JP2021111281A - Business operator classification device, method, program, business operator evaluation system, and credit risk evaluation system - Google Patents

Abstract

To classify business operators appropriately.SOLUTION: A business operator classification device 3A has a relevant information acquisition unit 5 and a group generation unit 6. The relevant information acquisition unit 5 acquires business operator related information, which is information including a plurality of items related to each of a plurality of business operators. The group generation unit 6 uses the plurality of business operators as elements of the whole set, executes clustering for dividing the whole set into a plurality of subsets based on the business operator related information to generate a plurality of clusters to which one or more businesses belong, and defines each of the generated plurality of clusters as a group for classifying the business operators.SELECTED DRAWING: Figure 1

Description

The present invention relates to a technique for classifying a target business operator.

There are various types of businesses, including sole proprietorships, corporations, and organizations, and the activities vary depending on these types of businesses and individual businesses. The credit risk score used for the credit of a business operator can be calculated based on various information obtained from the attributes and activities of the business operator. In order to obtain the credit risk score of a business operator with good accuracy, the business operator is classified by industry, a credit risk index calculation model is constructed for each industry, and various information of each business operator is used in the credit risk index calculation model. May be entered. In general, businesses such as individual businesses, corporations and groups are classified into industries such as construction, wholesale, real estate, retail food service, service, and manufacturing according to the type of business or business.
However, the type of business is often declared by the business operator, and the definition of the type of business is often ambiguous.

Patent Document 1 discloses a technique for classifying companies by existing industries. In the technology disclosed in Patent Document 1, the composition ratio of sales for each business field of the company subject to classification is calculated, and the business field having the highest sales composition ratio is determined as the industry classification of the company. ..

Japanese Unexamined Patent Publication No. 2004-227426

As described above, the technology disclosed in Patent Document 1 classifies each company into an industry having a high sales composition ratio. This fits a company into one of the pre-prepared industries as an existing framework.

However, in recent years, it seems that it cannot be classified into businesses that are developing multiple different businesses, businesses whose side businesses are actually the main source of revenue, or existing industries such as the sharing economy. The number of businesses that are engaged in such businesses is increasing. Therefore, there is a possibility that the business operators cannot be properly classified by the method of Patent Document 1.

An object of the present invention is to provide a technique for appropriately classifying business operators.

The business operator classification device according to one embodiment of the present invention includes a related information acquisition unit that acquires business operator-related information that is information on a plurality of items related to each of the plurality of business operators, and the plurality of business operators as a whole. As an element of the set, based on the business operator-related information, clustering is executed to divide the entire set into a plurality of subsets, a plurality of clusters to which one or more business operators belong are generated, and the generated plurality of clusters are generated. Each of these has a group generation unit, which is defined as a group for classifying businesses.

According to the present invention, it is possible to appropriately classify businesses.

Overall view of the business operator evaluation system according to the first embodiment. The physical block diagram of the business operator classification apparatus which concerns on Example 1. FIG. FIG. 6 is a flow chart of a business operator classification process according to the first embodiment. The flow chart of the group generation processing which concerns on Example 1. FIG. The figure which shows an example of the output screen which concerns on Example 1. FIG. The flow chart of the group name setting process which concerns on Example 1. FIG. The flow chart of the group update process which concerns on Example 1. FIG. Overall view of the business operator evaluation system according to the second embodiment. Overall view of the business operator evaluation system according to the third embodiment. Overall view of the credit risk assessment system according to Example 4. The flow chart of the group update process which concerns on Example 5.

Hereinafter, examples of the present invention will be described with reference to the drawings.

First, the business operator classification device of this embodiment divides the entire set including a plurality of business operators into a plurality of subsets by clustering based on the information related to each business operator, and the generated cluster is divided into a plurality of business operators. It is defined as a group for classifying. The following is an example of a business operator evaluation system equipped with a business operator classification device.

FIG. 1 is an overall view of the business operator evaluation system according to the first embodiment. The business operator evaluation system 1A is a system for evaluating business operators clustered in the same group. The business operator may be a sole proprietorship, various corporations, various organizations, or the like.

