JP6370434B1 - Company information provision system and program - Google Patents

Abstract

Provided is a company information providing system and program capable of easily performing a search operation of company information even by a user who is poor in search know-how and industry structure knowledge.
In a company information providing system 10, a company group / business attribute co-occurrence matrix X having a matrix element of the number of companies having each business attribute for each company group is used to determine each business attribute according to cosine similarity. The similarities between industries are calculated, and the similar business attributes with high similarity between industries are extracted from the target business attributes entered by the user, the target business attributes and the extracted similar business attributes, Create an extracted company list consisting of paired data that includes similar companies linked to these business attributes and the corporate group to which the similar companies belong, and use the extracted company list to show the relationship between the business groups and companies and the business attributes Output processing such as screen display is executed.
[Selection] Figure 1

Description

  The present invention relates to a company information providing system and program configured by a computer that provides company information. For example, for the purpose of analyzing the degree of similarity between companies in various ways in company research work, By using business segment classification labels, etc., it can be used when scoring the similarity of each company from the viewpoint of business composition.

  Conventionally, when conducting company analysis from the viewpoint of business structure in company research operations, search provided by each vendor for data held by company information vendors such as Reuter, Bloomberg, and Quick By accessing the web information using an interface or using an Internet search engine, we extracted target companies that seem to be closely related to the analysis target business.

  In addition, as a technique related to the present invention in that the information can be visually grasped and analyzed by performing an effective graphic display for the user, a query from the user for information is processed, A system that organizes and analyzes and displays related data in a graph for a user is known (see Patent Document 1).

Japanese translation of PCT publication No. 2002-532788 (FIG. 4)

  However, the extraction of target companies by the conventional method described above has the following problems. First of all, it is possible to grasp what kind of business a certain company is currently doing by using existing methods, but it is difficult to grasp potential needs for the analysis target business including future business development. In particular, beginners who do not have work experience are often able to extract target companies only by superficial information because their search know-how and industry structure knowledge are shallow.

  Moreover, in the existing method, the output of the search is a list format, and an overview of the entire target business and a multilateral comparison between companies require individual skills and experience. In particular, a beginner who does not have business experience can perform only one-sided prioritization, and therefore cannot extract not only companies that match the search conditions but also companies that are close to those companies.

  An object of the present invention is to provide a company information providing system and program that can easily perform a search operation of company information even for a user who is poor in search know-how and industry structure knowledge.

  In the present invention, a cosine similarity is calculated using a company group / business attribute co-occurrence matrix (hereinafter referred to as a cosine similarity method) and a latent dirichlet allocation method (LDA: Latency Dirichlet Allocation). There is a case where an operation is performed (hereinafter referred to as an LDA method).

  <Cosine similarity method>

The present invention is a company information providing system configured by a computer that provides company information,
A company information storage means for storing a company, a business attribute of the company, and a company group to which the company belongs;
Using a corporate group / business attribute co-occurrence matrix, where one of the rows or columns of the matrix is a corporate group, the other is a business attribute, and the number of companies with each business attribute for each corporate group is the matrix element The inter-industry similarity storage means for storing the inter-industry similarity indicating the similarity between the business attributes calculated by the similarity,
User input receiving means for executing processing for receiving user input including the target business attribute;
The number of business attributes similar to the target business attribute received by the user input receiving means, the number of cases received by the user input receiving means in the descending order of the industry similarity stored by the inter-industry similarity storage means, or a predetermined number of cases A similar business attribute extracting means for executing a process of extracting only as a similar business attribute,
The company having the target business attribute and the company group to which this company belongs, and the company having the similar business attribute extracted by the similar business attribute extracting means and the company group to which this company belongs from the company information storage means Similar company extraction means for executing a process of extracting as a company group to which the company belongs,
The group data including the target business attribute, the similar company extracted by the similar company extracting means, and the company group to which the similar company belongs, and the similar business attribute, the similar company extracted by the similar company extracting means, and the company group to which the similar company belongs An extraction company list creation means for executing a process of creating an extraction company list including the set data including;
Using the extracted company list created by the extracted company list creation means, screen display, print, three-dimensional display, or a combination thereof showing the relationship between the company group and the business attribute and / or the relationship between the company and the business attribute, Alternatively, output means for executing output processing by a combination of these and audio output is provided.

