formed around each of communities 4501 to 4507. In this way, by connecting further communities outside the spoke group, a three-tiered spoke conference can be configured, and similarly, when a large number of people are involved, even more tiered spoke conferences can be configured.
[0229] However, when using multiple layers, the advantages of equality-oriented spoke meetings are lost, and consideration must be given to increasing the number of meetings within and between layers, making each other's status public, etc., so as not to impair fairness and efficiency.Fair and efficient spoke meetings can be configured by utilizing the call, chat, and avatar functions of communication tools in social networking services and the metaverse.
[0230] As an example of an application of the consensus building support system 100, it is expected to support consensus building between local governments and their residents or businesses when considering policies in the local government. Also, when financial institutions such as banks consider investments and loans using the environment, society, and governance as one of their decision-making criteria, the consensus building support device 110 is expected to support consensus building between the investment recipient, or in some cases, the target of the business, and the financial institution.
[0231] The present invention is not limited to the above-described embodiments, and includes various modifications and equivalent configurations within the spirit and scope of the appended claims. For example, the above-described embodiments have been described in detail to clearly explain the present invention, and the present invention is not necessarily limited to configurations including all of the described configurations. Furthermore, part of the configuration of one embodiment may be replaced with the configuration of another embodiment. Furthermore, the configuration of another embodiment may be added to the configuration of one embodiment. Furthermore, part of the configuration of each embodiment may be added to, deleted from, or replaced with other configurations.
[0232] Furthermore, the aforementioned configurations, functions, processing units, processing means, etc. may be realized in part or in whole in hardware, for example by designing them as integrated circuits, or may be realized in software by having a processor interpret and execute a program that realizes each function.
[0233] Information such as programs, tables, and files that realize each function can be stored in a storage device such as a memory, hard disk, or SSD (Solid State Drive), or in a recording medium such as an IC (Integrated Circuit) card, SD card, or DVD (Digital Versatile Disc).
[0234] In addition, the control lines and information lines shown are those that are considered necessary for explanation, and do not necessarily represent all the control lines and information lines that are necessary for implementation. In reality, it can be assumed that almost all components are interconnected.
Claims
1. A consensus building support device having a processor that executes a program and a storage device that stores the program, wherein the processor executes an acquisition process to acquire human relationship information indicating the human relationships between a plurality of participants and dissimilarity information indicating the degree of difference in opinions between the plurality of participants, and a first generation process to generate a network graph having nodes indicating the participants, edges indicating that there is a human relationship between two of the plurality of participants, and weights of the edges indicating the dissimilarity, based on the human relationship information and dissimilarity information acquired by the acquisition process.
2. A consensus building support device as described in claim 1, characterized in that, in the acquisition process, the processor acquires opinion information indicating the opinion of each participant, and calculates the degree of difference in opinion between the two participants based on the opinion information.
3. A consensus building support device as described in claim 2, characterized in that, in the acquisition process, the processor calculates the degree of difference in opinions between the two participants based on the degree of agreement of opinions between the two participants.
4. A consensus building support device as described in claim 3, wherein the opinions are data indicating preference relationships between a plurality of options, and in the acquisition process, the processor calculates the number of times that the opinions of one of the two participants are reached by swapping the options to reach the opinion of the other participant as the degree of difference in opinions between the two participants.
5. A consensus building support device as described in claim 1, characterized in that the processor executes a determination process to determine whether or not there is a difference of opinion for all of the multiple participants based on the dissimilarity information, and in the first generation process, the processor generates the network graph if there is a difference of opinion for all of the multiple participants.
6. A consensus building support device as described in claim 1, characterized in that the processor executes a calculation process to calculate graph features of the network graph, and a second generation process to generate a network graph related to the graph features based on the network graph and the graph features calculated by the calculation process.
7. A consensus building support device as described in claim 1, characterized in that the processor executes an output process of outputting promotion information that promotes consensus on the opinions or the building of human relationships.
8. A consensus building support device as described in claim 7, characterized in that the processor executes a calculation process to calculate graph features of the network graph, and in the output process, the processor outputs the promotion information based on the graph features calculated by the calculation process.
9. A consensus building support device as described in claim 8, characterized in that, in the output processing, the processor outputs promotion information that promotes consensus on the opinions when the average distance between hub nodes among the group of nodes in the network graph is smaller than the average distance between nodes in the network graph.
10. A consensus building support device as described in claim 8, characterized in that in the output processing, the processor outputs promotion information that promotes the building of human relationships when the average distance between hub nodes among the group of nodes in the network graph is not smaller than the average distance between nodes in the network graph.
11. A consensus building support device as described in claim 5, wherein the processor executes an output process to output promotion information that promotes consensus on the opinions or building of human relationships, and re-executes the acquisition process, the judgment process, the first generation process, and the output process until the judgment process determines that there are no differences of opinion for all of the multiple participants.
12. A consensus building support device as described in claim 11, characterized in that the processor re-acquires the dissimilarity information in the acquisition process when promotion information for promoting consensus on the opinions is output by the output process.
13. A consensus building support device as described in claim 11, characterized in that the processor, when promotion information for promoting human relationship building is output by the output process, re-acquires the human relationship information and the dissimilarity information in the acquisition process.
14. A consensus building support method executed by a consensus building support apparatus having a processor that executes a program and a storage device that stores the program, wherein the processor executes an acquisition process to acquire human relationship information indicating the human relationships between a plurality of participants and dissimilarity information indicating the degree of difference in opinions between the plurality of participants, and a first generation process to generate a network graph having nodes indicating the participants, edges indicating that there is a human relationship between two of the plurality of participants, and weights of the edges indicating the dissimilarity, based on the human relationship information and dissimilarity information acquired by the acquisition process.
15. A consensus building support program characterized by having a processor execute: an acquisition process for acquiring human relationship information indicating the human relationships between a plurality of participants and dissimilarity information indicating the degree of difference in opinions between the plurality of participants; and a first generation process for generating a network graph having nodes indicating the participants, edges indicating that there is a human relationship between two of the plurality of participants, and weights of the edges indicating the dissimilarity, based on the human relationship information and dissimilarity information acquired by the acquisition process.
Citation Information
Patent Citations
Method and apparatus for supporting consolidation of ballot target
JP2011053920A
Discussion support device, and program for discussion support device
WO2020032125A1
Consensus-building support system
JP2004013256A
Adheat advertising model for social networks
JP2012519335A
Consensus building support device and consensus building support method
JP2021170187A