As shown in FIG. 1, the business operator evaluation system 1A includes a database 2 on the cloud, a business operator classification device 3A that classifies businesses by clustering, and an index calculation unit 4. The database 2 and the business classification device 3A are connected to each other so as to be able to communicate with each other via a telecommunication line such as a LAN or a network.

The database 2 records business operator-related information, which is information including a plurality of items related to each of the plurality of business operators. The business-related information includes a plurality of items that numerically represent the activities performed by the business. Examples of the item group including a plurality of items include a cost item group and an asset item group. The cost item group includes a plurality of cost items that are items related to money spent by the business operator. The asset item group includes a plurality of asset items which are items in which the assets owned by the business operator are expressed in money. As these cost items and asset items, the account items of the journal information or the total items of the account items of the journal information are used. For example, business-related information includes accounting-related information, which is a value related to business operator accounting, and asset-related information, which is a value related to business operator assets, as major items. The business-related information may include various information about the business, such as the size of the company, the business partner, and the number of reviews on the Internet such as the SNS site.

Accounting-related information is the business cost paid by the business operator. The project cost includes the purchase of goods for the current period, manufacturing cost, labor cost, and depreciation cost as sub-items. That is, the cost items include the current product purchase, the manufacturing cost, the labor cost, and the depreciation cost. The accounting-related information may include the sales and profits of the business operator.

Asset-related information includes current assets, tangible fixed assets, and intangible fixed assets (in Fig. 5, others) in the middle items. Liquid assets include cash and deposits, trade receivables, inventories, and other liquid assets in sub-items. Property, plant and equipment include land and buildings, machinery, vehicle carriers, and fixtures and fixtures in sub-items. That is, the asset items include cash and deposits, trade receivables, inventories, other current assets, land and buildings, machinery and equipment, vehicle carriers, equipment and fixtures, and intangible fixed assets. good. In particular, asset items preferably include cash and deposits, trade receivables, inventories, other current assets, land and buildings, and intangible fixed assets.

The business cost and asset-related information may be a value obtained by accumulating journal information within a predetermined period (for example, one year) of each business operator. The journal information is a value obtained by dividing the bookkeeping transactions generated by each business operator into credits and debits and classifying them into their respective accounts. Journal information is updated each time a bookkeeping transaction occurs. The journal information may be a private value that cannot be viewed by anyone other than each business operator and the accounting business operator holding the accounting-related information of each business operator. The journal information may be calculated and used for any period or area according to the purpose. As a result, the business-related information can be changed according to the purpose of use of the user.

The business operator classification device 3A has a related information acquisition unit 5, a group generation unit 6, and a group distribution unit 7.

The related information acquisition unit 5 acquires business-related information from the database 2 and outputs the acquired business-related information to the group generation unit 6.

The group generation unit 6 has the composition ratio of each cost item in the cost item group and the composition ratio of each asset item in the asset item group included in the business operator-related information input from the related information acquisition unit 5. calculate. Specifically, the group generation unit 6 calculates the total business cost by adding up the business costs for each sub-item of each business operator, and totals the asset-related information for each sub-item of each business operator to calculate the total assets. Calculate relevant information. The group generation unit 6 calculates the ratio of each sub-item to the calculated total project cost, and also calculates the ratio of each sub-item to the calculated total asset-related information.

Next, the group generation unit 6 executes clustering that divides the entire set into a plurality of subsets based on the operator-related information with a plurality of operators as elements of the overall set, and one or more operators belong to the group generation unit 6. Create multiple clusters. Here, the group generation unit 6 executes clustering with the respective composition ratios of the cost items in the cost item group and the respective composition ratios of the asset items in the asset item group as variables. The group generation unit 6 defines each of the generated plurality of clusters as a group for classifying the business operators. For example, the group generation unit 6 generates a plurality of clusters by clustering using the composition ratio of each business operator-related information as a variable, and defines each of the generated plurality of clusters as a group. The group generation unit 6 outputs a plurality of defined groups to the index calculation unit 4. The group generation unit 6 may generate a plurality of clusters by applying the latent Dirichlet allocation method or the k-means method to the business expenses and asset-related information of a plurality of businesses. As a result, it is possible to appropriately determine the group according to the business operator.

