Information processing system and information processing method

Abstract

To provide information processing systems and methods for appropriately evaluating effective control relationships between entities. [Solution] The information processing system includes an entity network acquisition unit that acquires an entity network representing the mutual investment relationships of multiple nodes corresponding to multiple entities, and a controlled node identification unit that identifies controlled nodes, which are nodes effectively controlled by a dominant node among the multiple nodes, based on the entity network. The controlled node identification unit identifies controlled nodes by performing an update process that adds to the controlled nodes one or more nodes in a second node group, which are directly connected to the first node group, and whose sum of investment ratios assigned to the edges between each node in the second node group is greater than a first threshold, when the dominant node and the already identified controlled nodes are considered a first node group.

Description

Technical Field

[0001] The present invention relates to an information processing system, an information processing method, and the like.

Background Art

[0002] Conventionally, a method of quantifying and evaluating the influence of one entity on another entity using the shareholding ratio (equity ratio, capital contribution ratio) has been known. Entities include, for example, countries, companies, people, and the like.

[0003] In Non-Patent Document 1, when one entity indirectly influences another entity, a method of simply accumulating the shareholding ratios to quantify the above influence has been proposed.

[0004] In addition, in Patent Document 1, when the capital contribution ratio exceeds half, a method of rewriting the capital contribution ratio to 100% and then quantifying the influence between entities has been proposed.

Prior Art Documents

Patent Documents

[0005]

Patent Document 1

Non-Patent Documents

[0006]

Non-Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0007] In conventional methods, depending on the structure of the network being analyzed, there may be a discrepancy between the effective dominance relationships between entities and the calculated influence values.

[0008] According to some aspects of this disclosure, it is possible to provide an information processing system and information processing method, etc., for appropriately evaluating effective control relationships between entities. [Means for solving the problem]

[0009] One aspect of this disclosure relates to an information processing system that identifies a controlled node by performing an update process to add to the controlled node a node that is directly connected to a second node group, which is one or more nodes included in the first node group, and whose sum of the investment ratios assigned to the edges between each node of the second node group is greater than a first threshold, when the first node group is defined as the first node group, the controlled node identification unit, which includes an entity network acquisition unit that acquires an entity network representing the mutual investment relationships of a plurality of nodes corresponding to a plurality of entities by connecting the lower nodes of the invested side and the upper nodes of the invested side with an edge assigned to an edge.

[0010] Other aspects of this disclosure relate to an information processing method that identifies a controlled node by performing an update process to add to the controlled node a node that is directly connected to a second node group, which is one or more nodes included in the first node group, and whose sum of the investment ratios assigned to the edges between each node of the second node group is greater than a first threshold, in which case the controlled node and the already identified controlled node are considered as a first node group. [Brief explanation of the drawing]

[0011] [Figure 1] This is an example of a system configuration that includes an information processing system. [Figure 2] This is an example of a server system configuration. [Figure 3] This is an example of a terminal device configuration. [Figure 4] This is an explanatory diagram of corporate holding network analysis. [Figure 5] This is an example of an entity network (subnetwork). [Figure 6A] This is an explanatory diagram of a comparative example to this embodiment. [Figure 6B] This is an explanatory diagram of a comparative example to this embodiment. [Figure 7] This is a specific example of a network where the impact is calculated to be excessively low in the comparative example. [Figure 8] This is a flowchart illustrating the processing of the information processing system in this embodiment. [Figure 9] This is a flowchart explaining the process of identifying controlled nodes. [Figure 10]It is a diagram for explaining the propagation of dominance by the specific process of the dominated node. [Figure 11] It is a diagram for explaining the propagation of dominance by the specific process of the dominated node. [Figure 12] It is a diagram for explaining the propagation of dominance by the specific process of the dominated node. [Figure 13] It is a diagram for explaining the propagation of dominance by the specific process of the dominated node. [Figure 14] It is a diagram for explaining the propagation of dominance by the specific process of the dominated node. [Figure 15A] Another example of an entity network (sub-network). [Figure 15B] An example of an adjacency matrix. [Figure 16] It is a flowchart for explaining the process using matrix operations. [Figure 17A] It is a diagram for explaining the first step of the vector update process. [Figure 17B] It is a diagram exemplifying the processing result of the first step. [Figure 18A] It is a diagram for explaining the second step of the vector update process. [Figure 18B] It is a diagram exemplifying the processing result of the second step. [Figure 19A] It is a diagram for explaining the third step of the vector update process. [Figure 19B] It is a diagram exemplifying the processing result of the third step. [Figure 20A] It is a diagram for explaining the fourth step of the vector update process. [Figure 20B] [[ID=4�]]It is a diagram exemplifying the processing result of the fourth step. [Figure 21A] It is a diagram for explaining the calculation process of the influence degree. [Figure 21B] It is a diagram exemplifying the result of the influence degree calculation process. [Figure 22A] It is a diagram for explaining the first step of the vector update process. [Figure 22B] It is a diagram exemplifying the processing result of the first step. [Figure 23A]This diagram illustrates the second step of the vector update process. [Figure 23B] This diagram illustrates the processing result of the second step. [Figure 24A] This diagram explains the process for calculating the degree of impact. [Figure 24B] This diagram illustrates the results of the impact calculation process. [Figure 25] This figure illustrates the processing results according to this embodiment. [Figure 26] This figure illustrates the processing results according to this embodiment. [Modes for carrying out the invention]

[0012] This embodiment will be described below with reference to the drawings. In the drawings, the same or equivalent elements are denoted by the same reference numerals, and redundant descriptions are omitted. This embodiment described below is not intended to unduly limit the content described in the claims. Furthermore, not all of the configurations described in this embodiment are essential components of this disclosure.

[0013] 1. OSINT System 1.1 Example System Configuration Figure 1 shows an example of the configuration of a system including the information processing system 10 according to this embodiment. The system according to this embodiment includes a server system 100 and terminal devices 200. However, the configuration of the system including the information processing system 10 is not limited to the example in Figure 1, and various modifications can be made, such as omitting some parts or adding other configurations. For example, Figure 1 shows two terminal devices 200, terminal device 200-1 and terminal device 200-2, as examples, but the number of terminal devices 200 is not limited to these. Furthermore, the same applies to Figures 2 and 3, which will be described later, in which modifications such as omitting or adding configurations can be made.

[0014] The information processing system 10 in this embodiment corresponds to, for example, a server system 100. The server system 100 corresponds to a computer. However, the method of this embodiment is not limited to this, and the processing of the information processing system 10 may be performed by distributed processing using the server system 100 and other devices. For example, the information processing system 10 in this embodiment may be realized by distributed processing between the server system 100 and a terminal device 200. An example where the information processing system 10 is the server system 100 will be described below.

[0015] The server system 100 may consist of one server or may include multiple servers. For example, the server system 100 may include a database server and an application server. The database server may store various types of data, such as the entity network 121 described later. The application server may perform various processes according to this embodiment. The multiple servers described above may be physical servers or virtual servers. If virtual servers are used, the virtual servers may be located on a single physical server or distributed across multiple physical servers. As described above, the specific configuration of the server system 100 in this embodiment can be modified in various ways.

[0016] Terminal device 200 is a device used by users of the information processing system 10. Terminal device 200 may be a PC (Personal Computer), a mobile terminal device such as a smartphone, or another device.

[0017] The server system 100 is connected to terminal devices 200-1 and 200-2 via a network, for example. Hereafter, if there is no need to distinguish between multiple terminal devices, they will simply be referred to as terminal device 200. The network will be described as a public communication network such as the Internet, for example, but it may also be a LAN (Local Area Network), etc.

[0018] The information processing system 10 of this embodiment is an OSINT (Open Source Intelligence) system that collects and analyzes data on a subject using, for example, publicly available information. Publicly available information includes a wide range of information that is widely available and legally obtainable, such as securities reports, input-output tables, official government announcements, and news reports concerning countries and companies. Note that the information processing system 10 of this embodiment is not limited to an OSINT system.

[0019] The server system 100 generates nodes containing various attributes based on publicly available information. Each node represents an entity. An entity may be a person, a company, or a country. Attributes are, for example, information determined based on publicly available information. Attributes include information about the entity and information about its shareholding ratio (investment ratio). Attributes may also include various other information such as nationality, business field, sales, number of employees, board members, and traded items.

[0020] If a given node has an attribute that includes a relationship with another node, the given node and the other node are connected by an edge that has direction. For example, suppose a given entity's shareholders include another entity. In this case, the node corresponding to the other entity and the node corresponding to the given entity are connected by an edge that represents the shareholding ratio. The edge has a direction from the influencing side to the affected side. For example, the edge has a direction from the investing side to the invested side. However, the edge may also have a direction from the affected side to the giving side.

