Information processing device, information processing method, and information processing program

Abstract

To estimate a transaction route for providing a specific transaction object.SOLUTION: An information processing device includes an operation unit and a storage unit. The operation unit includes: a connection source setting unit 1101 that sets a connection source company and a connection source keyword; a connection destination candidate selection unit 1102 that selects a connection destination candidate company and a connection destination candidate keyword from the connection source company and the connection source keyword; a type determination unit 1103 that determines a type of the transaction object indicated by the connection destination candidate; a category determination unit 1104 that determines a category indicating a transaction stage of the transaction object indicated by the connection destination candidate keyword; a necessity determination unit 1105 that determines the necessity between the transaction object indicated by the connection destination candidate keyword and a transaction object indicated by the connection source keyword; a route connection determination unit 1106 that determines whether to perform connection of a transaction route from a type of a transaction object indicated by the connection source keyword and from a determination result obtained by each determination unit; and a route connection unit 1107 that when it is determined that connection of the transaction route is to be performed, performs connection to the connection source company and the connection source keyword.SELECTED DRAWING: Figure 2

Description

[Technical Field]

[0001] The present disclosure relates to an information processing device, an information processing method, and an information processing program that estimate a transaction path. [Background technology]

[0002] Conventionally, techniques for estimating a procurement route have been known. In this conventional technology, for example, a supplier list and a customer list are used as shown on the left side of Fig. 21. On the left side of Fig. 21, the upper row shows the supplier list and the lower row shows the customer list.

[0003] The supplier list links and lists supplier companies, companies that do business with the supplier companies, and keywords that represent goods traded between these companies (traded goods). The customer list links and lists the selling company, the companies that do business with the selling company (customers), and keywords that represent the goods traded between these companies (traded goods).

[0004] In the prior art, a procurement route is estimated based on such a supplier list and a customer list.

[0005] In this case, first, based on the supplier list and the customer list, a transaction network such as that shown in the center of Figure 21 is obtained. In the example of Figure 21, a transaction network is obtained for "Company A," in which "apples" are purchased from "Company A," "fabric" is purchased from "Company I," "lunch boxes" are sold to "Company Alpha," and "clothes" are sold to "Company Beta."

[0006] Next, based on this transaction network, a procurement route is created, for example, as shown on the right side of Figure 21. In the example of Figure 21, for "Company A," a procurement route has been created in which "apples" are purchased from "Company A" and "lunch boxes" are sold to "Company Alpha," and a procurement route in which "fabric" is purchased from "Company I" and "clothes" are sold to "Company Beta." In other words, as shown by the gray arrows on the right side of Figure 21, the procurement route of purchasing "apples" from "Company A" and selling "clothes" to "Company β", and the procurement route of purchasing "fabric" from "Company I" and selling "lunch boxes" to "Company α" are not created and are considered to be non-existent routes.

[0007] Furthermore, a known technology for estimating such procurement routes is disclosed in, for example, Patent Document 1. In this conventional technology, a supply chain network is acquired and a graph embedding method is applied to the supply chain network to calculate feature quantities between nodes. In this graph embedding method, for example, as shown in Figure 22, it is found that the peripheral information of link (1) is closer to the peripheral information of link (3) than to the peripheral information of link (4), and as a result, the route connecting link (1) and link (3) is selected as the procurement route. [Prior art documents] [Patent documents]

[0008] [Patent Document 1] Japanese Patent Publication No. 2023-050018 Summary of the Invention [Problem to be solved by the invention]

[0009] However, in the conventional technology, a transaction route (procurement route or sales route) is estimated while ignoring the relationship between the transaction objects in purchasing or selling. Therefore, the conventional technology cannot estimate a transaction route for providing a specific transaction object.

[0010] The present disclosure has been made to solve the above-mentioned problems, and aims to provide an information processing device that can estimate a transaction path for providing a specific transaction object. [Means for solving the problem]

[0011] The information processing device according to the present disclosure includes a connection source setting unit that sets keywords representing a company that is a starting point and a transaction target at the company as a connection source company and a connection source keyword; a connection destination candidate selection unit that selects keywords representing a company that is a connection destination candidate and a transaction target at the company as a connection destination candidate company and a connection destination candidate keyword based on the connection source company and the connection source keyword set by the connection source setting unit; a type determination unit that determines, using a large-scale language model, a type of transaction target represented by the connection destination candidate keyword selected by the connection destination candidate selection unit; a category determination unit that determines, using a large-scale language model, a category representing a transaction stage of the transaction target represented by the connection destination candidate keyword selected by the connection destination candidate selection unit; The system is characterized by comprising: a necessity determination unit that determines, using a large-scale language model, the necessity of a transaction object represented by the candidate destination keyword and the transaction object represented by the source keyword set by the source setting unit, a route connection determination unit that determines whether to connect a transaction route based on the type of transaction object represented by the source keyword set by the source setting unit, the determination result by the type determination unit, the determination result by the category determination unit, and the determination result by the necessity determination unit, and a route connection unit that, when the route connection determination unit determines that a transaction route should be connected, sets the candidate destination company and candidate destination keyword selected by the candidate destination selection unit as the destination company and destination keyword, and connects the destination company and the destination keyword to the source company and source keyword set by the source setting unit. [Effects of the Invention]

[0012] According to the present disclosure, with the above configuration, it is possible to estimate a transaction path for providing a specific transaction object. [Brief explanation of the drawings]

[0013] [Figure 1] 1 is a block diagram showing an example of the configuration of an information processing device according to a first embodiment. [Figure 2] FIG. 2 is a block diagram showing an example of the configuration of a calculation unit according to the first embodiment. [Figure 3] 1 is a block diagram showing an example of a hardware configuration of an information processing device according to a first embodiment. [Figure 4] 4 is a flowchart showing an example of the operation of the information processing device according to the first embodiment. [Figure 5] 5A to 5D are diagrams illustrating an example of the operation of the information processing device according to the first embodiment. [Figure 6] FIG. 4 is a diagram illustrating a specific example of the operation of the information processing device according to the first embodiment. [Figure 7] FIG. 10 is a diagram showing a specific example of the operation of the information processing device according to the first embodiment (when estimating the procurement route of HDDs in company A). [Figure 8] FIG. 10 is a diagram showing a specific example of the operation of the information processing device according to the first embodiment (in the case where a procurement route of an SDD in company A is estimated). [Figure 9] FIG. 10 is a block diagram showing an example of the configuration of a calculation unit according to a second embodiment. [Figure 10] 10 is a flowchart showing an example of operation of the information processing device according to the second embodiment; [Figure 11] FIG. 10 is a diagram illustrating an example of the operation of the information processing device according to the second embodiment. [Figure 12] FIG. 11 is a block diagram showing an example of the configuration of a calculation unit according to a third embodiment. [Figure 13] FIG. 11 is a block diagram showing an example of the configuration of a calculation unit according to a fourth embodiment. [Figure 14] FIG. 13 is a block diagram showing an example of the configuration of a calculation unit in the fifth embodiment. [Figure 15] FIG. 20 is a block diagram showing an example of the configuration of a calculation unit in the sixth embodiment. [Figure 16] 13 is a diagram showing an example of a category determination table and a necessity determination table used in the calculation unit in the sixth embodiment. FIG. [Figure 17] FIG. 20 is a diagram illustrating an example of the operation of the information processing device according to the sixth embodiment. [Figure 18]FIG. 20 is a block diagram showing another example of the configuration of the calculation unit in the sixth embodiment. [Figure 19] FIG. 20 is a block diagram showing an example of the configuration of a calculation unit in the seventh embodiment. [Figure 20] FIG. 20 is a diagram showing an example of a substitutability determination table used in the calculation unit in the seventh embodiment. [Figure 21] FIG. 10 is a diagram for explaining an outline of estimation of a procurement route. [Figure 22] FIG. 1 is a diagram for explaining a conventional estimation of a procurement route. DETAILED DESCRIPTION OF THE INVENTION

[0014] Hereinafter, embodiments will be described in detail with reference to the drawings. Embodiment 1 FIG. 1 is a block diagram showing an example of the configuration of an information processing device 1 according to the first embodiment. The information processing device 1 estimates a transaction route. The transaction route is a procurement route or a sales route. The information processing device 1 includes a calculation unit 11 and a storage unit 12, as shown in FIG. The information processing device 1 may be physically configured from one terminal or multiple terminals.

[0015] The calculation unit 11 controls the overall operation of the information processing device 1. The calculation unit 11 executes applications for the information processing device 1 to realize various functions.

[0016] The storage unit 12 stores applications for the information processing device 1 and stores data used in the calculation process of the calculation unit 11. The storage unit 12 may be provided outside the information processing device 1 as long as it is accessible from the information processing device 1.

[0017] 2, the calculation unit 11 includes a connection source setting unit 1101, a connection destination candidate selection unit 1102, a type determination unit 1103, a category determination unit 1104, a necessity determination unit 1105, a route connection determination unit 1106, and a route connection unit 1107. When the calculation unit 11 executes the information processing device 1 application, the functions of the connection source setting unit 1101, the connection destination candidate selection unit 1102, the type determination unit 1103, the category determination unit 1104, the necessity determination unit 1105, the route connection determination unit 1106, and the route connection unit 1107 are realized.

[0018] The connection source setting unit 1101 sets keywords representing the starting company and the trading object of that company as connection source keywords for the connection source company. Keywords representing the trading object include keywords that directly represent the trading object, such as product names or service names, as well as keywords that indirectly represent the trading object, such as industry names (for example, agriculture or the semiconductor industry).

[0019] In addition, if the category determined by the category determination unit 1104 for the destination company and destination keyword set by the route connection unit 1107 is an intermediate transaction stage, the connection source setting unit 1101 updates the setting of the destination company and the destination keyword as the source company and the source keyword.

[0020] The connection destination candidate selection unit 1102 selects connection destination candidate companies and keywords representing the business partners of the companies as connection destination candidate companies and connection destination candidate keywords based on the connection source companies and connection source keywords set by the connection source setting unit 1101. The connection destination candidate selection unit 1102 selects one or more connection destination candidate companies and connection destination candidate keywords.

[0021] Here, for example, the connection destination candidate selection unit 1102 may select keywords representing connection destination candidate companies and transaction targets of the companies from the transaction relationship data based on the connecting source company and the connecting source keyword as connection destination candidate companies and connection destination candidate keywords. The transaction relationship data is data indicating the relationship between multiple companies and keywords representing transaction targets between the companies. In other words, when using transaction relationship data that lists connection destination candidate companies that are directly connected to the connecting source company, the connection destination candidate selection unit 1102 can directly select connection destination candidate companies and connection destination candidate keywords from this transaction relationship data. Furthermore, for example, the connection destination candidate selection unit 1102 may select keywords representing connection destination candidate companies and their business partners as connection destination candidate companies and connection destination candidate keywords by performing a breadth-first search on the transaction relationship network based on the above-mentioned source company and the above-mentioned source keyword. A transaction relationship network is a network in which nodes are companies and edges represent transaction relationships. Breadth-first search is an algorithm for finding the connection part (a set of companies that have a route connecting from the source company on the network) to which a specific company (here, the source company) belongs in the transaction relationship network. Note that if it is desired to search only for companies directly connected to the source company, it is sufficient to refer to the transaction relationship list of the source company. Furthermore, the search method is not limited to breadth-first search, and depth-first search may also be used.

[0022] When the transaction route to be estimated is a procurement route, the connection destination candidate selection unit 1102 performs the above selection in the upstream direction of the transaction relationship. Furthermore, when the transaction route to be estimated is a sales route, the connection destination candidate selection unit 1102 performs the above selection in the downstream direction of the transaction relationship.

[0023] The type determination unit 1103 determines the type of transaction object represented by the connection destination candidate keyword selected by the connection destination candidate selection unit 1102 using a large-scale language model. In this case, the type determination unit 1103 uses a large-scale language model such as ChatGPT to input information representing the connection destination candidate keyword and inquire about the type, thereby acquiring information representing the type. The type of transaction object represented by the connection destination candidate keyword determined by this type determination unit 1103 can be goods, services, or others. When the connection destination candidate selection unit 1102 selects a plurality of connection destination candidate keywords, the type determination unit 1103 makes the above determination for each of the connection destination candidate keywords.

[0024] The category determination unit 1104 determines, using a large-scale language model, a category indicating the transaction stage of the transaction object represented by the connection destination candidate keyword selected by the connection destination candidate selection unit 1102. In this case, the category determination unit 1104 acquires information indicating the category by inputting information indicating the connection destination candidate keyword using a large-scale language model such as ChatGPT and querying the category. The categories of the transaction object represented by the connection destination candidate keyword determined by the category determination unit 1104 include the final transaction stage, the intermediate transaction stage, and the initial transaction stage. In particular, when the type of transaction object is a good, a good in the final transaction stage may be referred to as a final good, a good in the intermediate transaction stage may be referred to as an intermediate good, and a good in the initial transaction stage may be referred to as a raw material. Furthermore, when the type of transaction object is a service, a service in the final transaction stage may be referred to as a final service, a service in the intermediate transaction stage may be referred to as an intermediate service, and a service in the initial transaction stage may be referred to as an initial service. When a plurality of connection destination candidate keywords are selected by the connection destination candidate selection unit 1102, the category determination unit 1104 makes the above determination for each connection destination candidate keyword.