After the group generation unit 6 determines the group, the group distribution unit 7 distributes the group to which the new business operator belongs from the plurality of groups already determined. For example, when the related information acquisition unit 5 acquires the business cost and asset-related information of a new business operator, the acquired business cost and asset-related information of the new business operator are input to the group distribution unit 7. , A plurality of groups defined by the group generation unit 6 are input. Based on the business cost and asset-related information of the new business operator, the group distribution department 7 divides the group to which the new business operator belongs from among the plurality of groups already determined by the group generation department 6, and the new business operator The group to which is belongs is output to the index calculation unit 4.

The index calculation unit 4 evaluates each business operator belonging to the same group among the plurality of groups input from the group generation unit 6. Specifically, the index calculation unit 4 relatively compares the businesses belonging to the same group with each other based on the business costs and asset-related information of the plurality of businesses input from the related information acquisition unit 5. , The score as an example of the "index" about the business operator is calculated and output to the display device 29.

FIG. 2 is a physical configuration diagram of the business operator classification device according to the first embodiment.

As the business operator classification device (or terminal) 3A described above, a computer 20 having a processor (hereinafter, CPU) 21, a memory 22, an auxiliary storage device 23, and a communication interface (communication I / F) 24 is used. .. As an example, the computer 20 may be a general-purpose computer such as a personal computer (PC).

The CPU 21 executes a program stored in the memory 22. The number of CPUs 21 may be 1 or more. The computer 20 may have a plurality of CPUs 21. Further, the CPU 21 may be a so-called multi-core processor having a plurality of processor cores. The memory 22 includes a ROM (Read Only Memory) which is a non-volatile storage element and a RAM (Random Access Memory) which is a volatile storage element. The ROM stores an invariant program (for example, BIOS) and the like. The RAM is a high-speed and volatile storage element such as a DRAM (Dynamic Random Access Memory), and temporarily stores a program executed by the CPU 21 and data used when the program is executed.

The auxiliary storage device 23 is, for example, a large-capacity and non-volatile storage device such as a magnetic storage device (HDD) or a flash memory (SSD). The auxiliary storage device 23 stores a program executed by the CPU 21 and data used when the program is executed. That is, after the program is read from the auxiliary storage device 23, it is loaded into the memory 22 and executed by the CPU 21. The communication interface 24 is a network interface device that controls communication with other devices according to a predetermined protocol.

The computer 20 may have an input interface (input I / F) 25 and an output interface (output I / F) 28. A keyboard 26, a mouse 27, and the like are connected to the input interface 25. The input interface 25 is an interface that receives input from the operator. A display device 29 as an example of the “display unit”, a printer (not shown), or the like is connected to the output interface 28. The output interface and the like 28 are interfaces that output the execution result of the program in a format that can be visually recognized by the operator.

FIG. 3 is a flow chart of the business operator classification process according to the first embodiment.

In the business classification process, first, the related information acquisition unit 5 acquires the business costs and asset-related information of a plurality of business operators from the database 2 according to the purpose of classification, and the acquired business cost and assets of the plurality of business operators. The related information is output to the group generation unit 6 (S301). Next, the group generation unit 6 calculates a plurality of composition ratios for each sub-item of the input business expenses and asset-related information of the plurality of businesses (S302). Next, the group generation unit 6 generates a plurality of clusters by clustering based on the calculated business costs of the plurality of businesses and a plurality of composition ratios for each sub-item of the asset-related information, and each of the generated plurality of clusters is generated. , And a plurality of the determined groups are output to the index calculation unit 4 and the display device 29 (S303).

FIG. 4 is a flow chart of the group generation process according to the first embodiment.

In the group generation process, first, the group generation unit 6 divides a plurality of businesses into elements of the whole set, and divides the whole set into a plurality of subsets based on the business costs and asset-related information of the plurality of businesses. Is executed to generate a plurality of clusters to which one or more operators belong (S401). Next, the group generation unit 6 defines the generated plurality of clusters as a group (S402). Next, the group generation unit 6 sets a group name for each of the defined groups (S403).

At this time, the group generation unit 6 has the composition ratio of each cost item in the cost item group representing the group and each of the asset items in the asset item group based on the asset-related information of the business operator belonging to each group. The composition ratio of the above may be graphically displayed on the display device 29, and the user may be prompted to input the group name of the group. As the composition ratio of the cost items in the cost item group representing the group, the average value of the composition ratios of the businesses belonging to the group may be used. The user can set a group name that represents the characteristics of the cost and the asset by referring to the composition ratio of the business cost and the composition ratio of the asset.