[0021] In the method of this embodiment, the server system 100 obtains an entity network, which is a network in which multiple nodes representing multiple entities are connected by edges having an orientation based on attributes. That is, the entity network is a directed graph. The server system 100 performs analysis based on the entity network and processes the results of the analysis. For example, the terminal device 200 is a device used by a user who uses the services provided by the OSINT system. For example, the user requests some kind of analysis from the server system 100 (information processing system 10) using the terminal device 200. The server system 100 performs analysis based on the entity network and sends the analysis results to the terminal device 200 as a response.

[0022] Figure 2 is a block diagram showing a detailed configuration example of the server system 100. The server system 100 includes, for example, a processing unit 110, a storage unit 120, and a communication unit 130.

[0023] The processing unit 110 of this embodiment can be implemented by predetermined hardware. The hardware may include at least one of a circuit that processes digital signals and a circuit that processes analog signals. For example, the hardware may consist of one or more circuit devices or one or more circuit elements mounted on a circuit board. As circuit devices, ICs (Integrated Circuits) and FPGAs (Field-Programmable Gate Arrays) may be used. One or more circuit elements may be, for example, resistors, capacitors, etc.

[0024] The processing unit 110 may be implemented by one or more processors. The server system 100 of this embodiment includes, for example, a memory for storing information and a processor that operates based on the information stored in the memory. The information is, for example, a program and various types of data. The processor includes hardware. Various types of processors can be applied, such as a CPU (Central Processing Unit), a GPU (Graphics Processing Unit), or a DSP (Digital Signal Processor). The memory may be a semiconductor memory such as SRAM (Static Random Access Memory), DRAM (Dynamic Random Access Memory), or flash memory, or it may be a register. The memory may be a magnetic storage device such as a hard disk drive (HDD), or an optical storage device such as an optical disk drive. For example, the memory stores instructions that can be read by a computer, and the functions of the processing unit 110 are realized as processing when the processor executes these instructions. The instructions may be instructions from an instruction set that constitutes a program, or they may be instructions that instruct the hardware circuit of the processor to operate.

[0025] The processing unit 110 in the example in Figure 2 includes an entity network acquisition unit 111, a target node acquisition unit 112, a subnetwork extraction unit 113, a controlled node identification unit 114, an influence calculation unit 115, and a display processing unit 116.

[0026] The entity network acquisition unit 111 is an acquisition unit that acquires the entity network 121. For example, the entity network acquisition unit 111 may generate the entity network 121 based on publicly available information. The entity network acquisition unit 111 stores the generated entity network 121 in the storage unit 120. When executing the processing of this embodiment, including the process of identifying controlled nodes, the entity network acquisition unit 111 performs the process of reading (acquiring) the entity network 121 stored in the storage unit 120.

[0027] The generation of the entity network 121 may be performed in a system other than the information processing system 10 according to this embodiment. In this case, the entity network acquisition unit 111 may acquire the entity network from the other system via the communication unit 130.

[0028] The entity network acquisition unit 111 acquires, for example, an entity network 121 in which multiple entities are connected by investment relationships. The entity network 121 contains multiple entities. Each entity represents a node, as described above. Nodes are connected by edges based on investment relationships. Each edge is also assigned an investment ratio. The investment ratio represents the shareholding ratio. Shareholding ratio information can also be obtained based on the publicly available information described above.

[0029] The target node acquisition unit 112 performs the process of acquiring target nodes from multiple nodes included in the entity network 121. Here, a target node refers to a node whose influence from other nodes is to be determined. For example, the target node acquisition unit 112 may acquire target nodes based on the result of a target node selection operation performed on the terminal device 200.

[0030] The subnetwork extraction unit 113 performs a process to extract subnetworks consisting of higher-level nodes that are directly or indirectly connected to the target node when one of the multiple nodes of the entity network 121 is determined to be the target node.

[0031] However, extracting subnetworks related to the target node is not mandatory. For example, subnetworks related to the dominating node (the dominating node described below) may be extracted instead of the target node, which is the node being dominated. Alternatively, subnetwork extraction may be omitted altogether, and the following processing may be performed on the entire entity network. Furthermore, various variations are possible regarding the specific processing flow. Below, we will mainly describe an example in which the target node is acquired, subnetworks related to that target node are extracted, and processing is performed on those subnetworks.

[0032] The controlled node identification unit 114 performs the process of selecting a controlling node that effectively controls other nodes from the subnetwork, and the process of identifying controlled nodes that are effectively controlled by the controlling node. Here, effective control means being able to substantially control the actions of other entities. For example, if an entity can determine the actions of another entity based on its investment relationship, that entity effectively controls the other entity. More specifically, the actions here refer to actions in resolutions of other entities, and for example, effective control may mean being able to determine whether the resolution is passed and / or rejected. In this case, effectively controlling another entity may mean directly or indirectly owning a majority of the voting shares issued by the other entity. The processing of the controlled node identification unit 114 will be described later using Figures 9-14, etc. Note that in the entity network 121, each node corresponds to an entity, so the control relationships between entities are expressed as control relationships between nodes. In other words, in this specification, "the entity corresponding to the dominant node effectively controls the entity corresponding to the controlled node" is simply expressed as "the dominant node effectively controls the controlled node." Similarly, the term "node" in the following description is not limited to the node itself in the network (directed graph), but may also represent the entity corresponding to that node.

[0033] The influence calculation unit 115 calculates the influence level, which is an index representing the degree of influence a dominant node has on other nodes. In this embodiment, the influence level is an index value that takes the maximum value (e.g., 1) when the target node is a controlled node, the minimum value (e.g., 0) when the target node is not directly connected to either a dominant node or a controlled node, and an intermediate value (e.g., greater than 0 and less than 1) that changes according to the degree of dominance when the target node is not a controlled node and the node directly above it is either a dominant node or a controlled node. In this embodiment, other influence levels, such as the influence level shown in Patent Document 1 (hereinafter referred to as the Power Index), may be used in combination.

[0034] The display processing unit 116 performs the process of displaying the processing results in this embodiment on a display unit. The display unit here is, for example, the display unit 240 of the terminal device 200. However, the display target in the display processing unit 116 may be the display unit of the server system 100 or the display unit of another device.

[0035] The display processing unit 116 performs processing to display controlling nodes and controlled nodes in the entity network in a manner that allows them to be identified from other nodes. Details will be described later using Figures 13-14, 25-26, etc. The display processing unit 116 may also perform processing to display a screen that includes the influence calculated by the influence calculation unit 115, or it may perform processing to display other indicator values ​​such as the indirect shareholding ratio and Power Index.

[0036] The storage unit 120 is the work area of ​​the processing unit 110 and stores various information. The storage unit 120 can be implemented using various types of memory. The memory may be semiconductor memory such as SRAM, DRAM, ROM, or flash memory. The memory may also be magnetic storage devices such as registers or hard disk drives, or optical storage devices such as optical disk drives.

[0037] The storage unit 120 stores, for example, the entity network 121 acquired by the entity network acquisition unit 111. The storage unit 120 can store various information related to the processing of this embodiment.

[0038] The communication unit 130 is an interface for communication over a network. The communication unit 130 includes, for example, an antenna, an RF (radiofrequency) circuit, and a baseband circuit. The communication unit 130 may operate according to the control of the processing unit 110, or it may include a separate communication control processor. The communication unit 130 is, for example, an interface for communication compliant with TCP / IP (Transmission Control Protocol / Internet Protocol). Various variations of the specific communication method are possible.

[0039] Figure 3 is a block diagram showing a detailed configuration example of the terminal device 200. The terminal device 200 includes a processing unit 210, a storage unit 220, a communication unit 230, a display unit 240, and an operation unit 250.

[0040] The processing unit 210 is comprised of hardware including at least one of a circuit for processing digital signals and a circuit for processing analog signals. The processing unit 210 may also be implemented by a processor. Various types of processors can be used, such as a CPU, GPU, or DSP. The functions of the processing unit 210 are realized as processing when the processor executes instructions stored in the memory of the terminal device 200.

[0041] The memory unit 220 is the work area of ​​the processing unit 210 and is implemented by various types of memory such as SRAM, DRAM, and ROM.

[0042] The communication unit 230 is an interface for communication over a network. The communication unit 230 includes, for example, an antenna, an RF circuit, and a baseband circuit. The communication unit 230 communicates with the server system 100 over the network, for example.

[0043] The display unit 240 is an interface for displaying various information. The display unit 240 may be a liquid crystal display, an organic EL display, or another type of display. Based on control from the display processing unit 116 of the server system 100, the display unit 240 displays, for example, a screen including an entity network where the display modes of the controlling node and controlled nodes differ from the display modes of other nodes.

