KR20190125067A - Apparatus For Extracting Similar Enterprise And Method For Extracting Similar Enterprise Using The Same - Google Patents

Abstract

Disclosed are an apparatus for extracting a similar company and a method for extracting a similar company using the same. The apparatus of the present invention comprises: a communication unit transmitting and receiving data whether a transaction is made with an external first server and transmitting and receiving transaction history data with an external second server; and a control unit electrically connected to the communication unit and extracting at least one primary customer of an interesting company and at least one similar company of the interesting company based on the received transaction history data. The communication unit may transmit a list of similar companies to an external terminal when the control unit extracts the similar companies.

Description

Apparatus For Extracting Similar Enterprise And Method For Extracting Similar Enterprise Using The Same}

The present invention relates to an apparatus for recognizing a pattern of a company's transaction information and recognizing a company of a similar category as a company of interest and a method for recognizing a similar company using the same.

Recognizing similar categories of companies that do business similar to theirs in corporate activities is an important task in identifying a competitive environment, generating new business opportunities, and identifying risks. Accordingly, public institutions or market research organizations provide the basis for companies to recognize companies in similar industries by providing statistical data by industry, industry, and region as well as standard industry classification.

However, in recent years, the business environment of companies is complicated by the convergence and combination of industrial technologies. Accordingly, the data provided by public institutions or market research institutes may be limited data that can not reflect the market situation immediately. have.

Accordingly, the necessity of using transaction data such as an electronic tax invoice in establishing a corporate management strategy is increasing. Transaction data can confirm the real-time transaction information between companies and can be used as a company's management decision-making means as important information with timeliness. In particular, the analysis of transaction data to recognize companies with similar business structures and to establish a management strategy through flexible competition analysis can be very helpful in corporate activities.

An object of the present invention is to provide a method for extracting similar companies that can be used to establish a management strategy for realizing the recognition of the competitive environment of companies, new business opportunities and identification of business risks using transaction data in real time. do.

It is another object of the present invention to provide a method of improved accuracy for the extraction of a company similar to the company of interest.

Transaction data-based similar company extraction apparatus of the present invention for achieving the above object, the communication unit for transmitting and receiving transaction data with the external first server and the transaction history data with the external second server; And a control unit electrically connected to the communication unit and extracting at least one primary account of a company of interest and at least one similar company of the company of interest based on the received transaction history data. When the controller extracts the similar company, the controller may transmit the list of similar companies to an external terminal.

The controller may generate an input factor group based on the transaction history data, and extract the similar company based on the similarity of the input factor group.

The transaction history data may include a transaction item name that is atypical data, and the controller may form a transaction item group name through morphological analysis and clustering of the transaction item name to include in the input factor group.

The control unit may be configured to extract at least one second trading partner that deals with the first trading client based on the extracted transaction history data of the first trading partner, and to extract the similar company from the second trading partner. It is preferable.

On the other hand, it is preferable that the first client is a primary sub node having a direct transaction relationship with the interested company in the list of value chain companies having the interested company as the master node, or selected from among the primary sub nodes.

The primary subnode is divided into a front subnode located at the front of the master node and a rear subnode located at the rear of the master node, and the control unit includes the front subnode and the rear of the second account. A company having a trading relationship with a sub node at the same time may be extracted as the most similar company having a priority among the similar companies.

The control unit extracts a candidate node associated with the master node from the transaction presence data and the transaction history data, requests transaction details data of the candidate node from the second server, and requests transaction details data of the candidate node. The input factor group regarding the candidate node may be extracted by receiving from the second server. In addition, the control unit extracts the relevance of the candidate node from the extracted input factor group, and extracts sub-nodes from the candidate nodes based on the relevance to form the value chain company list

The control unit may extract a list of value chain companies through machine learning including updating the learning data one or more times based on the received transaction history data, wherein the learning data includes a plurality of learning nodes and the learning nodes. Each learning input factor group corresponding to each of the control units, the control unit requests the transaction history data corresponding to the learning node to the second server, and extracts the learning input factor group based on the received transaction history data. It is desirable to.