FIG. 5 is a diagram showing an example of an output screen according to the first embodiment.

The business costs of the plurality of businesses calculated by the group generation unit 6 and the plurality of composition ratios for each sub-item of the asset-related information are displayed on the display device 29. For each group, the ratio of each sub-item of project cost (right side in FIG. 4) and asset-related information (left side in FIG. 4) is displayed as a pie chart. In the group illustrated in FIG. 5, the ratio of personnel expenses is relatively large among the sub-items related to business expenses, and the ratio of cash and deposits is relatively large among the sub-items related to asset-related information. Therefore, this group is a "humanoid" business form (large group) with people as the source of business. The business form may include a "tangible asset type" in which tangible assets are the source of business and a "mono type" in which goods are the source of business.

FIG. 6 is a flow chart of the group name setting process according to the first embodiment.

In the group name setting process, the group generation unit 6 sets a plurality of clusters that are similar to each other into three large groups among the generated plurality of clusters. For example, in the large group, the ratio of manufacturing cost or product purchase is relatively high, and the ratio of labor cost is relatively high "human type" and the ratio of depreciation cost is relatively high "tangible asset type". It may be a large group with many "mono-types".

The group generation unit 6 may set a plurality of clusters similar to each other in a plurality of small groups or medium groups among the clusters included in the large group. Specifically, the group generation unit 6 sets a cluster having a relatively large proportion of cash and deposits among the clusters included in the large group belonging to the "human type" as a small group of the "group 1", and also sets the cluster to the small group of the "group 1". Clusters with relatively low cash and deposit ratios may be set as small groups in "Group 7". The group generation unit 6 may set a cluster included in a large group belonging to the “tangible asset type” as a small group of the “group 2”. The group generation unit 6 sets the clusters included in the large group belonging to the "monotype" to the middle group of "make and sell", which has a relatively large ratio of manufacturing cost, and the ratio of product purchase. Set a cluster with a relatively large number of items in the middle group of "Purchase and sell".

Further, the group generation unit 6 may set the cluster included in the middle group belonging to "make and sell" as a small group of "group 3". The group generation unit 6 sets the clusters included in the middle group belonging to "purchase and sell" with a relatively large proportion of trade receivables as a small group of "group 4", and the proportion of cash and deposits is high. A relatively large number of clusters may be set as a small group of "Group 5", and a cluster that does not correspond to any of them may be set as a small group of "Group 6". "Group 4" and "Group 5" are wholesale groups, and "Group 6" is a retail group. As a result, the group generation unit 6 can set a plurality of similar clusters into seven small groups among the generated plurality of clusters.

FIG. 7 is a flow chart of the group update process according to the first embodiment.

When the defined group is distorted, the group generation unit 6 of this embodiment executes a group update process for updating the group.

In the group update process, first, the group generation unit 6 redefines each of the plurality of clusters as a new group in the background (S701). Next, the group generation unit 6 determines whether or not the distortion of the plurality of new groups determined again with respect to the plurality of predetermined groups already determined is equal to or greater than a predetermined threshold value (S702). If the determination in S702 is true (S702: YES), the group generator 6 suggests to the user that the group should be updated (S703). When the determination in S702 is false (S702: NO), the group generation unit 6 waits until the distortion of the plurality of redefined groups with respect to the plurality of predetermined groups already determined becomes equal to or greater than a predetermined threshold value. The user determines whether or not to accept the proposal to update the group (S704). When the judgment of S704 is true (S704: YES), the group generation unit 6 updates the plurality of predetermined groups already determined to the plurality of new groups defined again in the background. If the determination in S704 is false (S704: NO), the group generator 6 waits until the user accepts the proposal to update the group.