[0025] The necessity determination unit 1105 determines, with respect to the candidate destination keyword selected by the candidate destination selection unit 1102, the necessity between the transaction object represented by the candidate destination keyword and the transaction object represented by the source keyword set by the source setting unit 1101, using a large-scale language model. In this case, the necessity determination unit 1105 uses a large-scale language model such as ChatGPT, inputs information indicating the source keyword and the candidate destination keyword, and inquires about the necessity, thereby acquiring information indicating the necessity. When a plurality of connection destination candidate keywords are selected by the connection destination candidate selection unit 1102, the necessity determination unit 1105 makes the above determination for each of the connection destination candidate keywords.

[0026] Here, if the transaction route to be estimated is a procurement route, the necessity determination unit 1105 determines whether the transaction object represented by the connection destination candidate keyword is necessary in providing the transaction object represented by the connection source keyword. In addition, if the transaction route to be estimated is a sales route, the necessity determination unit 1105 determines whether the transaction object represented by the connection source keyword is necessary in providing the transaction object represented by the connection destination candidate keyword.

[0027] Furthermore, when the settings are updated by the connection source setting unit 1101, the necessity determination unit 1105 determines, for the connection destination candidate keywords selected by the connection destination candidate selection unit 1102, the necessity between the transaction object represented by the connection destination candidate keyword and the transaction object represented by the connection source keyword updated by the connection source setting unit 1101, and the necessity between the transaction object represented by the keyword at the starting point company. At this time, the necessity determination unit 1105 acquires information indicating the above-mentioned necessity by inputting information indicating the keyword at the starting point company, the connection source keyword, and the connection destination candidate keyword, using a large-scale language model such as ChatGPT, for example. When a plurality of connection destination candidate keywords are selected by the connection destination candidate selection unit 1102, the necessity determination unit 1105 makes the above determination for each of the connection destination candidate keywords.

[0028] Here, if the transaction route to be estimated is a procurement route, the necessity determination unit 1105 determines whether the transaction object represented by the candidate connection keyword is necessary for each of the provision of the transaction object represented by the source keyword and the provision of the transaction object represented by the keyword at the starting company. In addition, if the transaction route to be estimated is a sales route, the necessity determination unit 1105 determines whether or not the transaction object represented by the connection source keyword and the transaction object represented by the keyword at the starting point company are necessary in providing the transaction object represented by the connection destination candidate keyword.

[0029] The route connection determination unit 1106 determines whether or not to connect a transaction route based on the type of transaction object represented by the connection source keyword set by the connection source setting unit 1101, the determination result by the type determination unit 1103, the determination result by the category determination unit 1104, and the determination result by the necessity determination unit 1105. When a plurality of connection destination candidate keywords are selected by the connection destination candidate selection unit 1102, the route connection determination unit 1106 makes the above determination for each of the connection destination candidate keywords.

[0030] Here, if the type of transaction object represented by the connection source keyword is a good, the route connection determination unit 1106 determines to connect the transaction route if the type of transaction object determined by the type determination unit 1103 is a good, the category of the transaction object determined by the category determination unit 1104 is not the final transaction stage, i.e., the intermediate transaction stage or the initial transaction stage, and the necessity determination unit 1105 determines that it is necessary. In addition, if the type of transaction object represented by the connection source keyword is a service, the route connection determination unit 1106 determines to connect the transaction route if the type of transaction object determined by the type determination unit 1103 is a service, the category of the transaction object determined by the category determination unit 1104 is not the final transaction stage, i.e., the intermediate transaction stage or the initial transaction stage, and the necessity determination unit 1105 determines that it is necessary. On the other hand, in cases other than those mentioned above, the route connection determining unit 1106 determines not to connect the transaction route.

[0031] Furthermore, when the setting is updated by the connection source setting unit 1101, the route connection determination unit 1106 determines to connect the transaction route if the type of transaction object represented by the connection source keyword is a good, the type of transaction object determined by the type determination unit 1103 is a good, the category of the transaction object determined by the category determination unit 1104 is not the final transaction stage, i.e., the intermediate transaction stage or the initial transaction stage, and the necessity determination unit 1105 determines that all are necessary. In addition, if the type of transaction object represented by the connection source keyword is a service, the route connection determination unit 1106 determines to connect the transaction route if the type of transaction object determined by the type determination unit 1103 is a service, the category of the transaction object determined by the category determination unit 1104 is not the final transaction stage, i.e., the intermediate transaction stage or the initial transaction stage, and the necessity determination unit 1105 determines that all are necessary. On the other hand, in cases other than those mentioned above, the route connection determining unit 1106 determines not to connect the transaction route.

[0032] When the route connection determination unit 1106 determines that a transaction route should be connected, the route connection unit 1107 sets the connection candidate company and connection candidate keyword selected by the connection candidate selection unit 1102 as the connection destination company and connection destination keyword, and connects the connection destination company and connection destination keyword to the connection source company and connection source keyword set by the connection source setting unit 1101.

[0033] Here, if the transaction route to be estimated is a procurement route, the route connection unit 1107 connects the connection destination company to the connection source company using a connection destination keyword. Furthermore, if the transaction route to be estimated is a sales route, the route connection unit 1107 connects the connection destination company to the connection source company using a connection source keyword.

[0034] The information processing device 1 may display information indicating the transaction route obtained by the route connection unit 1107 on an external display device (not shown). This allows the user to visually recognize the transaction route obtained by the information processing device 1.

[0035] FIG. 3 is a block diagram showing an example of a hardware configuration that realizes the functions of the information processing device 1. As shown in FIG.

[0036] For example, the functions of the information processing device 1 are realized by a calculation unit 11. The calculation unit 11 has, as its hardware configuration, an input / output interface 101, a processor 102, and a memory 103. The functions of a connection source setting unit 1101, a connection destination candidate selection unit 1102, a type determination unit 1103, a category determination unit 1104, a necessity determination unit 1105, a route connection determination unit 1106, and a route connection unit 1107 provided in the calculation unit 11 are realized by these hardware configurations.

[0037] The processor 102 controls the reading and writing of data from and to the storage unit 12 via the input / output interface 101 .

[0038] An information processing program (information processing application) for realizing each function of the connection source setting unit 1101, the connection destination candidate selection unit 1102, the type determination unit 1103, the category determination unit 1104, the necessity determination unit 1105, the route connection determination unit 1106, and the route connection unit 1107 is stored in the memory 103. The memory 103 is, for example, a semiconductor memory provided in a computer that functions as the information processing device 1. The processor 102 executes a program loaded from the storage unit 12 to the memory 103 via the input / output interface 101. In this way, the processor 102 realizes each function of the connection source setting unit 1101, the connection destination candidate selection unit 1102, the type determination unit 1103, the category determination unit 1104, the necessity determination unit 1105, the route connection determination unit 1106, and the route connection unit 1107.

[0039] Next, an example of the operation of the information processing device 1 according to the first embodiment will be described with reference to FIG.

[0040] In an example of operation of the information processing device 1 according to embodiment 1, as shown in FIG. 4, first, the connection source setting unit 1101 sets keywords representing the starting company and the trading target of that company as connection source keywords for the connection source company (step ST101).

[0041] Next, the connection destination candidate selection unit 1102 selects, based on the connection source company and connection source keyword set by the connection source setting unit 1101, keywords representing the connection destination candidate company and the business partner of the company as connection destination candidate company and connection destination candidate keyword (step ST102). The connection destination candidate selection unit 1102 selects one or more connection destination candidate companies and connection destination candidate keywords.

[0042] When the transaction route to be estimated is a procurement route, the connection destination candidate selection unit 1102 performs the above selection in the upstream direction of the transaction relationship. Furthermore, when the transaction route to be estimated is a sales route, the connection destination candidate selection unit 1102 performs the above selection in the downstream direction of the transaction relationship.

[0043] Next, the type determination unit 1103 determines the type of transaction object represented by the connection destination candidate keyword selected by the connection destination candidate selection unit 1102 using a large-scale language model (step ST103). At this time, the type determination unit 1103 uses a large-scale language model such as ChatGPT, inputs information indicating the connection destination candidate keyword, and queries the type, thereby acquiring information indicating the type. When the connection destination candidate selection unit 1102 selects a plurality of connection destination candidate keywords, the type determination unit 1103 makes the above determination for each of the connection destination candidate keywords.

[0044] Furthermore, the category determination unit 1104 determines a category indicating a transaction stage of a transaction object represented by the connection destination candidate keyword selected by the connection destination candidate selection unit 1102 using a large-scale language model (step ST104). At this time, the category determination unit 1104 uses a large-scale language model such as ChatGPT, inputs information indicating the connection destination candidate keyword, and queries the category, thereby acquiring information indicating the category. When a plurality of connection destination candidate keywords are selected by the connection destination candidate selection unit 1102, the category determination unit 1104 makes the above determination for each connection destination candidate keyword.

[0045] Furthermore, the necessity determination unit 1105 determines, with respect to the connection destination candidate keyword selected by the connection destination candidate selection unit 1102, the necessity between the transaction object represented by the connection destination candidate keyword and the transaction object represented by the connection source keyword set by the connection source setting unit 1101, using a large-scale language model (step ST105). At this time, the necessity determination unit 1105 uses a large-scale language model such as ChatGPT, inputs information indicating the connection source keyword and the connection destination candidate keyword, and inquires about the necessity, thereby acquiring information indicating the necessity. When a plurality of connection destination candidate keywords are selected by the connection destination candidate selection unit 1102, the necessity determination unit 1105 makes the above determination for each of the connection destination candidate keywords.

[0046] Here, if the transaction route to be estimated is a procurement route, the necessity determination unit 1105 determines whether the transaction object represented by the connection destination candidate keyword is necessary in providing the transaction object represented by the connection source keyword. In addition, if the transaction route to be estimated is a sales route, the necessity determination unit 1105 determines whether the transaction object represented by the connection source keyword is necessary in providing the transaction object represented by the connection destination candidate keyword.

[0047] Next, the route connection determination unit 1106 determines whether or not to connect the transaction route based on the type of transaction object represented by the connection source keyword set by the connection source setting unit 1101, the determination result by the type determination unit 1103, the determination result by the category determination unit 1104, and the determination result by the necessity determination unit 1105 (step ST106). When a plurality of connection destination candidate keywords are selected by the connection destination candidate selection unit 1102, the route connection determination unit 1106 makes the above determination for each of the connection destination candidate keywords.

[0048] Here, if the type of transaction object represented by the connection source keyword is a good, the route connection determination unit 1106 determines to connect the transaction route if the type of transaction object determined by the type determination unit 1103 is a good, the category of the transaction object determined by the category determination unit 1104 is not the final transaction stage, i.e., the intermediate transaction stage or the initial transaction stage, and the necessity determination unit 1105 determines that it is necessary. In addition, if the type of transaction object represented by the connection source keyword is a service, the route connection determination unit 1106 determines to connect the transaction route if the type of transaction object determined by the type determination unit 1103 is a service, the category of the transaction object determined by the category determination unit 1104 is not the final transaction stage, i.e., the intermediate transaction stage or the initial transaction stage, and the necessity determination unit 1105 determines that it is necessary. On the other hand, in cases other than those mentioned above, the route connection determining unit 1106 determines not to connect the transaction route.

[0049] Next, if the route connection determination unit 1106 determines that a transaction route should be connected, the route connection unit 1107 sets the connection candidate company and connection candidate keyword selected by the connection candidate selection unit 1102 as the connection destination company and connection destination keyword, and connects the connection destination company and connection destination keyword to the connection source company and connection source keyword set by the connection source setting unit 1101 (step ST107).

[0050] Here, if the transaction route to be estimated is a procurement route, the route connection unit 1107 connects the connection destination company to the connection source company using a connection destination keyword. Furthermore, if the transaction route to be estimated is a sales route, the route connection unit 1107 connects the connection destination company to the connection source company using a connection source keyword.

[0051] Next, when the category determined by the category acquisition unit is the intermediate transaction stage for the destination company and destination keyword set by the route connection unit 1107, the source setting unit 1101 updates the setting of the destination company and destination keyword as the source company and source keyword (steps ST108 and ST109). After that, the sequence returns to step ST102, and the information processing device 1 repeats the above operations.

[0052] When the settings are updated by the connection source setting unit 1101, the necessity determination unit 1105 determines, for the connection destination candidate keywords selected by the connection destination candidate selection unit 1102, the necessity between the transaction object represented by the connection destination candidate keyword and the transaction object represented by the connection source keyword updated by the connection source setting unit 1101, and the necessity between the transaction object and the transaction object represented by the keyword at the starting point company. The necessity determination unit 1105 acquires information indicating the necessity by inputting information indicating the keyword at the starting point company, the connection source keyword, and the connection destination candidate keyword, for example, using a large-scale language model such as ChatGPT, and inquiring about the necessity. When a plurality of connection destination candidate keywords are selected by the connection destination candidate selection unit 1102, the necessity determination unit 1105 makes the above determination for each of the connection destination candidate keywords.