  Here, the “number of cases accepted by the user input acceptance means” as the number of extracted “similar business attributes” in the “similar business attribute extraction means” is the number obtained by subtracting 1 from the number actually inputted by the user (that is, the target (The number of cases excluding business attributes) or the number actually entered by the user. In the former case, since the user inputs and specifies the number of extractions including the target business attributes, the number obtained by subtracting one-tenth of the target business attributes is used as the number of extracted “similar business attributes”. For example, if the user inputs 6 cases, the number of extracted “similar business attributes” accepted by the system is 5 cases. In the latter case, since the user inputs and designates the number of extractions without including the target business attribute, the input number is directly used as the extraction number of “similar business attribute”.

  In addition, “screen display” is mainly a flat screen display, but when the screen of the display device (display) is curved or refracted, the display is in accordance with the screen shape of the display device. It does not necessarily have to be a plane. Further, the “stereoscopic display” may be a method with glasses or a method without glasses. The same applies to the following inventions.

  In such a company information providing system of the present invention, using the company group / business attribute co-occurrence matrix, the similarity between industries indicating the similarity between each business attribute is calculated based on the cosine similarity. Use similarities between industries to extract similar business attributes for the target business attributes entered by the user, create an extracted company list based on that, and then use the extracted company list to output company information. Even users with little know-how and knowledge of industry structure can easily search for company information, thereby achieving the object.

  In the above-described invention of the cosine similarity method, the preprocessing for calculating the similarity between industries indicating the similarity between the business attributes can be systemized as follows.

That is, in the above-described invention of the cosine similarity method,
A matrix storage means for storing a corporate group / business attribute co-occurrence matrix;
Using the company group / business attribute co-occurrence matrix stored in this matrix storage means, the similarity between industries indicating the similarity between each business attribute is calculated by cosine similarity, and the calculated similarity between industries is It can be set as the structure provided with the industry type similarity calculation means which performs the process memorize | stored in a between similarity storage means.

  When the matrix storage unit and the inter-industry similarity calculation unit are configured as described above, the data content stored in the company information storage unit is updated, and the content of the company group / business attribute co-occurrence matrix changes. Even when it occurs, it becomes possible to easily prepare a new similarity between industries according to the changed corporate group / business attribute co-occurrence matrix.

  <LDA method>

The present invention is a company information providing system configured by a computer that provides company information,
A company information storage means for storing a company, a business attribute of the company, and a company group to which the company belongs;
Using a corporate group / business attribute co-occurrence matrix, where one of the rows or columns of the matrix is a business group, the other is a business attribute, and the number of companies with each business attribute for each business group is a matrix element, An appearance probability storage means for storing the appearance probability of each business attribute in each company group, calculated by a dynamic Dirichlet distribution method,
User input receiving means for executing processing for receiving user input including the target business attribute;
For the target business attributes accepted by the user input accepting means, the number of cases received by the user input accepting means or the predetermined number of company groups having a high appearance probability in descending order of appearance probabilities stored in the appearance probability storage means A similar company group extracting means for executing a process of extracting as a similar company group;
For each similar company group extracted by the similar company group extraction means, the number of cases in which the business attributes having a high appearance probability are received by the user input reception means in the descending order of appearance probabilities stored in the appearance probability storage means or predetermined. Similar business attribute extraction means for executing a process of extracting as many similar business attributes as the number of cases,
From the company information storage means, out of the companies belonging to the similar company group extracted by the similar company group extraction means, companies having similar business attributes extracted by the similar business attribute extraction means are extracted as similar companies, and the extracted similar companies and similar An extracted company list creating means for executing a process of creating an extracted company list including group data including a similar company group to which the company belongs and similar business attributes;
Using the extracted company list created by the extracted company list creation means, screen display, print, three-dimensional display, or a combination thereof showing the relationship between the company group and the business attribute and / or the relationship between the company and the business attribute, Alternatively, output means for executing output processing by a combination of these and audio output is provided.