According to this embodiment, the business operator classification device 3A has a related information acquisition unit 5 and a group generation unit 6. The related information acquisition unit 5 acquires business-related information, which is information including a plurality of items related to each of the plurality of business operators. The group generation unit 6 uses a plurality of businesses as elements of the whole set, executes clustering that divides the whole set into a plurality of subsets based on the business-related information, and executes clustering to which one or more businesses belong. Is generated, and each of the generated clusters is defined as a group. As a result, the business operator classification device 3A distributes the entire set including the plurality of business operators into a plurality of subsets by clustering based on the information related to each of the business operators, and classifies the generated clusters into the business operators. It can be defined as a group to do. Therefore, businesses can be appropriately classified according to the framework of the group based on business-related information.

The business operator-related information includes a plurality of items that numerically represent the activities performed by the business operator, and the group generation unit 6 calculates the composition ratio of each of the plurality of items in the item group including the plurality of items. Perform clustering with the composition ratio as a variable. As a result, the business classification device 3A executes clustering with the composition ratio of each of the plurality of items in the item group as a variable for each business, so that the influence of the difference in the scale of the business is reduced and the activities are similar. Depending on the degree, it is possible to generate an appropriate group that reflects the nature of the business's activities. Therefore, the business operator evaluation system 1A can improve the evaluation accuracy of business operators when comparing business operators clustered in the same group among appropriately classified business operators.

The item group includes a cost item group that includes a plurality of cost items that are items related to money spent by the business operator, and an asset item that includes a plurality of asset items that represent the assets owned by the business operator in monetary terms. There is a group. As a result, by focusing on the expenditure and savings of the business operator, it is possible to grasp the characteristics of the business operator and perform good clustering that reflects the constitution of the business operator.

The group generation unit 6 calculates each composition ratio of the cost item in the cost item group and each composition ratio of the asset item in the asset item group, and each composition ratio of the cost item in the cost item group and the asset item. Clustering is executed with the composition ratio of each asset item in the group as a variable. As a result, clustering is executed with the business operator using the composition ratio of each item in the cost breakdown and the composition ratio of each item in the asset breakdown as variables, so that an appropriate group that reflects the nature of the cost composition and asset composition can be obtained. Can be generated.

As the cost item and the asset item, the account of the journal information or the item which is the sum of multiple accounts of the journal information is used. As a result, clustering is performed using the items obtained from the detailed information collected as the journal information, so that an appropriate group can be generated.

Since cost items include current product purchases, manufacturing costs, labor costs, and depreciation costs, expenditures that represent the characteristics of businesses are used as clustering variables, enabling highly accurate clustering. Become.

Asset items include cash and deposits, trade receivables, inventories, other current assets, land and buildings, and fixed assets. Highly accurate clustering is possible.

When the related information acquisition unit 5 acquires the business-related information of the new business operator, the group distribution unit 7 that distributes the group to which the new business operator belongs from among the plurality of default groups already determined by the group generation unit 6 is set. Have more. This makes it possible to classify new businesses into appropriate groups.

When the group generation unit 6 redefines each of the plurality of clusters as a new group in the background, if the distortion of the new group with respect to the plurality of default groups already determined is equal to or greater than a predetermined threshold value, the group generation unit 6 has already determined the plurality of defaults. Update the group to multiple new groups that have been redefined. As a result, when the distortion of the plurality of groups is large, the plurality of groups can be updated.

The business operator evaluation system 1A includes a database 2 that records accounting-related information on the cloud, and a business operator classification device 3A that has a related information acquisition unit 5 that acquires accounting-related information from the database 2. It can be managed on the cloud and can be easily used for business classification. The database 2 may be various local devices.

The business operator evaluation system 1B according to the second embodiment will be described. Each of the following examples including this example corresponds to a modified example of the first embodiment. The business operator evaluation system 1B according to the second embodiment is different from the business operator evaluation system 1A according to the first embodiment only in the configuration of the related information acquisition unit, and the other configurations are the business operator evaluation according to the first embodiment. It is the same as the system 1A. Therefore, the differences from the first embodiment will be mainly described.

FIG. 8 is an overall view of the business operator evaluation system according to the second embodiment.

The related information acquisition unit 5 of the business operator classification device 3B of this embodiment has an item proposal unit 5a that proposes a sub-item of the business operator-related information acquired from the database 2 to the user.

According to this embodiment, since the related information acquisition unit 5 has an item proposal unit 5a that proposes a sub-item of the business operator-related information to be acquired to the user, the user can arbitrarily select the information to be used for clustering. , User's knowledge and experience can be reflected in the classification result of the business operator.