[0053] Here, if the transaction route to be estimated is a procurement route, the necessity determination unit 1105 determines whether the transaction object represented by the candidate connection keyword is necessary for each of the provision of the transaction object represented by the source keyword and the provision of the transaction object represented by the keyword at the starting company. In addition, if the transaction route to be estimated is a sales route, the necessity determination unit 1105 determines whether or not the transaction object represented by the connection source keyword and the transaction object represented by the keyword at the starting point company are necessary in providing the transaction object represented by the connection destination candidate keyword.

[0054] Furthermore, when the setting is updated by the connection source setting unit 1101, the route connection determination unit 1106 determines to connect the transaction route if the type of transaction object represented by the connection source keyword is a good, the type of transaction object determined by the type determination unit 1103 is a good, the category of the transaction object determined by the category determination unit 1104 is not the final transaction stage, i.e., the intermediate transaction stage or the initial transaction stage, and the necessity determination unit 1105 determines that all are necessary. In addition, if the type of transaction object represented by the connection source keyword is a service, the route connection determination unit 1106 determines to connect the transaction route if the type of transaction object determined by the type determination unit 1103 is a service, the category of the transaction object determined by the category determination unit 1104 is not the final transaction stage, i.e., the intermediate transaction stage or the initial transaction stage, and the necessity determination unit 1105 determines that all are necessary. On the other hand, in cases other than those mentioned above, the route connection determining unit 1106 determines not to connect the transaction route.

[0055] On the other hand, if the category of the connection destination keyword is not the intermediate transaction stage, the information processing device 1 ends the sequence.

[0056] The information processing device 1 may display on an external display device information indicating the transaction path obtained by the path connection unit 1107. This allows the user to visually recognize the transaction path obtained by the information processing device 1.

[0057] 5A to 5D are diagrams showing an example of the operation of the information processing device 1 according to embodiment 1. Here, a case where the information processing device 1 estimates a procurement route will be described as an example. In this example, first, as shown in FIG. 5A, the connection source setting unit 1101 sets keywords representing company A and asset (1) as the starting point as the connection source company and connection source keyword.

[0058] In this case, as shown in Fig. 5B, the connection destination candidate selection unit 1102 selects connection destination candidate companies and connection destination candidate keywords based on the connection source keywords representing company A and asset (1). In the example of Fig. 5B, the connection destination candidate selection unit 1102 selects keywords representing company B and transaction target (2) as connection destination candidate companies and connection destination candidate keywords.

[0059] Next, using a large-scale language model, the type determination unit 1103 determines the type of transaction object (2), the category determination unit 1104 determines the category of the transaction object (2), and the necessity determination unit 1105 determines whether the transaction object (2) is necessary in the provision (manufacturing process) of the good (1).

[0060] At this time, for example, the information processing device 1 inputs "You are a business relationship network analyst" as an initial input to the large-scale language model. Then, for example, the type determination unit 1103 inputs a first question to the large-scale language model, such as, "What type of transaction object (2) is? Goods, services, or other? Please answer in one word." Then, the type determination unit 1103 receives "goods," "services," or "other" as a first answer from the large-scale language model. Also, for example, the category determination unit 1104 inputs the following as a second question to the large-scale language model: "What is the category of the transaction object (2): raw materials, intermediate goods, or final goods? Please answer in one word. If it is not a physical product, please answer 'I don't really know.'" Then, the category determination unit 1104 receives "raw materials," "intermediate goods," "final goods," or "I don't really know" as a second answer from the large-scale language model. Furthermore, for example, the necessity determination unit 1105 inputs the following as a third question to the large-scale language model: "Is trading object (2) necessary in the manufacturing process of good (1)? Please answer 'yes' or 'no'. If you are not sure, please answer 'I don't know'." Then, the necessity determination unit 1105 receives "yes," "no," or "I don't know" as the third answer from the large-scale language model.

[0061] Next, the route connection determination unit 1106 determines to connect the procurement route when the type determination unit 1103 receives "goods" as the first answer, the category determination unit 1104 receives "intermediate goods" or "raw materials" as the second answer, and the necessity determination unit 1105 receives "yes" as the third answer. Next, the route connection unit 1107 connects company B to company A using a keyword representing the transaction object (2) (=goods (2)). On the other hand, in cases other than those mentioned above, the route connection determining unit 1106 determines not to connect the procurement routes.

[0062] FIG. 5C shows a case where the route connection unit 1107 connects company A with company B using a keyword representing good (2).

[0063] Next, the connection source setting unit 1101 determines whether the category of the good (2) is an intermediate good. If the connection source setting unit 1101 determines that the category is intermediate goods, it updates the setting to the connection destination keywords representing company B and goods (2) as the connection source company and connection source keywords. On the other hand, if the connection source setting unit 1101 determines that the category is not an intermediate good, that is, a raw material, the information processing device 1 ends the process.

[0064] Figure 5C shows the case where the category of goods (2) is intermediate goods.

[0065] In this case, as shown in Fig. 5C, the connection destination candidate selection unit 1102 selects connection destination candidate companies and connection destination candidate keywords based on the connection source keywords representing company B and asset (2). In the example of Fig. 5C, the connection destination candidate selection unit 1102 selects keywords representing company C and transaction target (3) as connection destination candidate companies and connection destination candidate keywords.

[0066] Next, using a large-scale language model, the type determination unit 1103 determines the type of transaction object (3), the category determination unit 1104 determines the category of the transaction object (3), and the necessity determination unit 1105 determines whether the transaction object (3) is necessary for each of the provision of good (1) (manufacturing process) and the provision of good (2) (manufacturing process).

[0067] At this time, for example, the information processing device 1 inputs "You are a business relationship network analyst" as an initial input to the large-scale language model. Then, for example, the type determination unit 1103 inputs a first question to the large-scale language model, such as, "What type of transaction object (3) is? Goods, services, or other? Please answer in one word." Then, the type determination unit 1103 receives "goods," "services," or "other" as a first answer from the large-scale language model. Also, for example, the category determination unit 1104 inputs the following as a second question to the large-scale language model: "Is the transaction object (3) a raw material, an intermediate good, or a final good? Please answer in one word. If it is not a physical product, please answer 'I don't know.'" Then, the category determination unit 1104 receives "raw material," "intermediate good," "final good," or "I don't know" as a second answer from the large-scale language model. Furthermore, for example, the necessity determination unit 1105 inputs the following as a third question to the large-scale language model: "Is trading object (3) necessary in the manufacturing process of good (1)? Please answer 'yes' or 'no'. If you are not sure, please answer 'I don't know'." Then, the necessity determination unit 1105 receives "yes," "no," or "I don't know" as the third answer from the large-scale language model. Furthermore, for example, the necessity determination unit 1105 inputs the following as a fourth question to the large-scale language model: "Is trading object (3) necessary in the manufacturing process of good (2)? Please answer 'yes' or 'no'. If you are not sure, please answer 'I don't know'." Then, the necessity determination unit 1105 receives "yes," "no," or "I don't know" as the fourth answer from the large-scale language model.

[0068] Next, the route connection determination unit 1106 determines to connect the procurement route when the type determination unit 1103 accepts "goods" as the first answer, the category determination unit 1104 accepts "intermediate goods" or "raw materials" as the second answer, the necessity determination unit 1105 accepts "yes" as the third answer, and the necessity determination unit 1105 accepts "yes" as the fourth answer. Next, the route connection unit 1107 connects company C to company B using a keyword representing the trading object (3) (=goods (3)). On the other hand, in cases other than those mentioned above, the route connection determining unit 1106 determines not to connect the procurement routes.

[0069] FIG. 5D shows a case where the route connection unit 1107 connects company C to company B using a keyword representing good (3).

[0070] Next, the connection source setting unit 1101 determines whether the category of the good (3) is an intermediate good. If the connection source setting unit 1101 determines that the category is intermediate goods, it updates the setting to the connection destination keyword representing company C and good (3) as the connection source company and connection source keyword. On the other hand, if the connection source setting unit 1101 determines that the category is not an intermediate good, that is, a raw material, the information processing device 1 ends the process.

[0071] Figure 5D shows the case where the category of goods (3) is intermediate goods.

[0072] In this case, as shown in Fig. 5D, the connection destination candidate selection unit 1102 selects connection destination candidate keywords for the connection destination candidate companies based on the connection source keywords representing company C and asset (3). In the example of Fig. 5D, the connection destination candidate selection unit 1102 selects keywords representing company D and trading target (4) as the connection destination candidate companies and connection destination candidate keywords.

[0073] Next, using a large-scale language model, the type determination unit 1103 determines the type of trading object (4), the category determination unit 1104 determines the category of trading object (4), and the necessity determination unit 1105 determines whether trading object (4) is necessary in each of the manufacturing processes of good (1) and good (3).

[0074] At this time, for example, the information processing device 1 inputs "You are a business relationship network analyst" as an initial input. Then, for example, the type determination unit 1103 inputs, as a first question to the large-scale language model, "Is the transaction object (4) a good, a service, or something else? Please answer in one word." Then, the type determination unit 1103 receives, as a first answer to the large-scale language model, "goods," "services," or "other." Also, for example, the category determination unit 1104 inputs the following as a second question to the large-scale language model: "Is the transaction object (4) a raw material, an intermediate good, or a final good? Please answer in one word. If it is not a physical product, please answer 'I don't know.'" Then, the category determination unit 1104 receives "raw material," "intermediate good," "final good," or "I don't know" as a second answer from the large-scale language model. Furthermore, for example, the necessity determination unit 1105 inputs the following as a third question to the large-scale language model: "Is trading object (4) necessary in the manufacturing process of good (1)? Please answer 'yes' or 'no'. If you are not sure, please answer 'I don't know'." Then, the necessity determination unit 1105 receives "yes," "no," or "I don't know" as the third answer from the large-scale language model. Furthermore, for example, the necessity determination unit 1105 inputs the following as a fourth question to the large-scale language model: "Is trading object (4) necessary in the manufacturing process of good (3)? Please answer 'yes' or 'no'. If you are not sure, please answer 'I don't know'." Then, the necessity determination unit 1105 receives "yes," "no," or "I don't know" as the fourth answer from the large-scale language model.

[0075] Next, the route connection determination unit 1106 determines to connect the procurement route when the type determination unit 1103 accepts "goods" as the first answer, the category determination unit 1104 accepts "intermediate goods" or "raw materials" as the second answer, the necessity determination unit 1105 accepts "yes" as the third answer, and the necessity determination unit 1105 accepts "yes" as the fourth answer. Next, the route connection unit 1107 connects company D to company C using a keyword representing the trading object (4) (=goods (4)). On the other hand, in cases other than those mentioned above, the route connection determining unit 1106 determines not to connect the procurement routes.

[0076] Thereafter, the information processing device 1 repeats the above process.

[0077] For example, assume that there is transaction relationship data such as that shown in Figure 6. This transaction relationship data shows transaction relationships between multiple companies, but does not allow for understanding the transaction route for a specific transaction target. In response to this, it is assumed that the information processing device 1 estimates the procurement route of HDDs in company A. In this case, for example, a procurement route such as that shown in FIG. On the other hand, suppose that the information processing device 1 estimates the procurement route of SSDs in, for example, company A. In this case, for example, the procurement route shown in FIG. In this way, the information processing device 1 according to the first embodiment can estimate a transaction path for a specific transaction object by determining the type, category, and necessity of the transaction object using prompt processing.

[0078] 7 and 8, the display of each transaction object is changed according to the category (the color of the connecting line is changed in the figures). That is, in the information processing device 1, the category determination unit 1104 can determine the category of each transaction object, and therefore the display of the transaction object can also be changed according to this category.

[0079] In the above example, the keywords representing the transaction object are mainly keywords that directly represent the transaction object, such as product names or service names. However, the transaction relationship data does not necessarily include keywords that directly represent the transaction object, and the keywords representing the transaction object may be keywords that indirectly represent the transaction object, such as industry names (e.g., agriculture or the semiconductor industry). On the other hand, the information processing device 1 according to embodiment 1 can also process cases where the keyword representing the transaction object is a keyword that indirectly represents the transaction object, and can estimate the transaction path for a specific transaction object.

[0080] In addition, in the above, when the settings are updated by the connection source setting unit 1101, the necessity determination unit 1105 determines the necessity between the transaction object represented by the connection destination candidate keyword and the transaction object represented by the connection source keyword updated by the connection source setting unit 1101, and the necessity between the transaction object represented by the keyword at the starting point company. However, without being limited to this, the necessity determination unit 1105 may also determine the necessity between the transaction object represented by the connection destination candidate keyword and the transaction object represented by the connection source keyword previously updated by the connection source setting unit 1101, in addition to the above. For example, in FIG. 5D, the necessity determination unit 1105 may determine whether or not the trading object (4) is necessary in each of the manufacturing process of the good (1), the manufacturing process of the good (2), and the manufacturing process of the good (3).