  In such a company information providing system of the present invention, the appearance probability of each business attribute in each company group is calculated by the latent Dirichlet distribution method (LDA) using the company group / business attribute co-occurrence matrix. Using this appearance probability, similar business attributes are extracted from the target business attributes entered by the user, an extracted company list is created based on the extracted business attributes, and the company information is output using the extracted company list. Even users with little know-how and knowledge of industry structure can easily search for company information, thereby achieving the object.

  Further, in the above-described LDA method invention, preprocessing for calculating the appearance probability of each business attribute in each company group can be systemized as follows.

That is, in the above-described LDA method invention,
A matrix storage means for storing a corporate group / business attribute co-occurrence matrix;
Using the company group / business attribute co-occurrence matrix stored in this matrix storage means, the occurrence probability of each business attribute in each company group is calculated by the latent Dirichlet distribution method, and the calculated appearance probability is stored as the appearance probability storage means. And an appearance probability calculating means for executing a process to be stored.

  In this way, when the configuration includes the matrix storage means and the appearance probability calculation means, the data contents stored in the company information storage means are updated, and the contents of the company group / business attribute co-occurrence matrix change. Even now, it becomes possible to easily prepare new appearance probabilities according to the changed corporate group / business attribute co-occurrence matrix.

  <Mode of output common to cosine similarity method and LDA method>

In the corporate information provision system described above,
The output means is
Using the extracted company list created by the extracted company list creation means, the company group node corresponding to the company group and the business attribute node corresponding to the business attribute are represented by the side indicating the existence of the relationship between the company group and the business attribute. It is desirable that the output processing by screen display, printing, or three-dimensional display by the network representation connected by the above is executed.

  In this way, when it is configured to output in the network representation in which the company group node and the business attribute node are connected by edges, the relationship between the company group and the business attribute is visualized as a network plot. It is possible to get a comprehensive overview of the industry structure in the peripheral field. For this reason, for example, it becomes possible to support the direction of the advisory of the corporate research work for the extracted companies.

  <Details of output mode>

In addition, in the case of a configuration that performs output by the network expression described above,
A company evaluation information storage means for storing a company and financial information, management indicators, or other company evaluation information about the company in association with each other;
The company evaluation information for each company group included in each set of data constituting the extracted company list created by the extracted company list creation means is acquired from the company evaluation information storage means, and the acquired company evaluation information is added to each set of data. Extraction company list editing means for executing the processing to be performed,
Using the extracted company list with the company evaluation information added by the extracted company list editing means, the size and color of the company group node corresponding to each company group included in each set of data constituting the extracted company list Output element creation means for executing processing for creating setting information and a node list including size and color setting information of the business attribute node corresponding to each business attribute included in each set of data constituting the extracted company list; With
Output element creation means
For the size of the company group node, execute a setting process to set the value after conversion using the company evaluation information for the company group,
Or, for the size of the enterprise group node corresponding to the enterprise group that constitutes the set data including the same business attribute as the target business attribute among the size of the enterprise group node, the enterprise evaluation information for the enterprise group is used. It is configured to execute the setting process to be a fixed value for the size of the company group node other than the converted value,
The output means is
It is desirable that the output processing by the network expression is executed according to the setting information of the node list created by the output element creating means.