[0044] The operation unit 250 is an interface for the user to operate the terminal device 200. The operation unit 250 may be a button or the like provided on the terminal device 200. The display unit 240 and the operation unit 250 may be a touch panel that is configured as a single unit.

[0045] 1.2 Examples of Services Next, we will describe specific examples of services provided by the information processing system 10, which is an OSINT system. Below, we will explain an example of a specific service: corporate holding network analysis.

[0046] Figure 4 illustrates corporate holding network analysis and is an example of an entity network representing investment relationships. As shown in Figure 4, a network representing investment relationships between countries and companies is formed based on publicly available information, including shareholders and investment ratios.

[0047] The controlled node identification unit 114 identifies controlled nodes that are effectively controlled by a controlled node, for example, when one of the nodes in the network to be processed is designated as the controlling node. The influence calculation unit 115 analyzes the degree of influence that various countries and companies have on other companies. In this case, influence refers to the degree to which control is exerted through investment.

[0048] For example, if a particular country is considered a controlling node, it is possible to determine the extent to which that country controls the supply of products in a given industry by determining the controlled nodes that are effectively controlled by that controlling node, or the degree of influence that controlling node has on companies in a given industry. For example, if a major incident occurs in that country, it is possible to evaluate the impact that incident will have on the stable supply of products. The controlled node identification unit 114 and the influence calculation unit 115 may also determine whether each country has effective control or influence over global companies. This makes it possible to grasp the balance of power between nations. Furthermore, by determining the time-series changes in each country's influence over global companies, it is also possible to grasp the changes in the above balance of power.

[0049] Alternatively, the controlled node identification unit 114 and the impact calculation unit 115 may determine whether another country has effective control over and influence over companies related to infrastructure in a given country. Companies related to infrastructure may be energy-related companies such as electricity companies, or companies that provide mobile communication networks. This makes it possible to assess the risk of infrastructure failure.

[0050] Alternatively, the controlled node identification unit 114 and the influence calculation unit 115 may determine whether they have effective control over and influence over companies that possess technologies that can be used for military purposes. This makes it possible to detect security risks.

[0051] The controlled node identification unit 114 and the impact calculation unit 115 may determine the change in the target of effective control and the change in impact when a country or company takes a specific action. For example, assuming that a given country's foreign policy has been changed, it is possible to simulate the impact that foreign policy has on countries around the world by determining the scope of effective control and the impact before and after the change.

[0052] The controlled node identification unit 114 and the influence calculation unit 115 utilize corporate holding network analysis, enabling analysis based on complex investment relationships that would be difficult to detect manually.

[0053] In recent years, relationships between nations, corporations, and key figures have become more global and complex than ever before. Therefore, manual analysis has its limitations. In this respect, the aforementioned OSINT system can analyze networks representing corporate control through investment, for example. Because OSINT systems can decipher complex relationships, they enable governments and corporations to formulate optimal strategies.

[0054] 2. Processing Details The details of the processing in this embodiment will be described below. The method of this embodiment is, in a narrow sense, corporate holding network analysis. However, the method of this embodiment is applicable to any method other than corporate holding network analysis.

[0055] 2.1 Indicators representing the degree of impact First, let's consider the indicators that represent the degree of impact. Figure 5 shows an example of an entity network 121 acquired by the entity network acquisition unit 111, or a subnetwork extracted by the subnetwork extraction unit 113. The network in Figure 5 includes nine nodes, nodes A to I, which correspond to entities A to I, respectively. In Figure 5, the root of the edge of the arrow is connected to the node corresponding to the investing side (upper node, upstream node), and the tip is connected to the node corresponding to the receiving side (lower node, downstream node).

[0056] In the network shown in Figure 5, nodes E, F, G, and H are each connected to node I at the edge, and each edge is associated with a 51% ownership stake. Therefore, entity I has acquired more than half of the shares in each of entities E through H, and thus effectively controls entities E through H. As mentioned above, to simplify the explanation, expressions such as "the ownership stake of entity E, corresponding to node E, by entity I, corresponding to node I, is 51%" will be written as "the ownership stake of node E, corresponding to node I, is 51%."

[0057] In the network shown in Figure 5, Node B is owned 30% by Node E and 30% by Node F. Node C is owned 30% by Node E and 30% by Node G. Node D is owned 30% by Node F and 30% by Node H. Node A is owned 25% by Node B, 20% by Node C, and 30% by Node D.

[0058] For example, the widely known indirect ownership ratio is determined based on the product of the ownership ratios along the path connecting two nodes. For instance, between node I and node A, there exists a path from node I → node E → node B → node A. Hereafter, we will simply denote the path between nodes as path IEBA, etc., by listing the node names. Since the ownership ratios assigned to the edges along this path are 0.51, 0.30, and 0.25, the indirect ownership ratio of node I to node A via path IEBA can be calculated as 0.51 × 0.30 × 0.25. Since other paths also exist between node I and node A, the same values ​​are calculated for all paths, and the sum of the values ​​obtained for each path determines the indirect ownership ratio of node I to node A.

[0059] Furthermore, in the Power Index disclosed in Patent Document 1, if the shareholding ratio exceeds half, the shareholding ratio is changed to 1 and the same calculation as for the indirect shareholding ratio is performed. Figure 6A is a diagram illustrating the method of Patent Document 1. In the example of Figure 6A, consider the Power Index of node C with respect to node A. In this case, since the shareholding ratio of node C in node D exceeds half, the value is replaced with 1. Therefore, the Power Index of node C with respect to node A is 0.60 (60%), which is the sum of 0.30 in path CA and 1 × 0.30 in path CDA.

[0060] In the network shown in Figure 6A, Node C effectively controls Node D because it holds a majority of Node D's shares. Therefore, Node C can freely exercise the 30% of Node A's shares that it directly holds and the 30% held by Node D. As described above, the Power Index is 60 percent, making it a suitable indicator of Node C's influence on Node A. In other words, the Power Index reflects the actual degree of influence more accurately than the previous indirect shareholding ratio.

[0061] Figure 6B shows an example of applying the method described in Patent Document 1 to the network described in Figure 5. In this case, since node I's ownership ratio in nodes E to H exceeds half, these ownership ratios are replaced with 1. Therefore, the indirect ownership ratio along the IEBA route is corrected to 1 × 0.30 × 0.25. The same applies to other routes. When calculated for all routes, the Power Index representing node I's influence on node A becomes 45%. In this case, since the Power Index is less than half, it gives the impression that node I does not effectively control node A.

[0062] However, when considering the network in Figure 6B (Figure 5), such a judgment is not necessarily correct. For example, in this network, node I owns more than half of the shares in nodes E through H, so node I effectively controls nodes E through H.

[0063] Furthermore, Node B is held at a 30% stake by Node E and Node F, and since Node E and Node F are effectively controlled by Node I, this is equivalent to Node I owning a majority of the shares in Node B. In other words, Node I effectively controls Node B. Similarly, Node C and Node D are also effectively controlled by Node I because the combined shareholdings of nodes effectively controlled by Node I constitute a majority.

[0064] Furthermore, nodes B through D are effectively controlled by node I, and the combined shareholding ratio of these three nodes in node A is 75%, meaning that node A is also effectively controlled by node I. Thus, in the network shown in Figure 6B (Figure 5), although node I effectively controls node A, the Power Index, which represents node I's influence on node A, is relatively small.

[0065] The method described in Patent Document 1 has the advantage of reflecting the effective control relationship by updating the value to 1 if one node holds a majority stake in another node. However, it does not accurately reflect the effective control relationship when multiple nodes jointly hold a majority stake. For example, in Figure 6B, Node E and Node F jointly hold 60% of Node B's shares, but since their individual stakes do not exceed 50%, the value is not updated and the Power Index is calculated. Although Node E and Node F are both effectively controlled by Node I and could cooperate on resolutions concerning Node B, this is not taken into consideration by the method described in Patent Document 1.

[0066] Figure 7 shows another example of a network where the Power Index tends to be calculated as lower than the actual degree of influence. In the network in Figure 7, node I effectively controls nodes J through M. By combining the ownership ratios of two of nodes J through M, node I effectively controls nodes E through H. Furthermore, by combining the ownership ratios of two of nodes E through H, node I effectively controls nodes B through D. And, similar to the example in Figure 6B, by combining the ownership ratios of nodes B through D, node I effectively controls node A. Thus, in the example in Figure 7, node I also effectively controls node A. However, downstream of nodes J through M, the ownership ratio assigned to each edge does not exceed half, so the above effective control relationships are not taken into consideration, and the Power Index is calculated using the original value (30%). As a result, the Power Index, which represents the degree of influence of node I on node A, is lower than 50%.