Further, the controller generates a value chain graph based on the transaction presence data, the transaction history data, and the value chain company list, and the communication unit generates the value chain graph on the external terminal when the controller generates the value chain graph. A graph, wherein the value chain graph comprises: a plurality of nodes; A plurality of lines connecting the plurality of nodes, the plurality of nodes may include the master node and the sub-node.

The control unit generates the extracted similar company and a line connecting the similar company and the first customer to the value chain graph, wherein the line connecting the similar company and the first customer is a plurality of lines. In contrast to the plurality of lines connecting the nodes, it is desirable to vary at least one of color, thickness and end type.

An extraction method of a similar company using a similar company extracting apparatus including a communication unit and a control unit according to an embodiment of the present invention may include: receiving an extraction request of the similar company of a company of interest from an external terminal; Extracting, by the control unit, at least one primary trading partner of the interested company when the communication unit receives transaction status data of the master node from an external first server; Extracting, by the control unit, at least one second trading partner trading with the first trading partner when the communication unit receives transaction history data of the first trading partner from an external second server; The control unit may be configured to include the step of extracting the similar company from the second account.

In addition, the first client may be a primary sub node having a direct transaction relationship with the interested company in the list of value chain companies having the interested company as the master node, or selected from among the primary sub nodes.

According to various embodiments of the present invention, by using the transaction data in real time to recognize similar companies that can be used to establish the company's management strategy, it is advantageous to recognize the competitive environment, to derive new business opportunities and to identify business risks. Can provide.

Furthermore, in recognizing companies similar to the companies of interest, it may provide a method with improved accuracy.

1 is a schematic diagram of a similar company extraction apparatus according to an embodiment of the present invention.
2 is a schematic diagram showing a trading relationship between a company of interest, a primary customer and a similar company of the present invention.
3 is a diagram illustrating an example of a value chain related to the present invention.
4 is a diagram illustrating an example of a value chain related to the present invention.
5 is a flowchart illustrating a value chain extraction process according to an embodiment of the present invention.
6 is a data structure diagram of learning data according to an embodiment of the present invention.
FIG. 7 illustrates a graph visualizing correlations between nodes in a value chain according to an embodiment of the present invention. FIG.

The following merely illustrates the principles of the invention. Therefore, those skilled in the art, although not explicitly described or illustrated herein, can embody the principles of the present invention and invent various devices that fall within the spirit and scope of the present invention. In addition, all conditional terms and embodiments listed herein are in principle clearly intended to be understood only for the purpose of understanding the concept of the invention and are not to be limited to the specifically listed embodiments and states. do.

In addition, it is to be understood that all detailed descriptions, including the principles, aspects, and embodiments of the present invention, as well as listing specific embodiments, are intended to include structural and functional equivalents of these matters. In addition, these equivalents should be understood to include not only equivalents now known, but also equivalents to be developed in the future, that is, all devices invented to perform the same function regardless of structure.

Thus, for example, the block diagrams herein should be understood to represent a conceptual view of example circuitry embodying the principles of the invention. Similarly, all flowcharts, state transitions, pseudocodes, and the like are understood to represent various processes performed by a computer or processor, whether or not the computer or processor is substantially illustrated on a computer readable medium and whether the computer or processor is clearly shown. Should be.

The functionality of the various elements shown in the figures, including functional blocks represented by a processor or similar concept, can be provided by the use of dedicated hardware as well as hardware capable of executing software in association with appropriate software. When provided by a processor, the functionality may be provided by a single dedicated processor, by a single shared processor or by a plurality of individual processors, some of which may be shared.

In addition, the explicit use of terms presented in terms of processor, control, or similar concept should not be interpreted exclusively as a citation to hardware capable of running software, and without limitation, ROM for storing digital signal processor (DSP) hardware, software. (ROM), RAM, and non-volatile memory are to be understood to implicitly include. Other hardware for the governor may also be included.