  Here, the “company evaluation information” is information for evaluating the size and situation of the company, such as financial information and management indicators about the company. Specifically, for example, sales, market capitalization, Cash and equivalents, corporate value (EV), EBITDA (EBITDA) margin, sales growth rate, cash growth rate, etc. This “company evaluation information” is mainly objective numerical data, but it is subjective or subject matter mixed by a specific group (for example, an evaluation organization) or a specific person (for example, an economic critic). (For example, numerical data corresponding to ranking such as A, B, C,..., ◎, ○, Δ, ×,...) May be used. Further, it may be data collected from consumers (for example, numerical data obtained by collecting company brand preferences from a large number of consumers and collecting them).

  The “company evaluation information storage unit” is described separately from the “company information storage unit” for convenience of explanation, but may be included in the “company information storage unit”.

  Further, the “company evaluation information about each company group” added to each set of data constituting the extracted company list by the “extracted company list editing means” may be the company evaluation information about the parent company of the company group. Companies obtained by summing up or aggregating each company's evaluation information of parent companies and affiliated companies (which may be affiliated companies as long as they are included in the extracted company list or all affiliated companies included in the company information storage means) Evaluation information may be used. However, from the viewpoint of simplification of processing, it is preferable to use the company evaluation information about the parent company of the company group. In this case, for example, the determination as to whether or not the parent company is a parent company Information indicating the type of affiliated company may be stored, or the company name of the parent company (identification information of the parent company) may be used as the company group name (identification information of the company group). On the other hand, “total” when summarizing or aggregating each company's evaluation information of the parent company and affiliated companies is a process in which numerical values can be simply added, such as sales and market capitalization. For example, “aggregation” refers to processing when numerical values cannot be simply added, such as a weighted average with weight attached to the parent company and affiliated companies, for example. .

  Further, the “value after conversion using the company evaluation information” in the setting process by the “output element creation means” is the company evaluation information (in the case of a rank or the like, it is assumed to be a numerical data state). It is obtained by performing linear transformation (linear transformation) by a function (y = ax) multiplied by a factor a (this factor a is a factor determined in advance according to the numerical value indicating the company evaluation information). Or a value obtained by performing conversion using a linear function having a constant term (y = ax + b; a function that adds a constant term b to a proportional portion ax based on company evaluation information), or It may be a value obtained by performing conversion using a more complicated function. Note that the more complicated function is, for example, a factor (for example, 4 times) that a company group node having enterprise evaluation information having a numerical value twice as large as that of a company group node having enterprise evaluation information having a certain numeric value is greater than twice. A function that exaggerates the actual magnitude relationship of the numerical values of the company evaluation information so that it can be seen by the user, or on the contrary, relaxes the magnitude relationship so that it can be seen at a magnification smaller than 2 (for example, twice the root). It is. Further, the function is not limited to a function expressed by one simple mathematical expression, but may be a function expressed by a table. Furthermore, if the node becomes small enough that it is difficult to confirm with the naked eye by principle conversion alone, or conversely, it becomes too large compared to other company group nodes, etc., an exceptional conversion should be performed. Also good.

  In addition, in the setting process by the “output element creation means”, regarding the “other company group node sizes are fixed values”, the company group node size is changed using the company evaluation information (individual It means not to be set. Therefore, even if the overall size of the node is adjusted according to the screen size (number of pixels) of the display device (display), the number of dots of the printing device (printer), or the like, this is fixed here. Included in value.

  If the company group node is composed of characters (labels) indicating the name of the company group and figures (simple figures such as circles and squares, or company logos), both the letters and figures are displayed. Although conversion using company evaluation information is preferably performed, conversion using company evaluation information may be performed for either one of a character or a graphic. The company logo includes not only the official logo adopted by the company but also the logo created by the creator or user of this system or the company information provider stored in the company information storage means. included.

  In this way, when the size of the corporate group node is set using the corporate evaluation information, in addition to grasping the industry structure of the target business and its peripheral fields from a bird's-eye view, It is also possible to visually grasp the situation and the like.

Furthermore, in the case where the size of the company group node is set using the company evaluation information as described above,
User input acceptance means
When a plurality of types of company evaluation information are stored in the company evaluation information storage means, the process is configured to execute a process of accepting a user input for designating which company evaluation information is used,
Output element creation means
It is desirable that the setting process for calculating the converted value is executed using the company evaluation information specified by the user received by the user input receiving means.