[0067] For example, as shown in Figure 7, when a network structure in which the combined ownership ratios of multiple nodes effectively controlled by a particular node exceed half is referred to as an attenuation unit, the Power Index value tends to be smaller than the actual degree of influence in networks where such attenuation units exist. In the example in Figure 6B, the structure between nodes E-H and nodes B-D corresponds to an attenuation unit. Similarly, in the example in Figure 6B, the structure between nodes B-D and node A corresponds to an attenuation unit. Furthermore, in the example in Figure 7, there are three attenuation units between nodes J-M and nodes E-H, between nodes E-H and nodes B-D, and between nodes B-D and node A, so the Power Index value is even smaller compared to the network in Figure 6B.

[0068] In contrast, the information processing system 10 (for example, the server system 100) according to this embodiment includes an entity network acquisition unit 111 and a controlled node identification unit 114, as shown in Figure 2. The controlled node identification unit 114 identifies controlled nodes by performing an update process to add to the controlled nodes one or more nodes from among a plurality of nodes that are directly connected to a second node group which is included in the first node group, and whose sum of the investment ratios assigned to the edges between each node of the second node group is greater than a first threshold, when the controlling nodes and already identified controlled nodes are considered as a first node group.

[0069] Here, a controlling node refers to a node in an entity network or subnetwork that has the potential to control other nodes based on investment relationships (more specifically, a node that has the potential to control a target node). The first group of nodes is, as described above, the set of controlling nodes and nodes that are known to be effectively controlled by that controlling node. In other words, the first group of nodes is the set of nodes that act in coordination with the controlling node when the controlling node attempts to exert influence over a company (for example, when it attempts to exercise voting rights over a company).

[0070] Furthermore, the second node group refers to the nodes included in the first node group that are directly connected to any node other than the first node group (located directly above, i.e., one level upstream) when considering any node other than the first node group. For example, in the example in Figure 5, if node I is the dominant node and nodes E to H have been identified as the controlled nodes, then nodes E and F are the nodes directly above node B. In this case, the first node group is the set of nodes E to I, and the second node group for node B is the set of nodes I and F. Here, nodes other than the first node group being processed (for example, node B) are not included in the controlled nodes at the stage when processing is being performed, but since they receive investment from each node included in the second node group (for example, nodes E and F), it is possible to consider that they are effectively receiving investment from the dominant node (node ​​I).

[0071] Furthermore, the first threshold is, for example, 0.50 (50%). In this case, controlled nodes can be determined, for example, from the perspective of whether the controlling node has enough influence to unilaterally pass an ordinary resolution at a shareholders' meeting. However, the first threshold may also be 2 / 3 (66.7%), which corresponds to the shareholding ratio that allows a special resolution at a shareholders' meeting to be passed unilaterally. Alternatively, it may be 1 / 3 (33.3%), which corresponds to the shareholding ratio that allows a special resolution at a shareholders' meeting to be rejected unilaterally. Moreover, the first threshold may be set to a value different from any of these.

[0072] In this embodiment, the controlled node identification unit 114 uses the sum of the investment ratios (sum of shareholding ratios) by the second group of nodes to identify controlled nodes. Therefore, the method of this embodiment makes it possible to appropriately determine the control relationship. For example, as described above, it becomes possible to appropriately detect a control relationship in which multiple nodes effectively controlled by a controlling node jointly hold a majority of the shares. The specific processing will be described later. It should be noted that the method of this embodiment does not prevent the use of other influence indicators such as the Power Index.

[0073] Furthermore, some or all of the processing performed by the information processing system 10 in this embodiment may be implemented by a program. The processing performed by the information processing system 10 refers, in a narrow sense, to the processing performed by the processing unit 110 of the server system 100, but may also include the processing performed by the processing unit 210 of the terminal device 200.

[0074] The program according to this embodiment can be stored in a non-temporary information storage medium (information storage device), which is a medium readable by a computer. The information storage medium can be implemented by, for example, an optical disc, a memory card, an HDD, or a semiconductor memory. The semiconductor memory is, for example, ROM. The processing unit 110, etc., performs various processing according to this embodiment based on the program stored in the information storage medium. That is, the information storage medium stores a program that causes the computer to function as the processing unit 110, etc. A computer is a device that includes an input device, a processing unit, a storage unit, and an output unit. Specifically, the program according to this embodiment is a program that causes the computer to execute each of the steps described later using Figures 8, 9, 16, etc.

[0075] Furthermore, the method of this embodiment can be applied to an information processing method that includes the following steps. The information processing method involves an information processing device (e.g., an information processing system 10, a server system 100) acquiring an entity network that represents the mutual investment relationships of multiple nodes corresponding to multiple entities by connecting the lower-level nodes of the invested side and the upper-level nodes of the investor side with edges to which investment ratios are assigned. Based on the entity network, the device identifies controlled nodes, which are nodes that are effectively controlled by a dominant node among the multiple nodes. In identifying controlled nodes, when the dominant node and already identified controlled nodes are considered as the first node group, the device performs an update process to identify controlled nodes by adding to the controlled nodes one or more nodes from the multiple nodes that are directly connected to the second node group, which is one or more nodes included in the first node group, and whose sum of investment ratios assigned to the edges between each node of the second node group is greater than a first threshold.

[0076] 2.2 Overall Processing Figure 8 is a flowchart illustrating the general process performed in the information processing system 10 of this embodiment.

[0077] First, in step S101, the entity network acquisition unit 111 acquires the entity network 121. The entity network acquisition unit 111 stores the entity network 121 in the storage unit 120.

[0078] In step S102, the target node acquisition unit 112 acquires any node included in the entity network 121 as the target node. For example, in step S102, the display processing unit 116 may prompt the user of the terminal device 200 to select a target node by displaying the entity network 121, a list screen of entities included in the entity network 121, or an entity search screen on the display unit 240 of the terminal device 200. For example, the operation unit 250 of the terminal device 200 may receive input from a company of interest that the user wants to investigate the degree of control from other entities, and the target node acquisition unit 112 may acquire the node corresponding to that company of interest as the target node via the communication unit 130.

[0079] In step S103, the subnetwork extraction unit 113 extracts subnetworks that include the target node. For example, the subnetwork extraction unit 113 first initializes network N with the target node. Next, the subnetwork extraction unit 113 updates network N by adding nodes that have directly invested in the target node and edges representing the investment relationship. Then, for newly added nodes, the subnetwork extraction unit 113 updates network N by adding nodes that have directly invested in that node and edges representing the investment relationship. The subnetwork extraction unit 113 recursively repeats this process, and when network N no longer changes, it stores the current network N as a subnetwork in the storage unit 120. As described above, the subnetwork here is a network composed of nodes upstream of the target node and represents the ownership relationship, so it may also be referred to as the upstream ownership network.

[0080] In step S104, the controlled node identification unit 114 identifies controlled nodes that are effectively controlled by a controlled node, based on the assumption that a node potentially controlling the target node is designated as the controlling node among the multiple nodes included in the subnetwork. This process propagates the control relationship by the controlling node to the downstream side, and is therefore also referred to as the control propagation process. Details of the controlled node identification process will be described later.

[0081] In step S105, the influence calculation unit 115 determines the degree of influence that the controlling node has on other nodes based on the identification of the controlled node. Details of the influence calculation process will be described later.

[0082] In step S106, the display processing unit 116 performs a process to display the result of identifying the controlled node on the display unit 240 of the terminal device 200. The display processing unit 116 may also perform a process to display the result of calculating the degree of influence on the display unit 240 of the terminal device 200. For example, the display processing unit 116 may display the result of determining whether the target node selected by the user is effectively controlled by another node, or it may display the degree of influence of the controlling node on the target node.

[0083] 2.3 Identifying Controlled Nodes (Control Propagation Process) Figure 9 is a flowchart illustrating the process of identifying the controlled node shown in step S104 of Figure 8.

[0084] In step S201, the controlled node identification unit 114 selects a controlling node from among multiple nodes included in the subnetwork. For example, the controlled node identification unit 114 selects a node as the controlling node that, among the nodes other than the target node included in the subnetwork, has an investment ratio to at least one other node greater than the first threshold. Specifically, for each node other than the target node included in the subnetwork, the controlled node identification unit 114 determines whether there is an edge in the direction from that node to another node where the assigned investment ratio exceeds the first threshold. The controlled node identification unit 114 selects nodes that satisfy the conditions as candidate nodes for controlling nodes. Then, the controlled node identification unit 114 sequentially selects one of the candidate nodes as the controlling node. The first threshold here is, as mentioned above, for example, 0.50 (50%), but other values ​​may be used.