The business operator evaluation system 1C according to the third embodiment will be described. The business operator evaluation system 1C according to the third embodiment differs from the business operator evaluation system 1A according to the first embodiment only in the configuration of the index calculation unit and the first model generation unit, and the other configurations are the same as in the first embodiment. This is the same as the business evaluation system 1A according to the above. Therefore, the differences from the first embodiment will be mainly described.

FIG. 9 is an overall view of the business operator evaluation system according to the third embodiment.

The business operator evaluation system 1C of the present embodiment further includes a first model generation unit 9 as compared with the business operator evaluation system 1A according to the first embodiment. In the first model generation unit 9, the business operator-related information of the business operator is input from the related information acquisition unit 5, and a plurality of groups are input from the group generation unit 6. The first model generation unit 9 is a score as an example of an "index calculation model" for calculating a score as an example of an "index" regarding a business operator based on business operator-related information of a business operator for each of a plurality of groups. A calculation model is generated for each group.

The index calculation unit 4C inputs the business operator-related information of the business operator into the score calculation model generated by the first model generation unit 9, calculates the score related to the business operator, and outputs the score to the display device 29. This allows the user to check the score for the business operator.

According to this embodiment, the business operator evaluation system 1C has a first model generation unit 9 and an index calculation unit 4C. The first model generation unit 9 uses a score calculation model for calculating a score for a business operator for each of the plurality of groups based on the business operator-related information of the business operators for each of the plurality of groups set by the group generation unit 6. Generate for each group. The index calculation unit 4C inputs the business operator-related information of the business operator of the business operator into the score calculation model generated by the first model generation unit 9, and calculates the score of the business operator. As a result, the score for the business operator can be easily calculated, and the business operator can be appropriately evaluated based on the calculated score for the business operator.

The credit risk evaluation system 11 according to the fourth embodiment will be described. The credit risk evaluation system 11 according to the fourth embodiment differs from the business operator evaluation system 1A according to the first embodiment only in the configuration of the second model generation unit and the credit risk score calculation unit, and the other configurations are implemented. This is the same as the business evaluation system 1A according to Example 1. Therefore, the differences from the third embodiment will be mainly described.

FIG. 10 is an overall view of the credit risk evaluation system according to the fourth embodiment.

The credit risk evaluation system 11 of this embodiment is a system for evaluating the credit risk of businesses clustered in the same group. Compared to the business operator evaluation system 1A according to the first embodiment, the credit risk evaluation system 11 further includes a second model generation unit 10 and a credit risk score calculation unit 12 as an example of the “credit risk index calculation unit”. There is. In the second model generation unit 10, the business operator-related information of the business operator is input from the related information acquisition unit 5, and a plurality of groups are input from the group generation unit 6. The second model generation unit 10 is a credit risk score as an example of a “credit risk index calculation model” for calculating a credit risk score for a business operator based on business operator-related information of a business operator for each of a plurality of groups. A calculation model is generated for each group.

The credit risk score calculation unit 12 inputs the business operator-related information of the business operator into the credit risk score calculation model generated by the second model generation unit 10, and calculates the credit risk score for the business operator.

According to this embodiment, the credit risk evaluation system 11 has a second model generation unit 10 and a credit risk score calculation unit 12. The second model generation unit 10 is used to calculate the credit risk score for the business operators in each of the plurality of groups based on the business operator-related information of the business operators in each of the plurality of groups set by the group generation unit 6. A risk score calculation model is generated for each group. The index calculation unit 8 inputs the business operator-related information of the business operator into the credit risk score calculation model generated by the second model generation unit 10, and calculates the credit risk score of the business operator. As a result, the credit risk score for the business operator can be calculated, and the credit risk of the business operator can be appropriately evaluated based on the calculated credit risk score for the business operator.

The business operator evaluation system 1E according to the fifth embodiment will be described. Each of the following examples including this example corresponds to a modified example of the first embodiment. The business operator evaluation system 1E according to the fifth embodiment is different from the business operator evaluation system 1A according to the first embodiment only in the configuration of the group update process, and the other configurations are the business operator evaluation system according to the first embodiment. It is the same as 1A. Therefore, the differences from the first embodiment will be mainly described.