[0081] As described above, according to the first embodiment, the information processing device 1 includes a connection source setting unit 1101 that sets keywords representing a company that is a starting point and a transaction object at that company as a connection source company and a connection source keyword, a connection destination candidate selection unit 1102 that selects keywords representing a company that is a connection destination candidate and a transaction object at that company as a connection destination candidate company and a connection destination candidate keyword based on the connection source company and the connection source keyword set by the connection source setting unit 1101, a type determination unit 1103 that determines the type of transaction object represented by the connection destination candidate keyword selected by the connection destination candidate selection unit 1102 using a large-scale language model, a category determination unit 1104 that determines the category indicating the transaction stage of the transaction object represented by the connection destination candidate keyword selected by the connection destination candidate selection unit 1102 using a large-scale language model, and The information processing device 1 according to the first embodiment includes a necessity determination unit 1105 that determines, using a large-scale language model, the necessity of a transaction object represented by the connection destination candidate keyword and a transaction object represented by the connection source keyword set by the connection source setting unit 1101, a route connection determination unit 1106 that determines whether to connect a transaction path based on the type of transaction object represented by the connection source keyword set by the connection source setting unit 1101, the determination result by the type determination unit 1103, the determination result by the category determination unit 1104, and the determination result by the necessity determination unit 1105, and a route connection unit 1107 that sets the connection destination candidate company and connection destination candidate keyword selected by the connection destination candidate selection unit 1102 as a connection destination company and a connection destination keyword and connects the connection destination company and the connection destination keyword to the connection source company and connection source keyword set by the connection source setting unit 1101. This enables the information processing device 1 according to the first embodiment to estimate a transaction path for providing a specific transaction object.

[0082] Furthermore, according to the first embodiment, when the type of transaction object represented by the connection source keyword is goods, the route connection determination unit 1106 determines to connect a transaction route if the type of transaction object determined by the type determination unit 1103 is goods, the category of the transaction object determined by the category determination unit 1104 is not in the final transaction stage, and the necessity determination unit 1105 determines to need the transaction route; and when the type of transaction object represented by the connection source keyword is services, the type of transaction object determined by the type determination unit 1103 is services, the category of the transaction object determined by the category determination unit 1104 is not in the final transaction stage, and the necessity determination unit 1105 determines to connect a transaction route. This enables the information processing device 1 according to the first embodiment to estimate a transaction route for providing a specific transaction object.

[0083] Furthermore, according to the first embodiment, with respect to the destination company and destination keyword set by the path connection unit 1107, if the category determined by the category determination unit 1104 is the intermediate transaction stage, the connection source setting unit 1101 updates the setting of the destination company and destination keyword as the source company and source keyword, and when the setting is updated by the source setting unit 1101, the necessity determination unit 1105 determines, with respect to the destination candidate keyword selected by the destination candidate selection unit 1102, the necessity between the transaction object represented by the destination candidate keyword and the transaction object represented by the source keyword updated by the source setting unit 1101, and the necessity between the transaction object represented by the keyword of the starting point company. This enables the information processing device 1 according to the first embodiment to estimate a transaction path for providing a specific transaction object.

[0084] Furthermore, according to the first embodiment, when the type of transaction object represented by the connecting keyword is goods, the route connection determination unit 1106 determines to connect the transaction route if the type of transaction object determined by the type determination unit 1103 is goods, the category of the transaction object determined by the category determination unit 1104 is not in the final transaction stage, and the necessity determination unit 1105 determines that all of the transaction objects are required, and when the type of transaction object represented by the connecting keyword is services, the type determination unit 1103 determines to connect the transaction route if the category of the transaction object determined by the category determination unit 1104 is not in the final transaction stage, and the necessity determination unit 1105 determines that all of the transaction objects are required. This enables the information processing device 1 according to the first embodiment to estimate a transaction route for providing a specific transaction object.

[0085] Furthermore, according to the first embodiment, the information processing method executed by the computer includes the steps of: a source setting unit 1101 setting keywords representing a starting point company and a transaction object of the company as a source company and a source keyword; a destination candidate selection unit 1102 selecting keywords representing destination candidate companies and transaction objects of the company as destination candidate companies and destination candidate keywords based on the source company and source keyword set by the source setting unit 1101; a type determination unit 1103 determining, using a large-scale language model, the type of transaction object represented by the destination candidate keyword selected by the destination candidate selection unit 1102; a category determination unit 1104 determining, using a large-scale language model, the category representing the transaction stage of the transaction object represented by the destination candidate keyword selected by the destination candidate selection unit 1102; and a necessity determination unit 1105 The method includes the steps of: determining, using a large-scale language model, the necessity of a transaction object represented by the connection destination candidate keyword selected by the connection destination candidate selection unit 1102 and a transaction object represented by the connection source keyword set by the connection source setting unit 1101; determining, by the route connection determination unit 1106, whether to connect a transaction path based on the type of transaction object represented by the connection source keyword set by the connection source setting unit 1101, the determination result by the type determination unit 1103, the determination result by the category determination unit 1104, and the determination result by the necessity determination unit 1105; and, when the route connection determination unit 1106 determines that a transaction path should be connected, setting the connection destination candidate company and connection destination candidate keyword selected by the connection destination candidate selection unit 1102 as a connection destination company and connection destination keyword, and connecting the connection destination company and the connection destination keyword to the connection source company and connection source keyword set by the connection source setting unit 1101. This makes it possible for the information processing method according to the first embodiment to estimate a transaction path for providing a specific transaction object.

[0086] Furthermore, according to the first embodiment, the program causes the computer to function as the above-described information processing device 1. As a result, the program causes the computer to function as the information processing device 1 that is capable of estimating a transaction path for providing a specific transaction object.

[0087] Embodiment 2 Fig. 9 is a diagram showing an example of the configuration of the arithmetic unit 11 included in the information processing device 1 according to embodiment 2. In the example of the configuration of the arithmetic unit 11 included in the information processing device 1 according to embodiment 2 shown in Fig. 9, a condition acquisition unit 1108 and an abstraction level determination unit 1109 are added to the arithmetic unit 11 included in the information processing device 1 according to embodiment 1 shown in Fig. 2. The other example of the configuration of the arithmetic unit 11 included in the information processing device 1 according to embodiment 2 shown in Fig. 9 is the same as the example of the configuration of the arithmetic unit 11 included in the information processing device 1 according to embodiment 1 shown in Fig. 2, and the same reference numerals are used and only the different parts will be described.

[0088] The condition acquisition unit 1108 acquires a condition indicating the upper limit of the level of abstraction of the transaction route hierarchy and the connection destination candidate keywords.

[0089] The abstraction level determination unit 1109 determines the abstraction level of the connection destination candidate keyword selected by the connection destination candidate selection unit 1102 using a large-scale language model. In this case, the abstraction level determination unit 1109 uses a large-scale language model such as ChatGPT to acquire information in advance indicating the abstraction levels of various keywords that are assumed to be related to the connection destination candidate keyword, and determines the abstraction level of the connection destination candidate keyword selected by the connection destination candidate selection unit 1102 based on that information.

[0090] The route connection determination unit 1106 determines whether or not to connect a transaction route based on the type of transaction object represented by the connection source keyword set by the connection source setting unit 1101, the acquisition result by the condition acquisition unit 1108, the determination result by the type determination unit 1103, the determination result by the category determination unit 1104, the determination result by the necessity determination unit 1105, and the determination result by the abstraction level determination unit 1109.

[0091] In this case, in addition to the judgment shown in embodiment 1, the route connection judgment unit 1106 excludes from the connection candidates the candidate connection companies and candidate connection keywords selected by the candidate connection selection unit 1102 if the hierarchy of the transaction route to be connected is equal to or higher than the hierarchy indicated by the conditions acquired by the condition acquisition unit 1108, and if the abstraction level of the candidate connection keywords judged by the abstraction level judgment unit 1109 is equal to or higher than the abstraction level indicated by the conditions acquired by the condition acquisition unit 1108.

[0092] Next, an example of the operation of the information processing device 1 according to the second embodiment will be described with reference to FIG.

[0093] In an example of operation of the information processing device 1 according to the second embodiment, as shown in FIG. 10, the condition acquisition unit 1108 first acquires a condition indicating the hierarchy of the transaction route and the upper limit of the abstraction level of the connection destination candidate keywords (step ST201).

[0094] Furthermore, the connection source setting unit 1101 sets keywords representing the company that is the starting point and the business object of that company as connection source keywords for the connection source company (step ST202).

[0095] Next, the connection destination candidate selection unit 1102 selects, based on the connection source company and connection source keyword set by the connection source setting unit 1101, keywords representing the connection destination candidate company and the business partner of the company as connection destination candidate companies and connection destination candidate keywords (step ST203). The connection destination candidate selection unit 1102 selects one or more connection destination candidate companies and connection destination candidate keywords.

[0096] When the transaction route to be estimated is a procurement route, the connection destination candidate selection unit 1102 performs the above selection in the upstream direction of the transaction relationship. Furthermore, when the transaction route to be estimated is a sales route, the connection destination candidate selection unit 1102 performs the above selection in the downstream direction of the transaction relationship.

[0097] Next, the type determination unit 1103 determines the type of transaction object represented by the connection destination candidate keyword selected by the connection destination candidate selection unit 1102 using a large-scale language model (step ST204). At this time, the type determination unit 1103 uses a large-scale language model such as ChatGPT, inputs information indicating the connection destination candidate keyword, and inquires about the type, thereby acquiring information indicating the type. When the connection destination candidate selection unit 1102 selects a plurality of connection destination candidate keywords, the type determination unit 1103 makes the above determination for each of the connection destination candidate keywords.

[0098] Furthermore, the category determination unit 1104 determines a category indicating a transaction stage of a transaction object represented by the connection destination candidate keyword selected by the connection destination candidate selection unit 1102 using a large-scale language model (step ST205). At this time, the category determination unit 1104 uses a large-scale language model such as ChatGPT, inputs information indicating the connection destination candidate keyword, and queries the category, thereby acquiring information indicating the category. When a plurality of connection destination candidate keywords are selected by the connection destination candidate selection unit 1102, the category determination unit 1104 makes the above determination for each connection destination candidate keyword.

[0099] Furthermore, the necessity determination unit 1105 determines, with respect to the connection destination candidate keyword selected by the connection destination candidate selection unit 1102, the necessity between the transaction object represented by the connection destination candidate keyword and the transaction object represented by the connection source keyword set by the connection source setting unit 1101, using a large-scale language model (step ST206). At this time, the necessity determination unit 1105 uses a large-scale language model such as ChatGPT, inputs information indicating the connection source keyword and the connection destination candidate keyword, and inquires about the necessity, thereby acquiring information indicating the necessity. When a plurality of connection destination candidate keywords are selected by the connection destination candidate selection unit 1102, the necessity determination unit 1105 makes the above determination for each of the connection destination candidate keywords.

[0100] Here, if the transaction route to be estimated is a procurement route, the necessity determination unit 1105 determines whether the transaction object represented by the connection destination candidate keyword is necessary in providing the transaction object represented by the connection source keyword. In addition, if the transaction route to be estimated is a sales route, the necessity determination unit 1105 determines whether the transaction object represented by the connection source keyword is necessary in providing the transaction object represented by the connection destination candidate keyword.

[0101] Furthermore, the abstraction level determination unit 1109 determines the abstraction level of the connection destination candidate keyword selected by the connection destination candidate selection unit 1102 using a large-scale language model (step ST207). At this time, the abstraction level determination unit 1109 uses a large-scale language model such as ChatGPT to acquire information indicating the abstraction levels of various keywords that are assumed to be related to the connection destination candidate keyword in advance, and determines the abstraction level of the connection destination candidate keyword selected by the connection destination candidate selection unit 1102 based on the information.

[0102] Next, the route connection determination unit 1106 determines whether or not to connect the transaction route based on the type of transaction object represented by the connection source keyword set by the connection source setting unit 1101, the acquisition result by the condition acquisition unit 1108, the determination result by the type determination unit 1103, the determination result by the category determination unit 1104, the determination result by the necessity determination unit 1105, and the determination result by the abstraction level determination unit 1109 (step ST208). When a plurality of connection destination candidate keywords are selected by the connection destination candidate selection unit 1102, the route connection determination unit 1106 makes the above determination for each of the connection destination candidate keywords.

[0103] Here, if the type of transaction object represented by the connection source keyword is a good, the route connection determination unit 1106 determines to connect the transaction route if the type of transaction object determined by the type determination unit 1103 is a good, the category of the transaction object determined by the category determination unit 1104 is not the final transaction stage, i.e., the intermediate transaction stage or the initial transaction stage, and the necessity determination unit 1105 determines that it is necessary. In addition, if the type of transaction object represented by the connection source keyword is a service, the route connection determination unit 1106 determines to connect the transaction route if the type of transaction object determined by the type determination unit 1103 is a service, the category of the transaction object determined by the category determination unit 1104 is not the final transaction stage, i.e., the intermediate transaction stage or the initial transaction stage, and the necessity determination unit 1105 determines that it is necessary. On the other hand, in cases other than those mentioned above, the route connection determining unit 1106 determines not to connect the transaction route.