[0085] In the network shown in Figure 5, node I's ownership ratios with respect to each of the other nodes, E through H, exceed the first threshold of 0.50. Therefore, node I is selected as a candidate for the dominant node. Node E's ownership ratio with respect to node B is 0.30, which does not exceed the first threshold, and its ownership ratio with respect to node C is also 0.30, which does not exceed the first threshold. Therefore, node E is not a candidate for the dominant node. The same applies to the other nodes; in Figure 5, only node I is a candidate for the dominant node. Therefore, the controlled node identification unit 114 performs the following processing targeting node I to identify the controlled nodes that are effectively controlled by node I.

[0086] In step S202, the controlled node identification unit 114 initializes the set P of controlled nodes that are effectively controlled by the controlling node with an empty set.

[0087] In the following process, the set of the controlling node itself and the controlled nodes that are effectively controlled by that controlling node is referred to as the first node group. The first node group is the union of the controlling node and set P. Figure 10 shows the first node group after the processing in step S202. As shown in Figure 10, since no controlled nodes are identified in step S202, only node I, which is the controlling node, is included in the first node group.

[0088] In step S203, the controlled node identification unit 114 performs a first update process to update the set of controlled nodes P. Specifically, it performs an update process to add nodes controlled by the controlling node with a shareholding ratio exceeding a first threshold to the set of controlled nodes P.

[0089] In the case of the network shown in Figure 5, node I's ownership ratio in each of the other nodes, E to H, exceeds the first threshold of 0.5. Therefore, in step S203, the controlled node identification unit 114 executes the process of adding nodes E to H to the set P.

[0090] Figure 11 shows the first group of nodes after the processing in step S203. As shown in Figure 11, the first update process in step S203 adds nodes E to H, which are directly controlled by node I, the dominant node, with a shareholding ratio exceeding the first threshold, to the set of controlled nodes P. Therefore, the first group of nodes becomes the set of five nodes, E to I. In other words, the first update process propagates the dominance relationship from the dominant node down one level downstream.

[0091] In step S204, the controlled node identification unit 114 performs a second update process to update the set of controlled nodes P. Specifically, the controlled node identification unit 114 performs an update process to add to the set of controlled nodes P any nodes that are controlled by nodes included in the first node group, either individually or jointly, with a shareholding ratio exceeding the first threshold.

[0092] More specifically, the controlled node identification unit 114 identifies nodes that are directly connected from the subnetwork to the second node group of the first node group. In other words, the nodes identified here are nodes one step downstream of one or more nodes in the first node group. The controlled node identification unit 114 then calculates the sum of the shareholding ratios by the second node group for each identified node, and if this sum exceeds the first threshold, it adds the node to be processed to the set of controlled nodes P.

[0093] For example, the first group of nodes immediately before the first second update process consists of five nodes, E to I, as described above using Figure 11. In this case, nodes B to D are directly connected to the nodes included in the first group of nodes. In other words, each of nodes B to D is a node directly below (one level downstream) the node included in the first group of nodes. Therefore, the controlled node identification unit 114 calculates the sum of the shareholding ratios for each of nodes B to D.

[0094] Specifically, node B is directly connected to nodes E and F, which are included in the first node group (it is a node directly under node E and node F). Therefore, the second node group for node B is node E and node F. The controlled node identification unit 114 calculates the sum of node E's shareholding ratio with node B and node F's shareholding ratio with node B. In this case, since each shareholding ratio is 30%, the sum of the shareholding ratios is 60%. Since the sum of the shareholding ratios exceeds the first threshold of 50%, the controlled node identification unit 114 adds node B to the set of controlled nodes P.

[0095] Similarly, the second group of nodes for node C is nodes E and G. The sum of the shareholding ratios by nodes E and G is 60%, which exceeds the first threshold, so the controlled node identification unit 114 adds node C to the set of controlled nodes P. Also, the second group of nodes for node D is nodes F and H. The sum of the shareholding ratios by nodes F and H is 60%, which exceeds the first threshold, so the controlled node identification unit 114 adds node D to the set of controlled nodes P.

[0096] Figure 12 shows the first group of nodes after the first second update process. As shown in Figure 12, the first second update process adds nodes B to D, which are jointly controlled by the controlled nodes E to H with a shareholding ratio exceeding the first threshold, to the set of controlled nodes P. Therefore, the first group of nodes becomes the set of eight nodes, B to I. In other words, the second update process propagates the control relationship by the controlling nodes down to the node one level downstream from the state before the process. Note that although an example is shown here where the second group of nodes for each node includes multiple nodes, there is no prerequisite that the second group of nodes be a single node.

[0097] After the execution of the second update process, in step S205, the controlled node identification unit 114 determines whether the set of controlled nodes P has changed as a result of the second update process. In the example above, nodes B to D have been added to the set of controlled nodes P, so the set P has changed.

[0098] If the set of controlled nodes P has changed due to the second update process (step S205: Yes), the controlled node identification unit 114 returns to step S204 and recursively executes the second update process. For example, the first group of nodes after the first second update process is shown in Figure 12, and consists of nodes B to I. In this case, node A is directly connected to a node included in the first group of nodes (it is a node directly under a node included in the first group of nodes). Therefore, in the second second update process, the controlled node identification unit 114 calculates the sum of the shareholding ratios for node A.

[0099] Specifically, node A is connected to nodes B through D, which are included in the first node group. Therefore, the second node group for node A is nodes B through D. The controlled node identification unit 114 calculates the sum of node B's shareholding ratio with node A, node C's shareholding ratio with node A, and node D's shareholding ratio with node A. In this case, the sum of the shareholding ratios is 0.25 + 0.20 + 0.30 = 0.75 (75%). Since the sum of the shareholding ratios exceeds the first threshold of 50%, the controlled node identification unit 114 adds node A to the set of controlled nodes P.

[0100] Figure 13 shows the first group of nodes after the second update process has been performed. As shown in Figure 13, the second update process adds node A, which is jointly controlled by the controlled nodes B to D with a shareholding ratio exceeding the first threshold, to the set of controlled nodes P. Therefore, the first group of nodes becomes a set of nine nodes, A to I. In other words, the second update process propagates the control relationships of the controlling nodes down to the node one level downstream from the state before the process.

[0101] In this embodiment, the controlled node identification unit 114 may identify controlled nodes by performing an initialization process to initialize controlled nodes with an empty set (step S202), then performing a first update process to add nodes that are directly connected to a dominant node and whose investment ratio assigned to the edge between them and the dominant node is greater than a first threshold (step S203), and then repeatedly executing a second update process after the first update process (step S204). In this way, it becomes possible to efficiently determine the dominance path and dominance range by a dominant node through a top-down process that propagates the dominance relationship downstream starting from the dominant node.

[0102] After the execution of the second update process, in step S205, the controlled node identification unit 114 determines whether the set P of controlled nodes has changed as a result of the second update process. In the example above, since node A has been added to the set P of controlled nodes, the set P has changed. Therefore, the controlled node identification unit 114 returns to step S204 and executes the second update process for the third time.

[0103] As shown in Figure 13, in the network illustrated here, all nodes of the subnetwork are added to the first group of nodes during the second update process. Therefore, even if a third update process is performed, the set of controlled nodes P will not change.

[0104] Alternatively, the set of controlled nodes P remains unchanged in the network shown in Figure 14. The network shown in Figure 14 is similar to the network described above using Figures 10-13 for nodes A to I, with the addition of node J as a node directly under node A. Node A's shareholding ratio in node J is 5%.

[0105] In this case, as in the example above, the second update process results in nodes A to I being included in the first node group. Since node J is directly connected to node A, which is included in the first node group, the sum of the shareholding ratios can be calculated. For node J, the second node group consists only of node A, so the sum of the shareholding ratios is 5%. The controlled node identification unit 114 does not add node J to the set of controlled nodes P because the sum of the shareholding ratios does not exceed the first threshold. In this way, even if there are nodes that have not been added to the set of controlled nodes P, the set of controlled nodes P will not change as long as there are no nodes whose sum of shareholding ratios by the second node group exceeds the first threshold.

[0106] The controlled node identification unit 114 terminates the second update process if the controlled node does not change even after performing the second update process (step S205: No). This makes it possible to properly terminate the recursively executed second update process. At this stage, the controlled node identification unit 114 identifies the nodes included in the set P as controlled nodes that are effectively controlled by the controlling node.

[0107] Then, in step S206, the controlled node identification unit 114 determines whether processing has been completed for all candidate nodes of the controlling node. For example, if multiple candidate nodes were found in step S201 and there are still candidate nodes that are not subject to processing in steps S202-S205 (step S206: No), the controlled node identification unit 114 returns to step S201 and selects the unprocessed candidate node as the new controlling node. The processing in steps S202-S205 after the new controlling node has been selected is the same as in the example described above.