FIG. 11 is a flow chart of the group update process according to the fifth embodiment.

The group generation unit 6 of the present embodiment executes the group update process when the ratio of the business operators who have moved to another group is large.

In the group update process, first, the group generation unit 6 resets a plurality of groups in the background (S1101). Next, the group generation unit 6 determines whether or not the ratio of businesses that have moved to another group is equal to or greater than a predetermined threshold value (S1102). Next, the group generation unit 6 executes S1103 to S1105 corresponding to S703 to S705 of the first embodiment.

According to this embodiment, the group generation unit 6 is set when the ratio of the business operators who have moved from the plurality of groups set when the plurality of groups are reset in the background to another group is equal to or more than a predetermined threshold value. Update multiple groups that have been set up to the reset group. This makes it possible to update a plurality of groups when the percentage of businesses that have moved to other groups is high.

The above-described examples of the present invention are examples for the purpose of explaining the present invention, and the scope of the present invention is not limited to those examples. One of ordinary skill in the art can practice the present invention in various other aspects without departing from the scope of the present invention.

For example, in each of the above embodiments, the cost item and the asset item included an account of journal information. Not limited to this, at least one of the cost or asset items may include an account of the item or journal information contained in the published financial statements. As a result, since the financial statement item and the journal account item are used as at least one of the cost item or the asset item, the cost composition and the asset composition are appropriately used by the large classification item and the small classification item. It is possible to generate an appropriate group that reflects the properties of.

For example, in each of the above-described examples, the business classification devices 3A to 3D generated a plurality of clusters by one clustering. Not limited to this, the business classification devices 3A to 3D may generate a plurality of clusters by clustering a plurality of times based on different business-related information each time. Specifically, the business classification devices 3A to 3D generate a plurality of clusters by clustering based on the financial information included in the published financial statements among the business-related information, and the generated multiple clusters are grouped. Set. Next, the business operator classification devices 3A to 3D add or subtract a plurality of journal information according to the purpose for each set group, and a plurality of subclusters for each group set by clustering based on the added or subtracted journal information. May be generated and a plurality of generated subclusters may be set in the subgroup. As a result, for example, a cluster in which the business operator is not sufficiently divided by the first clustering based on financial information is further divided by the second clustering based on journal information, so that the group in which the business operator is appropriately divided while suppressing the amount of calculation. Can be determined.

For example, in each of the above embodiments, the business operators are set to 7 groups. Not limited to this, the business operators may be set to two or more groups.

1A, 1B, 1C Business evaluation system, 2 ... Database, 3A ... Business classification device, 3B ... Business classification device, 3C ... Business classification device, 5 ... Related information acquisition unit, 5a ... Item proposal department, 5 ... Composition ratio calculation unit, 6 ... group generation unit, 7 ... group distribution unit, 8C ... index calculation unit, 9 ... 1st model generation unit, 10 ... second model generation unit, 11 ... credit risk assessment system, 12 ... credit Risk calculation unit, 29 ... Display device

Claims (21)