[0104] Furthermore, in addition to the above judgment, the route connection judgment unit 1106 excludes from the connection candidates the candidate connection companies and candidate connection keywords selected by the candidate connection selection unit 1102 if the hierarchy of the transaction route to be connected is equal to or higher than the hierarchy indicated by the conditions acquired by the condition acquisition unit 1108, and if the abstraction level of the candidate connection keywords judged by the abstraction level judgment unit 1109 is equal to or higher than the abstraction level indicated by the conditions acquired by the condition acquisition unit 1108.

[0105] Next, if the route connection determination unit 1106 determines that a transaction route should be connected, the route connection unit 1107 sets the connection candidate company and connection candidate keyword selected by the connection candidate selection unit 1102 as the connection destination company and connection destination keyword, and connects the connection destination company and connection destination keyword to the connection source company and connection source keyword set by the connection source setting unit 1101 (step ST209).

[0106] Here, if the transaction route to be estimated is a procurement route, the route connection unit 1107 connects the connection destination company to the connection source company using a connection destination keyword. Furthermore, if the transaction route to be estimated is a sales route, the route connection unit 1107 connects the connection destination company to the connection source company using a connection source keyword.

[0107] Next, when the category determined by the category acquisition unit is the intermediate transaction stage for the destination company and destination keyword set by the route connection unit 1107, the source setting unit 1101 updates the setting of the destination company and destination keyword as the source company and source keyword (steps ST210 and ST211). After that, the sequence returns to step ST203, and the information processing device 1 repeats the above operations.

[0108] When the settings are updated by the connection source setting unit 1101, the necessity determination unit 1105 determines, for the connection destination candidate keywords selected by the connection destination candidate selection unit 1102, the necessity between the transaction object represented by the connection destination candidate keyword and the transaction object represented by the connection source keyword updated by the connection source setting unit 1101, and the necessity between the transaction object and the transaction object represented by the keyword at the starting point company. The necessity determination unit 1105 acquires information indicating the necessity by inputting information indicating the keyword at the starting point company, the connection source keyword, and the connection destination candidate keyword, for example, using a large-scale language model such as ChatGPT, and inquiring about the necessity. When a plurality of connection destination candidate keywords are selected by the connection destination candidate selection unit 1102, the necessity determination unit 1105 makes the above determination for each of the connection destination candidate keywords.

[0109] Here, if the transaction route to be estimated is a procurement route, the necessity determination unit 1105 determines whether the transaction object represented by the candidate connection keyword is necessary for each of the provision of the transaction object represented by the source keyword and the provision of the transaction object represented by the keyword at the starting company. In addition, if the transaction route to be estimated is a sales route, the necessity determination unit 1105 determines whether or not the transaction object represented by the connection source keyword and the transaction object represented by the keyword at the starting point company are necessary in providing the transaction object represented by the connection destination candidate keyword.

[0110] Furthermore, when the setting is updated by the connection source setting unit 1101, the route connection determination unit 1106 determines to connect the transaction route if the type of transaction object represented by the connection source keyword is a good, the type of transaction object determined by the type determination unit 1103 is a good, the category of the transaction object determined by the category determination unit 1104 is not the final transaction stage, i.e., the intermediate transaction stage or the initial transaction stage, and the necessity determination unit 1105 determines that all are necessary. In addition, if the type of transaction object represented by the connection source keyword is a service, the route connection determination unit 1106 determines to connect the transaction route if the type of transaction object determined by the type determination unit 1103 is a service, the category of the transaction object determined by the category determination unit 1104 is not the final transaction stage, i.e., the intermediate transaction stage or the initial transaction stage, and the necessity determination unit 1105 determines that all are necessary. On the other hand, in cases other than those mentioned above, the route connection determining unit 1106 determines not to connect the transaction route.

[0111] Furthermore, in addition to the above judgment, the route connection judgment unit 1106 excludes from the connection candidates the candidate connection companies and candidate connection keywords selected by the candidate connection selection unit 1102 if the hierarchy of the transaction route to be connected is equal to or higher than the hierarchy indicated by the conditions acquired by the condition acquisition unit 1108, and if the abstraction level of the candidate connection keywords judged by the abstraction level judgment unit 1109 is equal to or higher than the abstraction level indicated by the conditions acquired by the condition acquisition unit 1108.

[0112] On the other hand, if the category of the connection destination keyword is not the intermediate transaction stage, the information processing device 1 ends the sequence.

[0113] The information processing device 1 may display on an external display device information indicating the transaction path obtained by the path connection unit 1107. This allows the user to visually recognize the transaction path obtained by the information processing device 1.

[0114] FIG. 11 is a diagram showing an example of the operation of the information processing device 1 according to the second embodiment. In the information processing device 1 according to the second embodiment, the condition acquisition unit 1108 acquires the conditions in advance, the abstraction level determination unit 1109 determines the abstraction level of the transaction object, and the route connection determination unit 1106 performs an exclusion determination of the connection object based on this information.

[0115] In this case, for example, the condition acquiring unit 1108 acquires the following conditions: procurement route tier=3 and upper limit of abstraction level of connection destination candidate keyword=5.

[0116] Furthermore, the abstraction level determination unit 1109 presents examples of the abstraction levels of several keywords to the large-scale language model in advance, and then queries the model about the abstraction levels of keywords that are expected to be connection destination candidate keywords.

[0117] In this case, for example, the abstraction level determination unit 1109 inputs the following as the fifth question to the large-scale language model in advance: "You are a business relationship network analyst. A general term refers to a large class or a broad area. 'Automotive parts', 'units', and 'components' are general terms. A specific term refers to a specific item or an individual case. 'Carburetors', 'nutrunner systems', and 'shock absorbers' are specific terms. How abstract are fuel injection-related products, automotive parts, molds, components, fuel tanks, and carburetors? Please answer with one of the following: 1 = most specific, 2 = specific, 3 = slightly specific, 4 = normal, 5 = somewhat general, 6 = general, 7 = most general." In this case, the abstraction level determination unit 1109 receives information such as "Fuel injection related products: 5, automobile parts: 6, molds: 6, components: 7, fuel tanks: 4, carburetors: 2" from the large-scale language model as the fifth answer.

[0118] Then, the abstraction level determination unit 1109 determines the abstraction level of the connection destination candidate keyword selected by the connection destination candidate selection unit 1102 . Here, for example, if the connection destination candidate keyword selected by the connection destination candidate selection unit 1102 is a component, the abstraction level determination unit 1109 determines that the abstraction level is 7 from the information inquired in advance. Also, for example, if the connection destination candidate keyword selected by the connection destination candidate selection unit 1102 is "fuel tank," the abstraction level determination unit 1109 determines that the abstraction level is 4 based on the information inquired in advance.

[0119] Then, the route connection determining unit 1106 performs an exclusion determination for the connection destination candidate keywords based on the conditions acquired by the condition acquiring unit 1108 and the determination result by the abstraction level determining unit 1109 . For example, if the current search is performed at three or more levels and the connection candidate keyword is a component, its abstraction level is 7, which is higher than the upper limit of abstraction level (5), so this connection candidate keyword is excluded from the connection targets. On the other hand, if the currently searched layer is three or more layers and the candidate keyword is "fuel tank," the abstraction level is four, which is not higher than the upper limit of five, so this candidate keyword is not excluded from the list of candidates for connection. Furthermore, if the currently searched hierarchy is less than three levels, the candidate keywords are not excluded regardless of their abstraction level.

[0120] Although the above description has been given taking the case where the transaction route is a procurement route as an example, the same applies to the case where the transaction route is a sales route. For example, if a starting company sells battery modules for electric vehicles, then at the first level (direct sales destination), it would be natural for the connection candidate keyword to be "electric vehicle battery" and it would also be acceptable for it to be a connection target if it was "auto parts." However, at levels above the third level (sales destinations of the sales destinations of the sales destinations), for example, if the connection candidate keyword is "auto parts," it may be excluded from the connection target because it is too abstract. This is because as the levels go up, various auto parts flow into the company, and it becomes unclear which parts were used to make the "auto parts" that are the products being sold.

[0121] As described above, according to the second embodiment, the information processing device 1 includes a condition acquisition unit 1108 that acquires conditions indicating the hierarchy of the transaction route and the upper limit of the abstraction level of the connection destination candidate keywords, and an abstraction level determination unit 1109 that determines the abstraction level of the connection destination candidate keywords selected by the connection destination candidate selection unit 1102 using a large-scale language model, and the route connection determination unit 1106 determines whether to connect the transaction routes based on the type of transaction object represented by the connection source keyword set by the connection source setting unit 1101, the acquisition result by the condition acquisition unit 1108, the determination result by the type determination unit 1103, the determination result by the category determination unit 1104, the determination result by the necessity determination unit 1105, and the determination result by the abstraction level determination unit 1109. This enables the information processing device 1 according to the second embodiment to more appropriately estimate a transaction route for providing a specific transaction object.

[0122] Furthermore, according to the second embodiment, the route connection determination unit 1106 excludes from connection targets the connection target candidate companies and connection target candidate keywords selected by the connection target candidate selection unit 1102 when the hierarchical level of the transaction route to be connected is equal to or higher than the hierarchical level indicated by the condition acquired by the condition acquisition unit 1108, and when the abstraction level of the connection target candidate keyword determined by the abstraction level determination unit 1109 is equal to or higher than the abstraction level indicated by the condition acquired by the condition acquisition unit 1108. This enables the information processing device 1 according to the second embodiment to more appropriately estimate a transaction route for providing a specific transaction target.

[0123] Embodiment 3 Fig. 12 is a diagram showing a configuration example of the calculation unit 11 included in the information processing device 1 according to embodiment 3. In the configuration example of the calculation unit 11 included in the information processing device 1 according to embodiment 3 shown in Fig. 12, a risk determination unit 1110 is added to the calculation unit 11 included in the information processing device 1 according to embodiment 2 shown in Fig. 9. The other configuration example of the calculation unit 11 included in the information processing device 1 according to embodiment 3 shown in Fig. 12 is the same as the configuration example of the calculation unit 11 included in the information processing device 1 according to embodiment 2 shown in Fig. 9, and the same reference numerals are used and only the different parts will be described.

[0124] The risk determination unit 1110 determines the risk of at least one of the transaction targets represented by the destination company or the destination keyword using a large-scale language model for the destination company and destination keyword set by the route connection unit 1107. In this case, the risk determination unit 1110 uses a large-scale language model such as ChatGPT, inputs information indicating at least one of the destination company or the destination keyword, and queries the risk, thereby acquiring information indicating the risk of at least one of the transaction targets represented by the destination company or the destination keyword.

[0125] As a result, the information processing device 1 according to the third embodiment can identify risks, such as ESG-related issues, whether the company is subject to sanctions, or whether negative rumors are circulating in the news, for at least one of the companies or trading partners on the estimated transaction route.

[0126] In the above description, the risk determination unit 1110 is added to the information processing device 1 according to embodiment 2. However, the present invention is not limited to this, and the risk determination unit 1110 may be added to the information processing device 1 according to embodiment 1, and the same effects as those described above can be obtained.

[0127] As described above, according to the third embodiment, the information processing device 1 includes the risk determination unit 1110 that determines, using a large-scale language model, the risk associated with at least one of the connected companies and transaction targets represented by the connected keywords set by the route connection unit 1107. This enables the information processing device 1 according to the third embodiment to grasp the risk associated with at least one of the companies or transaction targets on the estimated transaction route.

[0128] In the above description, the risk determining unit 1110 determines the risk using a large-scale language model. However, the present invention is not limited to this, and the risk determination unit 1110 may determine the above-mentioned risk using, for example, a sanctions list or a black company list held by a company. Alternatively, a table showing risks such as ESG-related issues, sanctions targets, or whether negative rumors are circulating in the news may be created in advance using a large-scale language model, and the risk determination unit 1110 may use this table to determine the above-mentioned risk. In this way, the risk determination unit 1110 can also determine the above-mentioned risk without querying an external large-scale language model via an API each time.

[0129] Embodiment 4 Fig. 13 is a diagram showing an example of the configuration of the calculation unit 11 included in the information processing device 1 according to embodiment 4. In the example of the configuration of the calculation unit 11 included in the information processing device 1 according to embodiment 4 shown in Fig. 13, a reconnection proposal unit 1111 and a reconnection request acceptance unit 1112 are added to the calculation unit 11 included in the information processing device 1 according to embodiment 3 shown in Fig. 12. The other example of the configuration of the calculation unit 11 included in the information processing device 1 according to embodiment 4 shown in Fig. 13 is the same as the example of the configuration of the calculation unit 11 included in the information processing device 1 according to embodiment 3 shown in Fig. 12, and the same reference numerals are used and only the different parts will be described.