[0108] If the processing in steps S202-S205 is completed for all candidate nodes (step S206: Yes), the controlled node identification unit 114 terminates the controlled node identification process shown in Figure 9.

[0109] 2.4 Impact Calculation Process Next, the influence calculation process shown in step S105 of Figure 8 will be explained. The influence calculation unit 115 of this embodiment calculates an index value as the influence based on the state of effective control of other nodes by the controlling node.

[0110] For example, the influence calculation unit 115 sets the influence of the controlling node on the controlled nodes to the maximum value. In this way, it becomes possible to set an appropriate value as the influence for nodes that are effectively controlled by the controlling node. In the example in Figure 14, node I effectively controls nodes A to H. Therefore, the influence calculation unit 115 sets the influence of node I on each of nodes A to H to the maximum value (for example, 1).

[0111] Furthermore, if the target node among multiple nodes is not included in the first node group after the process of identifying the controlled node is completed, the influence calculation unit 115 may determine the influence of the controlling node on the target node based on the investment ratio assigned to the edge directly connecting the node included in the first node group and the target node. In this way, it becomes possible to set an influence based on the shareholding ratio for a target node whose shares have been acquired by a group consisting of a controlling node and a group of nodes effectively controlled by that controlling node (the first node group). In the example in Figure 14, node J is connected to node A, which is included in the first node group, and node A's shareholding ratio in node J is 5%. Therefore, the influence calculation unit 115 sets 0.05 (5%), which corresponds to the shareholding ratio, as the influence of node I on node J. In this way, the influence is a value corresponding to the voting rights (shareholding ratio) that the controlling node can effectively exercise, so it is possible to set an index value that is easy to understand in terms of its meaning.

[0112] For example, under Japanese company law, shareholders with a shareholding ratio exceeding 1% are granted the right to request proposals at general shareholders' meetings in companies with a board of directors. Shareholders with a shareholding ratio exceeding 3% are granted the right to request the convening of general shareholders' meetings, the right to inspect and copy accounting books, etc. By using the influence level of this embodiment, it becomes possible to easily determine whether a controlling node can exercise these rights over a target node.

[0113] The influence calculation unit 115 also identifies one or more edges that directly connect the target node to any of the nodes included in the first node group, and calculates the sum of the investment ratios assigned to the identified one or more edges as the influence of the controlling node on the target node.

[0114] For example, in Figure 14, node J is connected only to node A. Now consider a network in Figure 14 where an edge is added from node B to node J, with an associated ownership ratio of 10%. In this case, there are two edges that directly connect to the target node, node J: one from node A to node J, and another from node B to node J. The influence calculation unit 115 sets the sum of the ownership ratios assigned to each edge, 0.05 + 0.10 = 0.15 (15%), as the influence of node I on node J. Node I can exercise voting rights equivalent to 5% of all shares (specifically, voting shares) of node J via node A, and can also exercise voting rights equivalent to 10% of all shares of node J via node B. In other words, when multiple nodes included in the first node group are directly above the target node, it becomes possible to set an influence value corresponding to the voting rights (ownership ratio) that the controlling node can effectively exercise by using the sum of the ownership ratios of each edge.

[0115] Furthermore, the influence calculation unit 115 sets the influence of the controlling node on the target node to the minimum value if the target node is not directly connected to any node included in the first node group. For example, consider a network in which node K is added directly under node J to the network shown in Figure 14. In this case, node K is connected only to node A via node J, which is not included in the first node group, and is not directly connected to any node included in the first node group. In this case, the influence calculation unit 115 sets the influence of node I on node K to the minimum value (e.g., 0). As mentioned above, node I can only exercise voting rights equivalent to 5% of node J via node A, and cannot control the actions of node J in resolutions at node K. In other words, node I cannot definitively exert influence in resolutions at node K, so setting the influence to the minimum value makes it possible to set a value that is in line with the actual situation.

[0116] 2.5 Output Processing In step S106 of Figure 8, the display processing unit 116 may perform a process to display the controlling node and controlled node in the entity network in a manner that allows them to be distinguished from other nodes. For example, as shown in Figure 13 or Figure 14, the display processing unit 116 displays the first group of nodes in a first manner at the stage when the process of identifying controlled nodes (step S104) has been completed, and displays the nodes other than the first group of nodes in a second manner different from the first manner, and displays the network on the display unit 240 of the terminal device 200. In this way, it becomes possible to present to the user in an easy-to-understand manner the paths through which the controlling node effectively controls other nodes. The first and second manners are, for example, display manners in which the node colors are different, but are not limited to this, and the first manner may be a manner that is more visible than the second manner. For example, the node size may be larger in the first manner than in the second manner. Alternatively, the display processing unit 116 may display only the nodes in the first manner on the default screen, and then perform a process to display the nodes in the second manner when some user operation is received. Furthermore, various variations can be implemented to show the specific display configuration. The display processing unit 116 may also display edges where both end nodes are included in the first node group in a manner that is more visible than other edges (edges where at least one end node is not part of the first node group).

[0117] The display processing unit 116 may also perform processing to display the influence level values ​​of the controlling node for nodes other than the first group of nodes.

[0118] Furthermore, the display processing unit 116 may be configured to switch between displaying the impact level according to this embodiment and displaying other impact levels such as the Power Index. In addition, various modifications can be made to the specific screens displayed by the display processing unit 116.

[0119] 2.6 Matrix operations The process of identifying controlled nodes and calculating their influence, as described above, may also be performed using matrix and vector operations. Specific examples are described below.

[0120] Figure 15A shows an example of a network used for explanation. The network here is, for example, the subnetwork extracted in step S103 of Figure 8. Here, we consider a network containing nodes 1-7, which correspond to entities 1-7. The connection relationships of each node and the ownership ratios assigned to the edges are as shown in the figure.

[0121] The controlled node identification unit 114 sets up a connection matrix that represents the connection relationships of multiple nodes included in the network. For example, if the entity network includes a subnetwork that is directly or indirectly connected to the controlling node and contains the 1st to Nth nodes (where N is an integer greater than or equal to 2), the controlled node identification unit 114 calculates an N x N connection matrix in which the value of the element in row i (where i is an integer between 1 and N) and column i is 1, and the value of the element in row i and column j (where j is an integer between 1 and N that is different from i) represents the investment ratio of the i-th node to the j-th node.

[0122] Figure 15B shows an example of a connection matrix for the subnetwork in Figure 15A. As shown in Figure 15B, the row elements of the connection matrix represent the dominating nodes, and the column elements represent the controlled nodes. First, the controlled node identification unit 114 sets the values ​​of the diagonal elements of the connection matrix to 1.

[0123] The controlled node identification unit 114 determines the values ​​of the other components based on the connection relationships between nodes and the shareholding ratios assigned to the edges. For example, in the subnetwork shown in Figure 15A, node 1's shareholding ratio with node 3 is 0.6. Therefore, the value of the component in row 1, column 3 of the connection matrix is ​​set to 0.6. Similarly, node 1's shareholding ratio with node 4 is 0.9. Therefore, the value of the component in row 1, column 4 of the connection matrix is ​​set to 0.9. The same applies thereafter; based on the connection relationships of the subnetwork shown in Figure 15A, values ​​representing the shareholding ratios are set for the components in rows 2, column 4, row 3, column 5, row 4, column 5, row 5, column 6, and row 6, column 7 of the connection matrix. For the other elements, since node i does not have an investment relationship with node j, the values ​​are set to 0.

[0124] Figure 16 is a flowchart illustrating the process of identifying controlled nodes using matrix operations and calculating their influence. It is assumed that the connection matrix has already been determined before this process.

[0125] In step S301, the controlled node identification unit 114 selects node i as the controlling node. The process in step S301 is the same as the process in step S201 in Figure 9. That is, the controlled node identification unit 114 selects nodes other than the target node included in the subnetwork that control other nodes with a shareholding ratio exceeding the first threshold as candidate nodes for controlling nodes. The controlled node identification unit 114 then sequentially selects one of the candidate nodes as the controlling node. In the example in Figure 15A, nodes 1 and 5 are candidate nodes for controlling nodes, so the controlled node identification unit 114 sequentially selects 1 and 5 as i. Here, we first consider the example where i=1 is set.

[0126] In step S302, the controlled node identification unit 114 identifies the control vector v, which is a vector representing the control relationship by the controlling node (node ​​i), and vector e i Initialized by the vector e. iThis is an N-dimensional row vector where only the value of the i-th component is 1, and the values ​​of the other components are set to 0. When i=1, as shown in Figure 17A, the dominant vector v is set to e1, and e1 is a row vector where the first component is 1 and the other components are 0. The controlled node identification unit 114 performs the process of determining controlled nodes by multiplying the dominant vector and the connection matrix. In this way, it becomes possible to determine controlled nodes using operations with matrices and vectors.