  Here, in the “setting process for calculating the converted value using the company evaluation information” when the node list is created by the “output element creation means”, the extraction of the state to which the company evaluation information is added by the extraction company list editing means The company list is used. In this case, (1) the extracted company list includes only the company evaluation information related to the designation of the user received by the user input receiving means among the plurality of types of company evaluation information, and the output element The creation means may acquire it and use it for conversion, or (2) Multiple types of company evaluation information is added to the extracted company list, and the output element creation means specifies the user's designation The company evaluation information according to the above may be selectively acquired and used for conversion. In the former case (1), the extracted company list editing means acquires only the company evaluation information related to the user designation received by the user input receiving means from the company evaluation information storage means and adds it to the extracted company list. In the latter case (2), the extracted company list editing means acquires a plurality of types of company evaluation information from the company evaluation information storage means and adds it to the extracted company list.

  In this way, when the company group node size is set using the company evaluation information selected and specified by the user from a plurality of types of company evaluation information, the user can visually recognize the company group from various viewpoints. It becomes possible to grasp the corporate information from various perspectives.

And, as described above, in the case of a configuration that sets the size of the company group node using the company evaluation information,
Output element creation means
Regarding the color of the corporate group node, the color of the corporate group node corresponding to the corporate group that constitutes the set data including the same business attribute as the target business attribute is different from the color of the other corporate group node,
The size of the business attribute node is a fixed value,
As for the color of the business attribute node, it is desirable that the setting process is executed to make the color of the business attribute node corresponding to the target business attribute different from the colors of the other business attribute nodes.

  Here, as for “the size of the business attribute node is a fixed value”, the screen size (number of pixels) of the display device (display), as in the case of the size of the company group node described above, Even if the overall size of the node is adjusted in accordance with the number of dots of the printing apparatus (printer) or the like, this is included in the fixed value here.

  In addition, when the corporate group node or business attribute node is composed of characters (labels) indicating the name of the corporate group or business attributes, and figures (simple figures such as circles and squares, or company logos) May be different colors for both characters and graphics, or may be different colors for either characters or graphics.

  In this way, if the color scheme of the corporate group node and business attribute node can be easily distinguished from the analysis target related to the user's designation and other things, it is an overview of the industry structure of the target business and its surrounding fields It becomes much easier to do.

In addition, in the case of a configuration that performs output by the network expression described above,
User input acceptance means
It is configured to execute a process of accepting a user input indicating a company group node in a screen display or a three-dimensional display state,
The output means is
The company belonging to the company group corresponding to the company group node that has received the user instruction by the user input receiving means is acquired from the extracted company list, and the acquired company name is displayed on the screen or in a three-dimensional display in the vicinity of the company group node. It is desirable that a balloon display process for displaying a screen or a three-dimensional image in the balloon is executed.

  In this way, with a configuration that performs balloon display processing, it is possible to easily grasp the information of companies belonging to a company group when necessary, while overlooking the industry structure of the target business and its surrounding fields. Become.

  <Pre-processing common to cosine similarity method and LDA method>

Furthermore, in the corporate information providing system described above,
Using the association between the company stored in the company information storage means, the business attribute of the company and the company group to which the company belongs, the brute force combination of each company group and each business attribute is assigned to the same company group. A matrix creation means is provided for executing a process of creating a corporate group / business attribute co-occurrence matrix by extracting companies having the same business attribute and using the number of extracted companies as a matrix element. It is desirable to have a configuration.

  In the case of the configuration including the matrix creation means as described above, it is possible to easily prepare a new company group / business attribute co-occurrence matrix even when the data content stored in the company information storage means is updated. become able to.

  <Invention of program>

  The program of the present invention is for causing a computer to function as the company information providing system described above.