[0130] The reconnection proposal unit 1111 proposes, for the target transaction route obtained by the route connection unit 1107 and another transaction route different from the target transaction route, if there is a destination keyword on the other transaction route that is related to the destination keyword on the target transaction route, the related destination keyword and the destination company corresponding to the related destination keyword as the reconnection destination (first proposal).

[0131] Alternatively, for the target transaction route obtained by the route connection unit 1107 and another transaction route different from the target transaction route, if there is a destination keyword on the other transaction route that is related to the destination keyword on the target transaction route and the related destination keyword has a lower risk as determined by the risk determination unit 1110, the reconnection proposal unit 1111 proposes the related destination keyword and the destination company corresponding to the related destination keyword as a reconnection destination (second proposal).

[0132] The rewiring request receiving unit 1112 receives a rewiring request in response to a proposal by the rewiring proposing unit 1111. That is, the rewiring request receiving unit 1112 receives a rewiring request from a user.

[0133] When a reconnection request is received by the reconnection request receiving unit 1112, the route connection unit 1107 reconnects the target transaction route to the related destination keyword and destination company.

[0134] For example, consider the case where a certain product is electrified. In this case, the transaction route of the product (the target transaction route) and the transaction route of an electrical appliance equivalent to the electrified product (another transaction route) are investigated. Then, assume that there is a company that provides corresponding parts in both transaction routes. In this case, the reconnection proposal unit 1111 proposes the parts and the company on the transaction route of the electrical appliance as a reconnection destination (first proposal). This will make it possible to establish a trading route for the electrified products.

[0135] Also, for example, suppose that a company that provides a certain part in the transaction route (target transaction route) of a certain product is at risk. In this case, for example, the transaction route (other transaction route) of a competitor that provides a product equivalent to the above product is further investigated. Then, suppose that in the competitor's transaction route there is a company that provides a part corresponding to the above part, and that this company has a lower risk. In this case, the reconnection proposal unit 1111 proposes the above part and the above company on the competitor's transaction route as a reconnection destination (second proposal). This allows for the creation of lower-risk trading routes.

[0136] In the above description, the case has been described in which the rewiring proposing unit 1111 and the rewiring request receiving unit 1112 are added to the information processing device 1 according to the third embodiment. However, the present invention is not limited to this, and a rewiring proposal unit 1111 and a rewiring request reception unit 1112 may be added to the information processing device 1 according to embodiments 1 and 2. In this case, since the information processing device 1 does not have the risk determination unit 1110, the rewiring proposal unit 1111 makes a first proposal.

[0137] As described above, according to the fourth embodiment, the information processing device 1 includes a reconnection suggestion unit 1111 that, for a target transaction route obtained by the route connection unit 1107 and another transaction route different from the target transaction route, if a destination keyword related to a destination keyword on the target transaction route exists on the other transaction route, proposes the related destination keyword and the destination company corresponding to the related destination keyword as a reconnection destination. This allows the information processing device 1 according to the fourth embodiment to obtain a more appropriate transaction route.

[0138] Furthermore, according to the fourth embodiment, the information processing device 1 includes a reconnection suggestion unit 1111 that, for a target transaction route obtained by the route connection unit 1107 and another transaction route different from the target transaction route, proposes, as a reconnection destination, a connected company corresponding to the related connected destination keyword and the related connected destination keyword when a connected destination keyword related to the connected destination keyword on the target transaction route exists on the other transaction route and the related connected destination keyword has a lower risk as determined by the risk determination unit 1110. This enables the information processing device 1 according to the fourth embodiment to obtain a more appropriate transaction route.

[0139] Furthermore, according to the fourth embodiment, the information processing device 1 includes a reconnection request receiving unit 1112 that receives a reconnection request in response to a proposal by the reconnection proposing unit 1111, and the route connecting unit 1107, when a reconnection request is received by the reconnection request receiving unit 1112, reconnects the target transaction route to the related destination keyword and destination company. This allows the information processing device 1 according to the fourth embodiment to obtain a more appropriate transaction route.

[0140] Embodiment 5. Fig. 14 is a diagram showing a configuration example of the calculation unit 11 included in the information processing device 1 according to embodiment 5. In the configuration example of the calculation unit 11 included in the information processing device 1 according to embodiment 5 shown in Fig. 14, a trading object estimation unit 1113 is added to the calculation unit 11 included in the information processing device 1 according to embodiment 4 shown in Fig. 13. The other configuration example of the calculation unit 11 included in the information processing device 1 according to embodiment 5 shown in Fig. 14 is the same as the configuration example of the calculation unit 11 included in the information processing device 1 according to embodiment 4 shown in Fig. 13, and the same reference numerals are used, and only the different parts will be described.

[0141] The transaction object estimation unit 1113 estimates the transaction object when the destination keyword set by the path connection unit 1107 is not a keyword that directly represents the transaction object. In this case, for example, the transaction object estimation unit 1113 may acquire information indicating the transaction object by using a large-scale language model such as ChatGPT and inputting information indicating the destination keyword to inquire about the transaction object. For example, when the connection destination keyword is an industry name and the specific transaction object is unknown, the transaction object estimation unit 1113 estimates this transaction object.

[0142] In the above description, the case where the transaction target estimation unit 1113 is added to the information processing device 1 according to the fourth embodiment has been described. However, the present invention is not limited to this, and the trading object estimation unit 1113 may be added to the information processing device 1 according to the first to third embodiments, and the same effects as those described above can be obtained.

[0143] As described above, according to the fifth embodiment, the information processing device 1 includes the transaction object estimation unit 1113 that estimates the transaction object when the connection destination keyword set by the route connection unit 1107 is not a keyword that directly represents the transaction object. This allows the information processing device 1 according to the fifth embodiment to specifically grasp the transaction object on the transaction route.

[0144] Embodiment 6 In the first and second embodiments, various determinations are made using a large-scale language model. On the other hand, there may be a desire not to transmit keywords representing business partners in a company to the API (Application Programming Interface) of a large-scale language model. Therefore, in the sixth embodiment, a case will be described in which normalized keywords are used as keywords and various judgments are made on the normalized keywords using a judgment table.

[0145] 15 is a diagram showing an example of the configuration of the calculation unit 11 included in the information processing device 1 according to embodiment 6. In the example of the configuration of the calculation unit 11 included in the information processing device 1 according to embodiment 6 shown in FIG. 15, the connection source setting unit 1101, the connection destination candidate selection unit 1102, the type determination unit 1103, the category determination unit 1104, and the necessity determination unit 1105 included in the calculation unit 11 included in the information processing device 1 according to embodiment 1 shown in FIG. 2 are changed to a connection source setting unit 1101b, a connection destination candidate selection unit 1102b, a type determination unit 1103b, a category determination unit 1104b, and a necessity determination unit 1105b. The other example of the configuration of the calculation unit 11 included in the information processing device 1 according to embodiment 6 shown in FIG. 15 is the same as the example of the configuration of the calculation unit 11 included in the information processing device 1 according to embodiment 1 shown in FIG. 2, and the same reference numerals are used, and only the different parts will be described.

[0146] The connection source setting unit 1101b sets the company that is the starting point as the connection source company, and sets the keyword that represents the transaction target of the connection source company, normalized using a normalization table, as the connection source keyword. The normalization table is a table for converting a keyword into a keyword of the classification to which the keyword corresponds (normalized keyword), and an example of the normalization table is an export statistics commodity table. This normalization table may be created in advance using a large-scale language model such as ChatGPT. If the keyword representing the transaction object of the connecting company is already a normalized keyword, the connecting company setting unit 1101b sets the keyword representing the transaction object of the connecting company as the connecting company keyword as is.

[0147] The selection of keywords representing the starting company and the trading target of the starting company may be performed, for example, in response to a user operation, as in the first embodiment.

[0148] In addition, when the category determined by the category determination unit 1104b for the destination company and destination keyword set by the route connection unit 1107 is the intermediate transaction stage, the connection source setting unit 1101b updates the setting of the destination company and the destination keyword as the source company and the source keyword.

[0149] The connection destination candidate selection unit 1102b selects a connection destination candidate company as a connection destination candidate company based on the connection source company and connection source keyword set by the connection source setting unit 1101b, and selects keywords that represent the transaction target of the connection destination candidate company normalized using a normalization table as connection destination candidate keywords. In addition, if the keyword representing the business object of the above-mentioned connection destination candidate company is already a normalized keyword, the connection destination candidate selection unit 1102b selects the keyword representing the business object of the connection destination candidate company as the connection destination candidate keyword as is. The connection destination candidate selection unit 1102b selects one or more connection destination candidate companies and connection destination candidate keywords.

[0150] The selection method by the connection destination candidate selecting unit 1102b is the same as the selection method by the connection destination candidate selecting unit 1102 in the first embodiment.

[0151] Furthermore, when the transaction route to be estimated is a procurement route, the connection destination candidate selection unit 1102b performs the above selection in the upstream direction of the transaction relationship. Furthermore, when the transaction route to be estimated is a sales route, the connection destination candidate selection unit 1102b performs the above selection in the downstream direction of the transaction relationship.

[0152] The type determination unit 1103b determines the type of transaction object represented by the connection destination candidate keyword selected by the connection destination candidate selection unit 1102b using a type determination table. Here, the type determination table is a table showing the relationship between normalized keywords and the type of transaction object, and is created in advance using, for example, a large-scale language model such as ChatGPT. The type of transaction object represented by the connection destination candidate keyword determined by this type determination unit 1103b can be goods, services, or other. When a plurality of connection destination candidate keywords are selected by the connection destination candidate selection unit 1102b, the type determination unit 1103b makes the above determination for each connection destination candidate keyword.

[0153] The category determination unit 1104b determines, for a connection destination candidate keyword selected by the connection destination candidate selection unit 1102b, a category indicating the transaction stage of the transaction object represented by the connection destination candidate keyword using a category determination table. Here, the category determination table is a table showing the relationship between normalized keywords and categories indicating the transaction stage of the transaction object, and is created in advance using, for example, a large-scale language model such as ChatGPT. The category of the transaction object represented by the connection destination candidate keyword determined by this category determination unit 1104b can be the final transaction stage, intermediate transaction stage, or initial transaction stage. When a plurality of connection destination candidate keywords are selected by the connection destination candidate selection unit 1102b, the category determination unit 1104b performs the above determination for each of the connection destination candidate keywords.

[0154] The necessity determination unit 1105b determines the necessity between the transaction object represented by the connection destination candidate keyword selected by the connection destination candidate selection unit 1102b and the transaction object represented by the connection source keyword set by the connection source setting unit 1101b using a necessity determination table. Here, the necessity determination table is a table showing the necessity between the transaction objects represented by two normalized keywords, and is created in advance using, for example, a large-scale language model such as ChatGPT. When a plurality of connection destination candidate keywords are selected by the connection destination candidate selection unit 1102b, the necessity determination unit 1105b makes the above determination for each connection destination candidate keyword.

[0155] Here, if the transaction route to be estimated is a procurement route, the necessity determination unit 1105b determines whether the transaction object represented by the connection destination candidate keyword is necessary in providing the transaction object represented by the connection source keyword. In addition, if the transaction route to be estimated is a sales route, the necessity determination unit 1105b determines whether the transaction object represented by the connection source keyword is necessary in providing the transaction object represented by the connection destination candidate keyword.

[0156] Furthermore, when the settings are updated by the connection source setting unit 1101b, the necessity determination unit 1105b determines, for the connection destination candidate keywords selected by the connection destination candidate selection unit 1102b, the necessity between the transaction object represented by the connection destination candidate keyword and the transaction object represented by the connection source keyword updated by the connection source setting unit 1101b, and the necessity between the transaction object represented by the keyword at the starting point company. At this time, the necessity determination unit 1105b obtains information indicating the above-mentioned necessity using a necessity determination table. When a plurality of connection destination candidate keywords are selected by the connection destination candidate selection unit 1102b, the necessity determination unit 1105b makes the above determination for each connection destination candidate keyword.

[0157] Here, if the transaction route to be estimated is a procurement route, the necessity determination unit 1105b determines whether or not the transaction object represented by the candidate connection keyword is necessary for the provision of the transaction object represented by the source keyword and for the provision of the transaction object represented by the keyword at the starting company. In addition, if the transaction route to be estimated is a sales route, the necessity determination unit 1105b determines whether or not the transaction object represented by the connection source keyword and the transaction object represented by the keyword at the starting point company are necessary in providing the transaction object represented by the connection destination candidate keyword.

[0158] In addition, the route connection determination unit 1106 in embodiment 6 determines whether or not to connect a transaction route based on the type of transaction object represented by the connection source keyword set by the connection source setting unit 1101b, the determination result by the type determination unit 1103b, the determination result by the category determination unit 1104b, and the determination result by the necessity determination unit 1105b. When a plurality of connection destination candidate keywords are selected by the connection destination candidate selection unit 1102b, the route connection determination unit 1106 makes the above determination for each of the connection destination candidate keywords.