[0127] For example, the controlled node identification unit 114 performs a vector update process on each component of the N-dimensional vector, which is the product of the dominance vector and the connection matrix. It updates the values ​​of components whose values ​​are greater than the first threshold to the first value, and updates the values ​​of components whose values ​​are less than or equal to the first threshold to the second value which is smaller than the first value. The dominance vector is then updated using the N-dimensional vector after this process. In this way, the dominance propagation process described above, as shown in Figures 10-13, can be realized in the form of updating the dominance vector.

[0128] Specifically, in step S303, the controlled node identification unit 114 multiplies the controlling vector v by the connection matrix C from the right, and the resulting row vector is v'. As shown in Figure 17A, in the first vector update process, the controlling vector is e1, so the result of the multiplication, v', is v'=(1,0,0.6,0.9,0,0,0).

[0129] In step S304, the controlled node identification unit 114 applies the first function F to each component of the vector v', and the resulting row vector is v''. Here, the first function F is a function that satisfies equation (1) below. In equation (1) below, θ is the value corresponding to the first threshold described above. In other words, the first function F is a function that updates the values ​​of components whose values ​​are greater than the first threshold to a first value (for example, 1), and updates the values ​​of components whose values ​​are less than or equal to the first threshold to a second value (for example, 0).

[0130]

number

[0131] As shown in Figure 17A, the first function F(F) with θ=0.5 is given by the vector v'. 0.5 By applying ), v'' = (1,0,1,1,0,0,0).

[0132] In step S305, the controlled node identification unit 114 determines whether the controlling vector v and vector v'' are equal. This process corresponds to step S205 in Figure 9 and is equivalent to determining whether control propagation is continuing. If the controlling vector v and vector v'' are not equal (step S305: No), in step S306, the controlled node identification unit 114 updates the controlling vector v with vector v'' and then returns to step S303 to execute the vector update process again. Hereinafter, the xth vector update process will be referred to as the xth step.

[0133] As explained above, as a result of the first step, the dominance vector v is updated to v=(1,0,1,1,0,0,0). This represents the state in which the dominance by node 1 has propagated to nodes 3 and 4, which correspond to the third and fourth components of the dominance vector v, as shown in Figure 17B.

[0134] In the second step, as shown in Figure 18A, the dominant vector v=(1,0,1,1,0,0,0) updated in the first step is multiplied by the connection matrix (step S303). The result of the multiplication, v', is v'=(1,0,1.6,1.9,0.6,0,0). Furthermore, by applying the first function F to vector v', v''=(1,0,1,1,1,0,0) is obtained (step S304). In this case as well, the dominant vector v and vector v'' are not equal (step S305: No), so in step S306, the controlled node identification unit 114 updates the dominant vector v with vector v'' and then returns to step S303 to execute the vector update process again.

[0135] As a result of the second step, the dominance vector v is updated to v=(1,0,1,1,1,0,0). This represents the state in which dominance by node 1 has propagated up to node 5, which corresponds to the fifth component of the dominance vector v, as shown in Figure 18B.

[0136] In the third step, as shown in Figure 19A, v=(1,0,1,1,1,0,0) is multiplied by the connection matrix (step S303). The result of the multiplication, v', is v'=(1,0,1.6,1.9,1.6,0.6,0). By applying the first function F to the vector v', v''=(1,0,1,1,1,1,0) is obtained (step S304). In this case as well, the dominating vector v and the vector v'' are not equal (step S305: No), so in step S306, the controlled node identification unit 114 updates the dominating vector v with the vector v'', and then returns to step S303 to execute the vector update process again.

[0137] As a result of the third step, the dominance vector v is updated to v=(1,0,1,1,1,1,0). This represents the state in which dominance by node 1 has propagated up to node 6, which corresponds to the sixth component of the dominance vector v, as shown in Figure 19B.

[0138] In the fourth step, as shown in Figure 20A, v=(1,0,1,1,1,1,0) is multiplied by the connection matrix (step S303). The result of the multiplication, v', is v'=(1,0,1.6,1.9,1.6,1.6,0.3). By applying the first function F to the vector v', v''=(1,0,1,1,1,1,0) is obtained (step S304). In this case, the dominating vector v and the vector v'' are equal (step S305: Yes), so the controlled node identification unit 114 terminates the vector update process.

[0139] As a result of the fourth step, the dominance vector v is updated to the state v=(1,0,1,1,1,1,0). This represents the state where dominance by node 1 has propagated to node 6, which corresponds to the sixth component of the dominance vector v, as shown in Figure 20B, and no further propagation has occurred.

[0140] The controlled node identification unit 114 identifies the nodes corresponding to the component whose value is a first value (for example, 1) in the dominant vector after the vector update process is completed as controlled nodes. In this way, it becomes possible to efficiently determine the controlled nodes that are effectively controlled by the dominant nodes using matrix operations.

[0141] Furthermore, the influence calculation unit 115 performs a process on each component of the N-dimensional vector, which is the product of the dominant vector and the connection matrix after the vector update process is completed. For components whose value is greater than the first threshold, the value of the component whose value is less than or equal to the first threshold is maintained, and the value of each component of the N-dimensional vector after processing is determined as the degree of influence of the dominant node on each of the multiple nodes. In this way, it becomes possible to determine the degree of influence that takes into account the effective dominance relationship using matrix operations.

[0142] Specifically, after the vector update process is completed, in step S307, the influence calculation unit 115 obtains a row vector which is the product of the dominant vector v and the connection matrix, and applies the second function G to each component of the row vector. The second function G here is a function that satisfies equation (2) below. In equation (2) below, θ is the value corresponding to the first threshold described above. In other words, the second function G is a function that updates the values ​​of components whose values ​​are greater than the first threshold to the first value (for example, 1), and maintains the values ​​of components whose values ​​are less than or equal to the first threshold.

[0143]

number

[0144] As shown in Figure 21A, the dominant vector v=(1,0,1,1,1,1,0) at the end of the vector update process is multiplied by the connection matrix. The resulting row vector is (1,0,1.6,1.9,1.6,1.6,0.3). Furthermore, the result of applying the second function G to this row vector is o(i)=(1,0,1,1,1,1,0.3).

[0145] Here, each component of the vector o(i) represents the degree of influence that node i, the dominant node, exerts on other nodes. For example, since the values ​​of the 3rd to 6th components of o(i) are 1, the degree of influence of node 1 on nodes 3 to 6 is 1 (maximum value), indicating that nodes 3 to 6 are controlled nodes. Also, since the value of the 7th component of o(i) is 0.3, the degree of influence of node 1 on node 7 is 0.3. This reflects the connection relationship in the area enclosed by the dashed line in Figure 21B. That is, since node 7 is the 6th node, whose directly above it is a controlled node, it is possible that node 1 will exercise voting rights equivalent to 30% through the 6th node. According to the method of this embodiment, as a result of the action of the second function G, it becomes possible to obtain an influence value that takes into account the dominance relationship.

[0146] In step S308, the controlled node identification unit 114 determines whether or not all candidate nodes have been selected as controlling nodes. As mentioned above, there are two candidate nodes here, node 1 and node 5, and node 5 is unprocessed. If there are unprocessed candidate nodes remaining (step S308: No), in step S309, the controlled node identification unit 114 updates the value of i to the value of the unprocessed candidate node. In this case, i is updated to 5. After that, the process returns to step S301, and the above-described process is executed for the new controlling node.

[0147] Figures 22A to 24B illustrate the processing flow when node 5 is the dominant node. In this case, the initial value of the dominant vector v is a vector e5 where the value of the 5th component is 1 and the values ​​of the other components are 0. The connection matrix is ​​the same as in the example described above.

[0148] In the first step, as shown in Figure 22A, e5=(0,0,0,0,1,0,0) is multiplied by the connection matrix (step S303). The result of the multiplication, v', is v'=(0,0,0,0,1,0.6,0). By applying the first function F to the vector v', v''=(0,0,0,0,1,1,0) is obtained (step S304). Since the dominating vector v(e5) and the vector v'' are not equal (step S305: No), in step S306, the controlled node identification unit 114 updates the dominating vector v with the vector v'', and then returns to step S303 to execute the vector update process again.

[0149] As a result of the second step, the dominance vector v is updated to v=(0,0,0,0,1,1,0). This represents the state in which the dominance by node 5 has propagated up to node 6, which corresponds to the sixth component of the dominance vector v, as shown in Figure 22B.

[0150] In the second step, as shown in Figure 23A, v=(0,0,0,0,1,1,0) is multiplied by the connection matrix (step S303). The result of the multiplication, v', is v'=(0,0,0,0,1,1.6,0.3). Furthermore, by applying the first function F to the vector v', v''=(0,0,0,0,1,1,0) is obtained (step S304). Since the dominating vector v and the vector v'' are equal (step S305: Yes), the controlled node identification unit 114 terminates the vector update process.