In the claims of this specification, components expressed as means for performing the functions described in the detailed description include all forms of software including, for example, a combination of circuit elements or firmware / microcode, etc. that perform the functions. It is intended to include all methods of performing a function which are combined with appropriate circuitry for executing the software to perform the function. The invention, as defined by these claims, is equivalent to what is understood from this specification, as any means capable of providing such functionality, as the functionality provided by the various enumerated means are combined, and in any manner required by the claims. It should be understood that.

The above objects, features, and advantages will become more apparent from the following detailed description taken in conjunction with the accompanying drawings, whereby those skilled in the art may easily implement the technical idea of the present invention. There will be. In addition, in describing the present invention, when it is determined that the detailed description of the known technology related to the present invention may unnecessarily obscure the gist of the present invention, the detailed description thereof will be omitted.

Hereinafter, various embodiments of the present invention will be described in detail with reference to the accompanying drawings.

1 is a schematic diagram of a similar company extraction apparatus according to an embodiment of the present invention. Referring to FIG. 1, an apparatus for extracting a value chain according to an embodiment of the present invention may be a main server 100.

The main server 100 may be a server that builds modeling and extracts a value chain. The main server 100 may be connected to the first server 200 and / or the second server 300 by wire and / or wirelessly. The main server 100 may include a controller. The main server 100 may include a communication unit capable of communicating with the outside.

The first server 200 may hold and manage information regarding the presence or absence of a transaction of a specific company. The first server 200 may be, for example, a bank, a credit rating company, a company transaction information distribution company, an investment banking company, or the like. The communication unit of the main server 100 may communicate with the first server 200 to transmit and receive data.

The second server 300 may hold and manage transaction history data of a specific company, that is, information on transaction history. The second server 300 may be, for example, the IRS or the Statistics Office. For another example, the second server 300 may be integrated with the first server 200. The communication unit of the main server 100 may communicate with the second server 300 to transmit and receive data.

The terminal 400 may be connected to the main server 100 by wire or / and wirelessly. The terminal 400 may be a computer or a mobile terminal. For example, the terminal 400 may communicate with the main server 100 using an application. The communication unit of the main server 100 may communicate with the terminal 400 to transmit and receive data.

2 is a schematic diagram showing a trading relationship between a company of interest, a primary customer and a similar company of the present invention. As shown in FIG. 2, the control unit may extract a first accounting company that is at least one or more companies of interest and a similar company that is at least one or more companies of interest based on the received transaction history data. When the controller extracts the similar company, the controller may transmit the list of similar companies to an external terminal. In an embodiment of the present invention, the term “similar company” means a company having a similar business structure based on a company of interest and a transaction history, and a company that is in competition with the company of interest or may be included in the same industry group. Can be.

The controller may generate an input factor group based on the transaction history data, and extract the similar company based on the similarity of the input factor group. The transaction history data may include a transaction item name that is atypical data, and the controller may form a transaction item group name through morphological analysis and clustering of the transaction item name to include in the input factor group. The transaction history data may be, for example, a tax invoice. The input factor group may include at least one of a number of transactions, a transaction amount, a transaction period, a sales amount, a regional distance, a transaction dependency rate, a transaction monthly transaction amount, and an item synchronization index. The 'similarity' is the so-called structural equivalence, that is, the degree of overlap of elements or elements contained therein. Cosine similarity or jacquard similarity, as well as Pearson correlation coefficient may be used. In addition, as described below, a company in the value chain having the interested company as the master node may be considered as a company that satisfies a certain level of similarity.