The related information acquisition department that acquires business-related information, which is information that includes multiple items related to each of multiple businesses,
Using the plurality of businesses as elements of the total set, clustering that divides the entire set into a plurality of subsets is executed based on the business-related information, and a plurality of clusters to which one or more businesses belong are generated. , A group generation unit that defines each of the generated clusters as a group for classifying businesses,
Business classification device with. The business-related information includes a plurality of items that numerically represent the activities performed by the business.
The group generation unit calculates the composition ratio of each of the plurality of items in the item group including the plurality of items, and executes the clustering using the composition ratio as a variable.
The business classification device according to claim 1. The item group includes a cost item group including a plurality of cost items which are items related to money spent by the business operator, and an asset item including a plurality of asset items which are items representing the assets owned by the business operator in monetary terms. There is a group,
The business classification device according to claim 2. The group generation unit calculates the composition ratio of each of the cost items in the cost item group and the composition ratio of each of the asset items in the asset item group, and each of the cost items in the cost item group. The clustering is executed with the composition ratio and the composition ratio of each of the asset items in the asset item group as variables.
The business classification device according to claim 3. The cost item and / or the asset item includes an item contained in the financial statements and / or an account of journal information.
The business classification device according to claim 4. As the cost item and the asset item, an account item of the journal information or an item obtained by summing a plurality of accounts of the journal information is used.
The business classification device according to claim 5. The cost items include current product purchases, manufacturing costs, labor costs, and depreciation costs.
The business classification device according to claim 6. The asset items include cash and deposits, trade receivables, inventories, other current assets, land and buildings, and fixed assets.
The business classification device according to claim 6. The cost item and the asset item are items included in the financial statements.
The group generation unit executes the first clustering with the composition ratio of each of the cost items in the cost item group and the composition ratio of each of the asset items in the asset item group as variables, and a plurality of generated units. The second clustering is executed for one or more of the clusters with the account item of the journal information as a variable, and among the plurality of clusters generated by the first clustering, the clusters that are not the target of the second clustering and the clusters. The cluster generated by the second clustering is defined as the group.
The business classification device according to claim 4. When the related information acquisition department acquires business-related information of a new business operator, a group distribution department that distributes the group to which the new business operator belongs from among a plurality of default groups already defined by the group generation department is assigned. Have more
The business classification device according to claim 7. The related information acquisition unit has an item proposal unit that proposes items of business-related information to be acquired to the user.
The business classification device according to claim 7. The group generation unit has already been determined when a plurality of clusters to which one or more businesses belong are generated in the background and each of the plurality of clusters is redefined as a new group for reclassifying the businesses. When the distortion of the new group with respect to the plurality of default groups is equal to or greater than a predetermined threshold value, the plurality of default groups are updated to the plurality of new groups.
The business classification device according to claim 7. The group generation unit has already been determined when a plurality of clusters to which one or more businesses belong are generated in the background and each of the plurality of clusters is redefined as a new group for reclassifying the businesses. When the ratio of businesses that have moved from a plurality of default groups to another default group is equal to or greater than a predetermined threshold value, the plurality of default groups are updated to the plurality of new groups.
The business classification device according to claim 7. The group generation unit generates a plurality of groups by applying the latent Dirichlet allocation method to the business-related information.
The business classification device according to claim 7. The business operator classification device according to any one of claims 1 to 14 and
First model generation for generating an index calculation model for calculating an index related to the business operator for each of the plurality of groups based on the business operator-related information for each of the plurality of groups defined by the group generation unit. Department and
It is provided with an index calculation unit for inputting business-related information of the target business operator into the index calculation model generated by the first model generation unit and calculating the index for the target business operator.
Business evaluation system. Further having a display unit for prompting the user to input the group name of the group.
The business operator evaluation system according to claim 15. Equipped with a database on the cloud that records the business-related information
It has the related information acquisition unit that acquires the business related information from the database.
The business operator evaluation system according to claim 15. The business operator classification device according to any one of claims 1 to 14 and
Based on the business operator-related information for each of a plurality of groups defined by the group generation unit, a credit risk index calculation model for calculating a credit risk index for the business operator for each of the plurality of groups is generated for each group. 2nd model generator and
The credit risk index calculation unit generated by the second model generation unit is provided with a credit risk calculation unit for inputting business-related information of the target business operator and calculating the credit risk index for the target business operator. , Credit risk assessment system. It also has an output unit that displays the credit risk index.
The credit risk assessment system according to claim 18. The related information acquisition department provided in the computer acquires the business operator-related information, which is information including a plurality of items related to each of the multiple business operators.
The group generation unit provided in the computer uses the plurality of businesses as elements of the whole set, executes clustering that divides the whole set into a plurality of subsets based on the business-related information, and executes clustering to divide the whole set into a plurality of subsets. A business operator classification method in which a plurality of clusters to which a company belongs are generated, and each of the generated multiple clusters is defined as a group for classifying business operators. Have the related information acquisition department provided in the computer acquire business operator related information, which is information including a plurality of items related to each of the plurality of target business operators.
A group generation unit provided in a computer is made to have the plurality of businesses as elements of a whole set, and based on the business-related information, clustering is performed to distribute the whole set to a plurality of subsets, and one or more businesses. A business operator classification program that generates multiple clusters to which a person belongs and defines each of the generated clusters as a group for classifying business operators.

Images