[0151] As a result of the second step, the dominance vector v is updated to the state v=(0,0,0,0,1,1,0). This represents the state where dominance by node 5 has propagated to node 6, which corresponds to the 6th component of the dominance vector v, as shown in Figure 23B, and no further propagation has occurred.

[0152] After the vector update process is completed, in step S307, as shown in Figure 24A, the dominant vector v=(0,0,0,0,1,1,0) at the end of the vector update process is multiplied by the connection matrix. The resulting row vector is (0,0,0,0,1,1.6,0.3). Furthermore, the result of applying the second function G to this row vector is o(i)=(0,0,0,0,1,1,0.3).

[0153] Since the value of the sixth component of o(i) is 1, the influence of node 5 on node 6 is 1 (maximum value), indicating that node 6 is a controlled node. Also, since the value of the seventh component of o(i) is 0.3, the influence of node 5 on node 7 is 0.3. This reflects the connection relationship in the area enclosed by the dashed line in Figure 24B.

[0154] As a result, processing for all candidate nodes is complete (Step S308: Yes), and the matrix operation shown in Figure 16 is terminated.

[0155] In this example, it is known that node 5 is a controlled node effectively controlled by node 1 after processing node 1. Therefore, the control relationship (Figure 23B) and influence (Figure 24B) by node 5 are included in the processing results for node 1, and processing for node 5 can be omitted.

[0156] 2.7 Specific Examples Figures 25 and 26 illustrate the processing results when node X is the dominant node and node Y is the target node in a predetermined network. In the example shown in Figure 25, if the index described above in this embodiment is used as the degree of influence of node X on node Y, the value is 0.49. In other words, node X does not effectively control node Y. On the other hand, if the Power Index is used as the degree of influence, the value of the degree of influence of node X on node Y is 0.526. In other words, in the network shown in Figure 25, the Power Index value may become excessively large, but in this embodiment, the degree of influence can be set to a value that appropriately takes into account the dominance relationship.

[0157] In the example shown in Figure 26, if the index described above in this embodiment is used to measure the influence of node X on node Y, the value is 1. In other words, node X effectively controls node Y. On the other hand, if the Power Index is used as the influence measure, the value of the influence of node X on node Y is 0.428. In other words, in the network shown in Figure 26, the Power Index value may be excessively small, but in this embodiment, the influence measure can be set to a value that appropriately considers the dominance relationship. As described above, the method of this embodiment makes it possible to accurately determine the influence measure that represents the dominance relationship.

[0158] Although this embodiment has been described in detail above, it will be readily apparent to those skilled in the art that many modifications are possible without substantially departing from the novel aspects and effects of this embodiment. Therefore, all such modifications are included within the scope of this disclosure. For example, any term that appears at least once in the specification or drawings together with a broader or synonymous term may be replaced with that different term anywhere in the specification or drawings. Furthermore, all combinations of this embodiment and its modifications are also included within the scope of this disclosure. In addition, the configuration and operation of server systems, terminal devices, information processing systems, etc., are not limited to those described in this embodiment, and various modifications are possible. [Explanation of symbols]

[0159] 10... Information processing system, 100... Server system, 110... Processing unit, 111... Entity network acquisition unit, 112... Target node acquisition unit, 113... Subnetwork extraction unit, 114... Controlled node identification unit, 115... Influence calculation unit, 116... Display processing unit, 120... Storage unit, 121... Entity network, 130... Communication unit, 200, 200-1, 200-2... Terminal device, 210... Processing unit, 220... Storage unit, 230... Communication unit, 240... Display unit, 250... Operation unit

Claims

1. An entity network acquisition unit acquires an entity network that represents the mutual investment relationships of multiple nodes corresponding to multiple entities by connecting the lower-level nodes that are being invested in and the upper-level nodes that are investing in them with edges assigned investment ratios. A controlled node identification unit identifies controlled nodes that are effectively controlled by a controlling node among the plurality of nodes based on the entity network, Includes, The controlled node identification unit is: When the controlling node and the already identified controlled nodes are considered as the first node group, An information processing system that identifies a controlled node by performing an update process to add to the controlled node a node that is directly connected to a second node group, which is one or more nodes from the plurality of nodes included in the first node group, and whose sum of the investment ratios assigned to the edges between each node of the second node group is greater than a first threshold.

2. In claim 1, The controlled node identification unit is: The aforementioned controlled node is initialized with an empty set, After the initialization process, a first update process is performed to add to the controlled nodes any of the plurality of nodes that are directly connected to the controlling node and whose investment ratio assigned to the edge between them and the controlling node is greater than the first threshold, An information processing system that identifies the controlled node by repeatedly executing the second update process, which is the update process, after the first update process.

3. In claim 2, The controlled node identification unit is: An information processing system that terminates the second update process if the controlled node does not change even after performing the second update process.

4. In any one of claims 1 to 3, The system further includes an influence calculation unit that determines the degree of influence that the controlling node has on other nodes. The aforementioned impact calculation unit, An information processing system that, after the process of identifying the controlled node is completed, determines the degree of influence of the controlling node on the target node based on the investment ratio assigned to the edge directly connecting the node included in the first node group and the target node, if the target node among the plurality of nodes is not included in the first node group.

5. In claim 4, The aforementioned impact calculation unit, An information processing system that sets the degree of influence of the controlling node over the controlled nodes to the maximum value.

6. In claim 4, The aforementioned impact calculation unit, An information processing system that identifies one or more edges that directly connect the target node to any of the nodes included in the first node group, and determines the influence of the controlling node on the target node as the sum of the investment ratios assigned to the identified one or more edges.

7. In claim 4, The aforementioned impact calculation unit, An information processing system that, if the target node is not directly connected to any of the nodes included in the first node group, sets the influence of the controlling node on the target node to the minimum value.

8. In any one of claims 1 to 3, If any of the aforementioned plurality of nodes is determined to be a target node, the system further includes a subnetwork extraction unit that extracts subnetworks consisting of higher-level nodes directly or indirectly connected to the target node, The controlled node identification unit is: An information processing system that selects, as the dominant node, a node among the nodes other than the target node included in the subnetwork, whose investment ratio to at least one other node is greater than the first threshold.

9. In any one of claims 1 to 3, An information processing system further including a display processing unit that displays the controlling node and the controlled node in the entity network in a manner that allows them to be identified from other nodes.

10. In claim 1, The controlled node identification unit is: If the entity network includes a subnetwork that is directly or indirectly connected to the dominant node, and contains the first to the Nth node (where N is an integer of 2 or more), An N x N connection matrix in which the value of the element in row i (where i is an integer between 1 and N) and column i is 1, and the value of the element in row i and column j (where j is an integer between 1 and N, and different from i) represents the investment ratio of the i-th node to the j-th node, We find a dominant vector which is an N-dimensional vector in which the component corresponding to the dominant node is 1 and the components corresponding to other nodes are 0. An information processing system that determines the controlled node by multiplying the controlling vector and the connection matrix.

11. In claim 10, The controlled node identification unit is: For each component of the N-dimensional vector, which is the product of the governing vector and the connection matrix, the values ​​of components whose values ​​are greater than the first threshold are updated to a first value, and the values ​​of components whose values ​​are less than or equal to the first threshold are updated to a second value smaller than the first value. Then, a vector update process is performed to update the governing vector using the N-dimensional vector after the processing. An information processing system that identifies the node corresponding to the component whose value is the first value in the dominant vector after the vector update process has been completed as the controlled node.

12. In claim 11, An information processing system further comprising an influence calculation unit that, after the completion of the vector update process, updates the values ​​of components whose values ​​are greater than the first threshold to 1 and maintains the values ​​of components whose values ​​are less than or equal to the first threshold, and determines the values ​​of each component of the N-dimensional vector after processing as the degree of influence of the dominating node on each of the multiple nodes.

13. Information processing device, An entity network is obtained that represents the mutual investment relationships of multiple nodes corresponding to multiple entities, by connecting the lower-level nodes that are being invested in and the upper-level nodes that are investing in them with edges that are assigned investment ratios. Based on the entity network, a controlled node is identified, which is a node that is effectively controlled by the controlling node among the multiple nodes. In identifying the controlled node, When the controlling node and the already identified controlled nodes are considered as the first node group, the controlled node is identified by performing an update process to add to the controlled node one or more nodes from the plurality of nodes that are directly connected to the second node group, which is one or more nodes included in the first node group, and whose sum of the investment ratios assigned to the edges between each node of the second node group is greater than the first threshold. An information processing method that performs processing.

Images