[0159] Here, if the type of transaction object represented by the connection source keyword is a good, the route connection determination unit 1106 determines to connect the transaction route if the type of transaction object determined by the type determination unit 1103b is a good, the category of the transaction object determined by the category determination unit 1104b is not the final transaction stage, i.e., the intermediate transaction stage or the initial transaction stage, and it is determined to be necessary by the necessity determination unit 1105b. In addition, if the type of transaction object represented by the connection source keyword is a service, the route connection determination unit 1106 determines to connect the transaction route if the type of transaction object determined by the type determination unit 1103b is a service, the category of the transaction object determined by the category determination unit 1104b is not the final transaction stage, i.e., the intermediate transaction stage or the initial transaction stage, and the necessity determination unit 1105b determines that it is necessary. On the other hand, in cases other than those mentioned above, the route connection determining unit 1106 determines not to connect the transaction route.

[0160] Furthermore, when the settings are updated by the connection source setting unit 1101b, the route connection determination unit 1106 determines to connect the transaction route if the type of transaction object represented by the connection source keyword is goods, the type of transaction object determined by the type determination unit 1103b is goods, the category of the transaction object determined by the category determination unit 1104b is not the final transaction stage, i.e., the intermediate transaction stage or the initial transaction stage, and the necessity determination unit 1105b determines that all are necessary. In addition, if the type of transaction object represented by the connection source keyword is a service, the route connection determination unit 1106 determines to connect the transaction route if the type of transaction object determined by the type determination unit 1103b is a service, the category of the transaction object determined by the category determination unit 1104b is not the final transaction stage, i.e., the intermediate transaction stage or the initial transaction stage, and the necessity determination unit 1105b determines that all are necessary. On the other hand, in cases other than those mentioned above, the route connection determining unit 1106 determines not to connect the transaction route.

[0161] Next, FIG. 16 shows an example of a category determination table and a necessity determination table. For example, in the decision table in Figure 16, we can see that liquefied petroleum gas (propane gas) "2711" is an intermediate good and is not necessary in the manufacturing process of beer "2203" (NO), but is necessary in the manufacturing process of smartphones "8517" (YES).

[0162] 16, the category determination table and the necessity determination table are combined into one table. However, this is not limiting, and the category determination table and the necessity determination table may be separate tables.

[0163] Next, FIG. 17 shows a specific example of the operation of the information processing device 1 according to the sixth embodiment. 17, the source company is Company A, and the keyword for Company A is "Integral Microchip." The candidate company to connect to is Company B, and the keyword for Company B is "MOBILE PHONE MAIN BOARD." Here, the information processing device 1 according to the first embodiment uses "Integral Microchip" as a connection source keyword and "MOBILE PHONE MAIN BOARD" as a connection destination candidate keyword, and makes various determinations using a large-scale language model. In contrast, in the information processing device 1 according to the sixth embodiment, "Integral Microchip" is converted into a keyword normalized by the normalization table and used as a source keyword, and "MOBILE PHONE MAIN BOARD" is converted into a keyword normalized by the normalization table and used as a destination candidate keyword. In the example of Fig. 17, "Integral Microchip" is converted into "HS8542.39" (an integrated circuit for recording and reproducing audio) by the normalization table and used as a source keyword, and "MOBILE PHONE MAIN BOARD" is converted into "HS8517.70" (a part of a telephone communication device) by the normalization table and used as a destination candidate keyword. Then, the information processing device 1 according to the sixth embodiment uses the connection source keywords and connection destination candidate keywords, which are normalized keywords, to make various determinations using a determination table. As a result, the information processing device 1 according to the sixth embodiment can estimate a transaction path without transmitting keywords representing the transaction target in a company to the API of a large-scale language model.

[0164] In the above, a case has been shown in which the information processing device 1 according to embodiment 1 is modified so that normalized keywords are used as keywords and various judgments are made on the normalized keywords using a judgment table. However, the present invention is not limited to this, and the information processing device 1 according to the second embodiment may be modified so as to use normalized keywords as keywords and perform various judgments on the normalized keywords using a judgment table.

[0165] In this case, for example, as shown in FIG. 18, an abstraction level determination unit 1109b is added to the calculation unit 11 included in the information processing device 1 according to the sixth embodiment shown in FIG.

[0166] The abstraction level determination unit 1109b determines the abstraction level of the connection destination candidate keyword selected by the connection destination candidate selection unit 1102b using an abstraction level determination table. Here, the abstraction level determination table is a table showing the relationship between normalized keywords and abstraction levels, and is created in advance using a large-scale language model such as ChatGPT.

[0167] In this case, the route connection determination unit 1106 determines whether or not to connect the transaction route based on the type of transaction object represented by the connection source keyword set by the connection source setting unit 1101b, the acquisition result by the condition acquisition unit 1108, the determination result by the type determination unit 1103b, the determination result by the category determination unit 1104b, the determination result by the necessity determination unit 1105b, and the determination result by the abstraction level determination unit 1109b.

[0168] At this time, in addition to the judgment described above, the route connection judgment unit 1106 excludes from the connection candidates the candidate connection companies and candidate connection keywords selected by the candidate connection selection unit 1102b if the hierarchy of the transaction route to be connected is equal to or higher than the hierarchy indicated by the conditions acquired by the condition acquisition unit 1108, and if the abstraction level of the candidate connection keywords judged by the abstraction level judgment unit 1109b is equal to or higher than the abstraction level indicated by the conditions acquired by the condition acquisition unit 1108.

[0169] As described above, according to the sixth embodiment, the information processing device 1 includes a connection source setting unit 1101b that sets keywords representing a company that is a starting point and a transaction object at that company as a connection source company and a connection source keyword, a connection destination candidate selection unit 1102b that selects keywords representing a company that is a connection destination candidate and a transaction object at that company as a connection destination candidate company and a connection destination candidate keyword based on the connection source company and the connection source keyword set by the connection source setting unit 1101b, a type determination unit 1103b that determines the type of transaction object represented by the connection destination candidate keyword using a type determination table for the connection destination candidate keyword selected by the connection destination candidate selection unit 1102b, a category determination unit 1104b that determines the category indicating the transaction stage of the transaction object represented by the connection destination candidate keyword using a category determination table for the connection destination candidate keyword selected by the connection destination candidate selection unit 1102b, and a category determination unit 1105b that determines the category indicating the transaction stage of the transaction object represented by the connection destination candidate keyword using a category determination table for the connection destination candidate keyword selected by the connection destination candidate selection unit 1102b. The system is provided with: a necessity determination unit 1105b that determines, using a necessity determination table, the necessity between the transaction object represented by the connection destination candidate keyword and the transaction object represented by the connection source keyword set by the connection source setting unit 1101b; a route connection determination unit 1106 that determines whether to connect a transaction route based on the type of transaction object represented by the connection source keyword set by the connection source setting unit 1101b, the determination result by the type determination unit 1103b, the determination result by the category determination unit 1104b, and the determination result by the necessity determination unit 1105b; and a route connection unit 1107 that, when it is determined by the route connection determination unit 1106 that the transaction route should be connected, sets the connection destination candidate company and connection destination candidate keyword selected by the connection destination candidate selection unit 1102b as a connection destination company and connection destination keyword, and connects the connection destination company and the connection destination keyword to the connection source company and connection source keyword set by the connection source setting unit 1101b. Furthermore, according to this embodiment 6, the information processing device 1 has a connection source setting unit 1101b that sets, as a connection source keyword, a keyword that represents the business object of the starting company, normalized using a normalization table, and a connection destination candidate selection unit 1102b that selects, as a connection destination candidate keyword, a keyword that represents the business object of the connection destination candidate company, normalized using a normalization table. Furthermore, according to this embodiment 6, the information processing device 1 is provided with a condition acquisition unit 1108 that acquires conditions indicating the hierarchy of the transaction route and the upper limit of the abstraction level of the connection destination candidate keywords, and an abstraction level determination unit 1109b that determines the abstraction level of the connection destination candidate keywords selected by the connection destination candidate selection unit 1102b using an abstraction level determination table, and the route connection determination unit 1106 determines whether to connect the transaction route based on the type of transaction object represented by the connection source keyword set by the connection source setting unit 1101b, the acquisition result by the condition acquisition unit 1108, the determination result by the type determination unit 1103b, the determination result by the category determination unit 1104b, the determination result by the necessity determination unit 1105b, and the determination result by the abstraction level determination unit 1109b. As a result, the information processing device 1 according to the sixth embodiment can estimate a transaction path without transmitting keywords representing the transaction target in a company to the API of a large-scale language model.

[0170] Furthermore, according to the sixth embodiment, the information processing method includes the steps of: a connection source setting unit 1101b setting keywords representing a company that is a starting point and a transaction object of the company as a connection source company and a connection source keyword; a connection destination candidate selection unit 1102b selecting keywords representing a company that is a connection destination candidate and a transaction object of the company as a connection destination candidate company and a connection destination candidate keyword based on the connection source company and the connection source keyword set by the connection source setting unit 1101b; a type determination unit 1103b determining the type of transaction object represented by the connection destination candidate keyword using a type determination table for the connection destination candidate keyword selected by the connection destination candidate selection unit 1102b; a category determination unit 1104b determining the category indicating the transaction stage of the transaction object represented by the connection destination candidate keyword using a category determination table for the connection destination candidate keyword selected by the connection destination candidate selection unit 1102b; the step of determining, by a route connection determination unit 1106, whether or not to connect a transaction route based on the type of transaction route represented by the route connection determination unit 1101b, the determination result by the type determination unit 1103b, the determination result by the category determination unit 1104b, and the determination result by the necessity determination unit 1105b, for the route connection determination unit 1106, when it is determined by the route connection determination unit 1106 that the transaction route should be connected, setting the route connection candidate company and route connection candidate keyword selected by the route connection candidate selection unit 1102b as a destination company and a destination keyword, and connecting the destination company and the destination keyword to the source company and source keyword set by the source setting unit 1101b. As a result, the information processing method according to the sixth embodiment makes it possible to estimate a transaction path without transmitting keywords representing business objects in a company to the API of a large-scale language model.

[0171] Embodiment 7 Fig. 19 is a diagram showing an example of the configuration of the calculation unit 11 included in the information processing device 1 according to embodiment 7. In the example of the configuration of the calculation unit 11 included in the information processing device 1 according to embodiment 7 shown in Fig. 19, a replacement target acceptance unit 1114 and a replacement determination unit 1115 are added to the calculation unit 11 included in the information processing device 1 according to embodiment 6 shown in Fig. 15. The other example of the configuration of the calculation unit 11 included in the information processing device 1 according to embodiment 7 shown in Fig. 19 is the same as the example of the configuration of the calculation unit 11 included in the information processing device 1 according to embodiment 6 shown in Fig. 15, and the same reference numerals are used and only the different parts will be described.

[0172] The substitution target receiving unit 1114 receives a destination company and a destination keyword that are substitution targets from among the destination companies and destination keywords set by the route connection unit 1107. That is, the substitution target receiving unit 1114 receives a destination company and a destination keyword that are substitution targets specified by the user.

[0173] When the replacement target receiving unit 1114 receives a connection destination company and a connection destination keyword that are the replacement target, the substitutability determination unit 1115 searches for keywords that represent substitutable companies and trading targets of the companies using a substitution determination table. Here, the substitution determination table is a table that indicates the substitutability between two normalized keywords, and is created in advance using, for example, a large-scale language model such as ChatGPT. In this case, the substitution determination unit first extracts normalized keywords that can be substituted for the destination keyword (normalized keyword) to be substituted using a substitution determination table, and then searches for keywords that represent business partners in the company that correspond to the extracted normalized keywords.

[0174] Next, FIG. 20 shows an example of a substitutability determination table. For example, from Figure 20, it can be seen that for "0202" frozen beef (normalized keyword), "0203" frozen pork and "0207" frozen chicken are substitutable (YES) normalized keywords. By using such a substitutability determination table, it becomes possible to search for a trading object of a company that can be substituted for the trading object of a specific company on the trading route obtained by the route connection unit 1107.

[0175] In the above description, the case has been described in which the alternative target receiving unit 1114 and the alternativeness determining unit 1115 are added to the information processing device 1 shown in FIG. However, the present invention is not limited to this, and an alternative target receiving unit 1114 and an alternativeness determining unit 1115 may be added to the information processing device 1 shown in FIG.

[0176] As described above, according to the seventh embodiment, the information processing device 1 includes a substitution target receiving unit 1114 that receives destination companies and destination keywords to be substituted from among the destination companies and destination keywords set by the route connection unit 1107, and a substitutability determination unit 1115 that, when the destination companies and destination keywords to be substituted are received by the substitution target receiving unit 1114, searches for keywords that represent substitutable companies and business partners at the companies using a substitution determination table. This makes it possible for the information processing device 1 according to the seventh embodiment to search for keywords that represent substitutable companies and business partners at the companies for the destination companies and destination keywords to be substituted.