In addition, as shown in FIG. 2, the controller is configured to extract at least one second trading partner that deals with the first trading client based on the extracted transaction history data of the first trading client, and the second trading partner. It is desirable to extract the similar company from the customer. For example, if the first customer is a company that purchases a specific item from a company of interest (see the first customer (1) of FIG. 2), the company that sells a specific item to the buying company (similar company of FIG. 2 (1). ) May have a business structure similar to that of the company of interest. If the primary customer is a company that sells a specific item to a company of interest (see first account (2) in FIG. 2), see a company that buys a specific item from the selling company (similar company (2) in FIG. 2). ) There is also the possibility of having a business structure similar to the company of interest. By extracting similar companies from these secondary accounts, companies with high similarity in business structure can be selected. Such a method of extracting similar companies can be extended. In other words, there may be a third client dealing with a second client and a fourth client dealing with a third client. Among the fourth client, a company having a similar business structure to a company of interest may be derived. In this way, for the natural number n, the second-nth accounts can be candidates for similar companies. In this case, it is also possible to reduce the number of candidate groups and the number of operations by considering the direction of the transaction (item sale) in the process of selecting similar companies.

3 is a diagram illustrating an example of a value chain related to the present invention. Referring to FIG. 3, a value chain centering on a semiconductor manufacturing company appears. The central node in the value chain can be called the master node. For example, in the value chain of FIG. 1, the master node may be an X1 company or an X2 company, a semiconductor manufacturing company. In another example, in the value chain of FIG. 1, the master node may be an A1 company or an A2 company that handles semiconductor materials.

The arrow shown in Figure 3 may be referred to as the direction of delivery. For example, a semiconductor manufacturing company can be supplied with semiconductor materials and semiconductor equipment. For example, a semiconductor manufacturing company can supply semiconductors to computer companies, telecommunication equipment companies, and electronic set companies.

4 is a diagram illustrating an example of a value chain related to the present invention. As shown in FIG. 4, an example of a value chain related to the present invention may be divided into a front and a rear from a master node.

Nodes related to the master node may be referred to as subnodes. When the sub node is located in front of the master node, it may be referred to as a front sub node. The sub node may be referred to as the rear sub node if it is located behind the master node. The rear subnode may be an enterprise that delivers to the master node. The front subnode may be an enterprise that purchases a product from the master node.

Sub nodes may be classified into primary, secondary, and tertiary according to the degree of connection with the master node. For example, the primary subnodes F11, F12, F13, F14, R11, R12, R13, and R14 may be companies that directly deal with the master node. For example, the secondary subnodes R21, R22, R23, R24, F21, F22, F23, and F24 are master nodes through the primary subnodes F11, F12, F13, F14, R11, R12, R13, and R14. It is a company connected with.

In one embodiment of the present invention, the first account is a primary sub node having a direct transaction relationship with the interested company in the list of value chain companies having the interested company as a master node, or selected from among the primary sub nodes. It is preferable. In other words, by extracting the primary account from the subnodes in the list of value chain companies whose reliability has been verified in terms of relevance to the master node, it is possible to increase the accuracy of similar company extraction and provide a visualization service integrated with the value chain. Lose.

As illustrated in FIG. 4, the primary sub node is divided into a front sub node located in front of the master node and a rear sub node located behind the master node, and the control unit is configured as the second sub account. It is possible to extract a company having a trading relationship with the front sub-node and the rear sub-node simultaneously as the most similar company having a priority among the similar companies. In other words, a company that deals simultaneously with the front and rear primary customers of the secondary company among the secondary customers may be considered to have a core competitive relationship with the interested company or may belong to the same industry in terms of business structure.

5 is a flowchart illustrating a value chain extraction process according to an embodiment of the present invention. As shown in FIG. 5, the controller extracts a candidate node associated with the master node from the transaction presence data and the transaction history data, requests transaction details data of the candidate node from the second server, and The transaction history data of the candidate node may be received from the second server to extract an input factor group related to the candidate node. The control unit may extract a relevance of the candidate node from the extracted input factor group, and extract a sub node among the candidate nodes based on the relevance to form the value chain company list. The relevance may be a measure that includes the association with the main transaction item of the company that is the master node.

6 is a data structure diagram of learning data according to an embodiment of the present invention. As illustrated in FIG. 6, the learning data includes a plurality of learning nodes and a learning input factor group corresponding to each of the learning nodes, and the controller controls the transaction details data corresponding to the learning nodes to the second server. Request 300 and extract the learning input factor group based on the received transaction history data.