[0177] It should be noted that the embodiments may be freely combined, or any of the components in each embodiment may be modified, or any of the components in each embodiment may be omitted. [Explanation of symbols]

[0178] 1 Information processing device, 11 arithmetic unit, 12 memory unit, 101 Input / output interface, 102 Processor, 103 Memory, 1101, 1101b Connection source setting unit, 1102, 1102b Connection destination candidate selection unit, 1103, 1103b Type determination unit, 1104, 1104b Category determination unit, 1105, 1105b Necessity determination unit, 1106 Route connection determination unit, 1107 Route connection unit, 1108 Condition acquisition unit, 1109, 1109b Abstraction level determination unit, 1110 Risk determination unit, 1111 Reconnection proposal unit, 1112 Reconnection request reception unit, 1113 Transaction object estimation unit, 1114 Substitution object reception unit, 1115 Substitutability determination unit.

Claims

1. a connection source setting unit that sets keywords representing a company that is a starting point and a transaction target of the company as a connection source company and a connection source keyword; a connection destination candidate selection unit that selects, based on the connection source companies and connection source keywords set by the connection source setting unit, keywords that represent connection destination candidate companies and transaction targets of the companies as connection destination candidate companies and connection destination candidate keywords; a type determination unit that determines, with respect to the connection destination candidate keyword selected by the connection destination candidate selection unit, a type of transaction object represented by the connection destination candidate keyword using a large-scale language model; a category determination unit that determines, with respect to the connection destination candidate keyword selected by the connection destination candidate selection unit, a category indicating a transaction stage of a transaction object represented by the connection destination candidate keyword using a large-scale language model; a necessity determination unit that determines, with respect to the connection destination candidate keyword selected by the connection destination candidate selection unit, the necessity between a transaction object represented by the connection destination candidate keyword and a transaction object represented by the connection source keyword set by the connection source setting unit using a large-scale language model; a route connection determination unit that determines whether to connect a transaction route based on the type of transaction object represented by the connection source keyword set by the connection source setting unit, the determination result by the type determination unit, the determination result by the category determination unit, and the determination result by the necessity determination unit; a route connection unit that, when it is determined by the route connection determination unit that a transaction route should be connected, sets the connection destination candidate company and the connection destination candidate keyword selected by the connection destination candidate selection unit as a connection destination company and a connection destination keyword, and connects the connection destination company and the connection destination keyword to the connection source company and connection source keyword set by the connection source setting unit; An information processing device comprising:

2. a connection source setting unit that sets keywords representing a company that is a starting point and a transaction target of the company as a connection source company and a connection source keyword; a connection destination candidate selection unit that selects, based on the connection source companies and connection source keywords set by the connection source setting unit, keywords that represent connection destination candidate companies and transaction targets of the companies as connection destination candidate companies and connection destination candidate keywords; a type determination unit that determines the type of transaction object represented by the connection destination candidate keyword selected by the connection destination candidate selection unit using a type determination table; a category determination unit that determines a category indicating a transaction stage of a transaction object represented by the connection destination candidate keyword selected by the connection destination candidate selection unit using a category determination table; a necessity determination unit that determines, with respect to the connection destination candidate keyword selected by the connection destination candidate selection unit, the necessity between the transaction object represented by the connection destination candidate keyword and the transaction object represented by the connection source keyword set by the connection source setting unit, based on a necessity determination table; a route connection determination unit that determines whether to connect a transaction route based on the type of transaction object represented by the connection source keyword set by the connection source setting unit, the determination result by the type determination unit, the determination result by the category determination unit, and the determination result by the necessity determination unit; a route connection unit that, when it is determined by the route connection determination unit that a transaction route should be connected, sets the connection destination candidate company and the connection destination candidate keyword selected by the connection destination candidate selection unit as a connection destination company and a connection destination keyword, and connects the connection destination company and the connection destination keyword to the connection source company and connection source keyword set by the connection source setting unit; An information processing device comprising:

3. the connection source setting unit sets, as a connection source keyword, a keyword that is obtained by normalizing a keyword that represents a transaction target in a starting company using a normalization table; The connection destination candidate selection unit selects keywords that represent business partners of the connection destination candidate companies, normalized using a normalization table, as connection destination candidate keywords.

3. The information processing apparatus according to claim 2.

4. The route connection determination unit When the type of transaction object represented by the connection source keyword is a good, if the type of transaction object determined by the type determination unit is a good, and the category of the transaction object determined by the category determination unit is not in the final transaction stage but is determined to be necessary by the necessity determination unit, it is determined that a transaction path connection is to be made; When the type of transaction object represented by the connection source keyword is a service, if the type of transaction object determined by the type determination unit is a service, and the category of the transaction object determined by the category determination unit is not in the final transaction stage but is determined to be necessary by the necessity determination unit, it is determined that a transaction path connection is to be performed.

3. The information processing apparatus according to claim 1 or 2.

5. the connection source setting unit updates the setting of the connection destination company and the connection destination keyword as a connection source company and a connection source keyword when the category determined by the category determination unit is an intermediate transaction stage with respect to the connection destination company and the connection destination keyword set by the route connection unit; When the setting is updated by the connection source setting unit, the necessity determination unit determines, with respect to the connection destination candidate keyword selected by the connection destination candidate selection unit, the necessity between the transaction object represented by the connection destination candidate keyword and the transaction object represented by the connection source keyword updated by the connection source setting unit, and the necessity between the transaction object represented by the keyword in the starting point company.

3. The information processing apparatus according to claim 1, wherein the information processing apparatus is a computer.

6. The route connection determination unit When the type of transaction object represented by the connection source keyword is a good, if the type of transaction object determined by the type determination unit is a good, and the category of the transaction object determined by the category determination unit is not in the final transaction stage, and if the necessity determination unit determines that all of the categories are necessary, it is determined that a transaction path connection is to be performed; When the type of transaction object represented by the connection source keyword is a service, if the type of transaction object determined by the type determination unit is a service, and the category of the transaction object determined by the category determination unit is not in the final transaction stage, and if the necessity determination unit determines that all of the transaction object categories are necessary, it is determined that a transaction path is to be connected.

6. The information processing apparatus according to claim 5,

7. a condition acquisition unit that acquires conditions indicating the upper limit of the level of abstraction of transaction route hierarchies and connection candidate keywords; an abstraction level determination unit that determines the abstraction level of the connection destination candidate keyword selected by the connection destination candidate selection unit using a large-scale language model; The route connection determination unit determines whether to connect a transaction route based on the type of transaction object represented by the connection source keyword set by the connection source setting unit, the acquisition result by the condition acquisition unit, the determination result by the type determination unit, the determination result by the category determination unit, the determination result by the necessity determination unit, and the determination result by the abstraction level determination unit.

2. The information processing apparatus according to claim 1, wherein:

8. a condition acquisition unit that acquires conditions indicating the upper limit of the level of abstraction of transaction route hierarchies and connection candidate keywords; an abstraction level determination unit that determines the abstraction level of the connection destination candidate keyword selected by the connection destination candidate selection unit using an abstraction level determination table; The route connection determination unit determines whether to connect a transaction route based on the type of transaction object represented by the connection source keyword set by the connection source setting unit, the acquisition result by the condition acquisition unit, the determination result by the type determination unit, the determination result by the category determination unit, the determination result by the necessity determination unit, and the determination result by the abstraction level determination unit.

3. The information processing apparatus according to claim 2.

9. The route connection determination unit excludes the connection destination candidate companies and connection destination candidate keywords selected by the connection destination candidate selection unit from the connection targets when the hierarchical level of the transaction route to be connected is equal to or higher than the hierarchical level indicated by the condition acquired by the condition acquisition unit, and the abstraction level of the connection destination candidate keywords determined by the abstraction level determination unit is equal to or higher than the abstraction level indicated by the condition acquired by the condition acquisition unit.

9. The information processing device according to claim 7 or 8.

10. a risk determination unit that determines, with respect to the destination company and the destination keyword set by the route connection unit, the risk of at least one of the destination company or the transaction target represented by the destination keyword using a large-scale language model; 2. The information processing apparatus according to claim 1, wherein:

11. With regard to the target transaction route obtained by the route connection unit and another transaction route different from the target transaction route, if a connection destination keyword related to a connection destination keyword on the target transaction route exists on the other transaction route, the related connection destination keyword and the connection destination company corresponding to the related connection destination keyword are proposed as a connection destination.

2. The information processing apparatus according to claim 1, wherein:

12. a re-splicing request receiving unit that receives a re-splicing request in response to the proposal by the re-splicing proposal unit; When the reconnection request receiving unit receives a reconnection request, the route connection unit reconnects the target transaction route to the related destination keyword and destination company.

12. The information processing apparatus according to claim 11.

13. With regard to the target transaction route obtained by the route connection unit and another transaction route different from the target transaction route, if a connection destination keyword related to a connection destination keyword on the target transaction route exists on the other transaction route and the related connection destination keyword has a lower risk determined by the risk determination unit, a reconnection suggestion unit is provided which proposes the related connection destination keyword and the connection destination company corresponding to the related connection destination keyword as a reconnection destination.

11. The information processing apparatus according to claim 10.

14. a re-splicing request receiving unit that receives a re-splicing request in response to the proposal by the re-splicing proposal unit; When the reconnection request receiving unit receives a reconnection request, the route connection unit reconnects the target transaction route to the related destination keyword and destination company.

14. The information processing apparatus according to claim 13,

15. A transaction object estimation unit is provided for estimating a transaction object when the destination keyword set by the route connection unit is not a keyword that directly represents the transaction object.

2. The information processing apparatus according to claim 1, wherein:

16. a replacement target receiving unit that receives a replacement target connection destination company and a connection destination keyword from among the connection destination companies and connection destination keywords set by the route connection unit; a substitutability determination unit that, when the substitutable target receiving unit receives the connection destination company and the connection destination keyword that are the substitutable target, searches for a substitutable company and a keyword that represents a transaction target of the company using a substitution determination table.

3. The information processing apparatus according to claim 2.

17. a step in which a connection source setting unit sets keywords representing a company that is a starting point and a trading target of the company as a connection source company and a connection source keyword; a step in which a connection destination candidate selection unit selects, based on the connection source companies and connection source keywords set by the connection source setting unit, keywords representing connection destination candidate companies and transaction targets of the companies as connection destination candidate companies and connection destination candidate keywords; a type determination unit determining, with respect to the connection destination candidate keyword selected by the connection destination candidate selection unit, a type of transaction object represented by the connection destination candidate keyword using a large-scale language model; a step in which a category determination unit determines, with respect to the connection destination candidate keyword selected by the connection destination candidate selection unit, a category indicating a transaction stage of a transaction object represented by the connection destination candidate keyword using a large-scale language model; a necessity determination unit determining, with respect to the connection destination candidate keyword selected by the connection destination candidate selection unit, the necessity between the transaction object represented by the connection destination candidate keyword and the transaction object represented by the connection source keyword set by the connection source setting unit using a large-scale language model; a step in which a route connection determination unit determines whether or not to connect a transaction route based on the type of transaction object represented by the connection source keyword set by the connection source setting unit, the determination result by the type determination unit, the determination result by the category determination unit, and the determination result by the necessity determination unit; a step in which, when the route connection determination unit determines that a transaction route should be connected, the route connection unit sets the connection destination candidate company and the connection destination candidate keyword selected by the connection destination candidate selection unit as a connection destination company and a connection destination keyword, and connects the connection destination company and the connection destination keyword to the connection source company and connection source keyword set by the connection source setting unit; An information processing method comprising:

18. a step in which a connection source setting unit sets keywords representing a company that is a starting point and a trading target of the company as a connection source company and a connection source keyword; a step in which a connection destination candidate selection unit selects, based on the connection source companies and connection source keywords set by the connection source setting unit, keywords representing connection destination candidate companies and transaction targets of the companies as connection destination candidate companies and connection destination candidate keywords; a type determination unit determining, with respect to the connection destination candidate keyword selected by the connection destination candidate selection unit, the type of transaction object represented by the connection destination candidate keyword based on a type determination table; a step in which a category determination unit determines, with respect to the connection destination candidate keyword selected by the connection destination candidate selection unit, a category indicating a transaction stage of the transaction object represented by the connection destination candidate keyword based on a category determination table; a necessity determination unit determining, with respect to the connection destination candidate keyword selected by the connection destination candidate selection unit, the necessity between the transaction object represented by the connection destination candidate keyword and the transaction object represented by the connection source keyword set by the connection source setting unit, based on a necessity determination table; a step in which a route connection determination unit determines whether or not to connect a transaction route based on the type of transaction object represented by the connection source keyword set by the connection source setting unit, the determination result by the type determination unit, the determination result by the category determination unit, and the determination result by the necessity determination unit; a step in which, when the route connection determination unit determines that a transaction route should be connected, the route connection unit sets the connection destination candidate company and the connection destination candidate keyword selected by the connection destination candidate selection unit as a connection destination company and a connection destination keyword, and connects the connection destination company and the connection destination keyword to the connection source company and connection source keyword set by the connection source setting unit; An information processing method comprising:

19. An information processing program for causing a computer to function as the information processing device according to claim 1 or 2.

Images