The control unit may extract a list of value chain companies through machine learning including updating the learning data one or more times based on the received transaction history data, wherein the learning data includes a plurality of learning nodes and the learning nodes. Each learning input factor group corresponding to each of the control units, the control unit requests the transaction history data corresponding to the learning node to the second server, and extracts the learning input factor group based on the received transaction history data. It is desirable to.

The machine learning includes a random forest technique, and the relevance of the candidate node may be a posterior probability that the candidate node is a sub node described below. That is, the machine learning may include a first random forest learning step. In this step, the learning data may have a posterior probability by the first random forest learning. The machine learning may also include learning data updating. At this stage, the learning data can be updated by adding or deleting new companies based on the existing learning data based on the post-probability value. The machine learning may include a second random forest learning step. The training data updated at this stage may have a posterior probability. A value chain company list may be derived through the second random forest learning step. At this stage, the list of value chain companies can be sorted according to their post-probability.

The process of updating the learning data by the control unit at least once may generate a degree of relevance of each candidate node from the extracted input factor group through the machine learning, and match any of the learning nodes among the candidate nodes. When the degree of relevance of the candidate nodes that do not exist is equal to or greater than a first reference value, the training data includes the non-matching candidate node as a new learning node and the input factor group corresponding to the mismatching candidate node as a new learning input factor group. Adding to and excluding the matching node and the learning input factor group corresponding to the matching node from the learning data when the relevance of the node matching one of the learning nodes is less than a second reference value. Process may be included. The update process of the learning data may improve the accuracy of the learning even when the learning data includes a transaction history which is not substantially meaningful, thereby contributing to deriving a reliable value chain company list. As an example of the method for setting the first reference value, it is possible to list the relevance of the candidate node in large order, and then set the relevance corresponding to a predetermined percentage of the upper level as the first reference value. The predetermined percentage may be designated as a value such as 30%, for example. Preferably, the first reference value is set higher than the second reference value. This can prevent excessive fluctuation of the training data and ensure a stable update.

FIG. 7 illustrates a graph visualizing a correlation between subnodes (SN1, SN2, SN3, SN4, SN5, SN6, etc.) and a master node MN derived by a value chain extraction method according to an embodiment of the present invention. Drawing. 7 may be referred to as a value chain graph.

Referring to FIG. 7, 'SNx' may mean a primary x subnode, and 'SNx, y' may mean a secondary y subnode connected to the primary x subnode.

'Node' may be understood as a general term for a master node and a sub node. Lines between nodes LN may provide information between nodes. For example, the thickness of the line LN between the nodes may be greater as the relation between the nodes increases. For example, an arrow that appears in the line LN between nodes may indicate a delivery direction. For example, the distance of the line LN between nodes may mean a geographical distance between nodes.

The graph illustrated in FIG. 7 may be a screen displayed on the terminal 400 illustrated in FIG. 1. Although not shown in the drawing, when a touch input is applied to the node, financial information related to the node may be displayed on the screen of the terminal 400.

The controller may generate the value chain graph based on the transaction presence data, the transaction history data, and the value chain company list.

In addition, the control unit is generated by including the extracted similar company and the line connecting the similar company and the first customer in the value chain graph, the line connecting the similar company and the first customer is the In contrast to the plurality of lines connecting the plurality of nodes, it is preferable to change at least one of color, thickness and end type. For example, SN6,2 of FIG. 7 may be a similar company of the MN (master node) of interest.

An extraction method of a similar company using a transaction data-based similar company extracting device including a communication unit and a control unit according to an embodiment of the present invention may include: receiving an extraction request of the similar company of a company of interest from an external terminal; Extracting, by the control unit, at least one primary trading partner of the interested company when the communication unit receives transaction status data of the master node from an external first server; Extracting, by the control unit, at least one second trading partner trading with the first trading partner when the communication unit receives transaction history data of the first trading partner from an external second server; The control unit may be configured to include the step of extracting the similar company from the second account.

In addition, the first client may be a primary sub node having a direct transaction relationship with the interested company in the list of value chain companies having the interested company as the master node, or selected from among the primary sub nodes.

Some or other embodiments of the invention described above are not exclusive or distinct from one another. Certain embodiments or other embodiments of the invention described above may be combined or combined in each configuration or function.

The above detailed description should not be construed as limiting in all respects but should be considered as illustrative. The scope of the invention should be determined by reasonable interpretation of the appended claims, and all changes within the equivalent scope of the invention are included in the scope of the invention.

100: primary server 200: first server
300: second server 400: terminal

Claims (12)

A communication unit which transmits and receives transaction presence data with an external first server and transmits and receives transaction detail data with an external second server; And
A controller electrically connected to the communication unit and extracting at least one primary account of a company of interest and at least one similar company of the company of interest based on the received transaction history data;
And the communication unit transmits the similar company list to an external terminal when the controller extracts the similar company. The method of claim 1,
The controller generates an input factor group based on the transaction history data, and extracts the similar company based on the similarity of the generated input factor group. The method of claim 2,
The transaction history data includes a transaction item name that is unstructured data,
The control unit
And forming a trade item group name through morphological analysis and clustering of the trade item name to include in the input factor group. The method of claim 2,
The control unit,
And extracting at least one second trading partner trading with the first trading client based on the extracted transaction history data of the first trading client, and extracting the similar company from the second trading partner. Industry extraction device. The method of claim 4, wherein
The first trading partner may be a primary sub node having a direct transaction relationship with the interested company in the value chain company list having the interested company as a master node, or a similar company extracting device selected from among the primary sub nodes. . The method of claim 5,
The primary sub node is divided into a front sub node located in front of the master node and a rear sub node located behind the master node,
The control unit,
And extracting a company having a trading relationship with the front sub-node and the rear sub-node among the second trading partners as the most similar company having priority among the similar companies. The method of claim 5,
The control unit,
Extracting candidate nodes associated with the master node from the transaction presence data and the transaction history data, requesting transaction details data of the candidate node from the second server, and requesting transaction details data of the candidate node from the second server; Receiving and extracting the input argument group,
And extracting a relevance of the candidate node from the extracted input factor group, and extracting sub-nodes from the candidate nodes based on the relevance to form the value chain company list. The method of claim 6,
The control unit,
Extracting a value chain company list through machine learning including the process of updating the learning data one or more times based on the received transaction history data,
The learning data includes a plurality of learning nodes and a learning input factor group corresponding to each of the learning nodes,
The controller may request the transaction details data corresponding to the learning node to the second server, and extract the learning input factor group based on the received transaction details data. . The method of claim 5,
The control unit
Generating a value chain graph based on the transaction presence data, the transaction history data, and the value chain company list,
The communication unit transmits the value chain graph to an external terminal when the controller generates the value chain graph.
The value chain graph
A plurality of nodes;
A plurality of lines connecting the plurality of nodes,
And the plurality of nodes includes the master node and the sub-node. The method of claim 9,
The control unit,
The extracted similar company and a line connecting the similar company and the first account are generated by being included in the value chain graph,
And a line connecting the similar company to the first customer differs in at least one of color, thickness, and end type from the plurality of lines connecting the plurality of nodes. In the extraction method of a similar company using a similar company extraction device including a communication unit and a control unit,
Receiving an extraction request of the similar company of the company of interest from an external terminal;
Extracting, by the controller, at least one primary trading partner of the interested company when the communication unit receives transaction presence data of the interested company from an external first server;
Extracting, by the control unit, at least one second trading partner trading with the first trading partner when the communication unit receives transaction history data of the first trading partner from an external second server;
And extracting, by the controller, the similar company from the second account. The method of claim 11,
The primary trading partner is a primary sub node having a direct transaction relationship with the interested company in the value chain company list having the interested company as a master node, or a similar company extraction method, characterized in that selected from the primary sub nodes. .

Images