  The above-mentioned program or a part thereof is, for example, a magneto-optical disk (MO), a compact disk (CD), a digital versatile disk (DVD), a flexible disk (FD), a magnetic tape, or a read-only memory (ROM). , Electrically erasable and rewritable read-only memory (EEPROM), flash memory, random access memory (RAM), hard disk, solid state drive (SSD), flash disk, etc. And can be transmitted using a transmission medium such as a wired network such as LAN, MAN, WAN, Internet, Intranet, Extranet, or a wireless communication network, or a combination thereof. And It is also possible to transport and put in a carrier wave. Furthermore, the above program may be a part of another program, or may be recorded on a recording medium together with a separate program.

  As described above, according to the present invention, by using the corporate group / business attribute co-occurrence matrix, the similarity between industries indicating the similarity between each business attribute is calculated by the cosine similarity, or the potential The appearance probability of each business attribute in each company group is calculated by the dynamic Dirichlet distribution method (LDA), and the similar business attribute is calculated with respect to the target business attribute input by the user using the similarity between the industries or the appearance probability. Extraction and creation of an extracted company list based on the extracted company list and output of the company information using the extracted company list makes it possible for users with limited search know-how and industry structure knowledge to easily search for company information. There is an effect.

1 is an overall configuration diagram of a company information providing system according to a first embodiment of the present invention. The block diagram of the company information storage means of 1st Embodiment. The block diagram of the matrix memory | storage means of 1st Embodiment. The block diagram of the similarity memory | storage means between types of industries of 1st Embodiment. The block diagram of the extraction company list memory | storage means of 1st Embodiment. The block diagram of the output element memory | storage means of 1st Embodiment. The figure of the flowchart which shows the flow of the provision process of the company information by the company information provision system of 1st Embodiment. The figure which shows an example of the screen display of 1st Embodiment. The whole block diagram of the company information provision system of 2nd Embodiment of this invention. Explanatory drawing of the calculation process of the appearance probability by LDA of 2nd Embodiment. The figure of the flowchart which shows the flow of the provision process of the company information by the company information provision system of 2nd Embodiment.

  Embodiments of the present invention will be described below with reference to the drawings.

  [First Embodiment]

  FIG. 1 shows the overall configuration of a company information providing system 10 according to the first embodiment of the present invention. 2 shows the configuration of the company information storage unit 41, FIG. 3 shows the configuration of the matrix storage unit 42, and FIG. 4 shows the configuration of the inter-industry similarity storage unit 43. 5 shows the configuration of the extracted company list storage means 44, and FIG. 6 shows the configuration of the output element storage means 46. FIG. 7 is a flowchart showing the flow of the company information providing process performed by the company information providing system 10. Further, FIG. 8 shows an example of a screen display by the company information providing system 10.

  <Configuration of company information providing system 10>

  In FIG. 1, a company information providing system 10 includes a company information providing server 20 that executes various processes related to the provision of company information and stores various data necessary for the processes. A user terminal 2 operated by a user and an administrator terminal 3 operated by a system administrator are connected to the company information providing server 20 via a network 1.

  The company information providing server 20 is composed of one or a plurality of computers. The company information creating / editing means 21, the matrix creating means 22, the industry similarity calculation means 23, and the user input receiving means 24 are similar. Business attribute extraction means 25, similar company extraction means 26, extraction company list creation means 27, extraction company list editing means 28, output element creation means 29, output means 30, company information storage means 41, matrix The storage means 42, the inter-industry similarity storage means 43, the extracted company list storage means 44, the company evaluation information storage means 45, and the output element storage means 46 are configured.

  The network 1 includes various forms such as, for example, the Internet, an intranet, an extranet, a LAN, a MAN, a WAN, or a combination of these, and is wired or wireless, or a mixture of wired and wireless It doesn't matter if it's a mold.

  The user terminal 2 and the administrator terminal 3 are configured by a computer, and include input means such as a mouse and a keyboard and a display device such as a liquid crystal display. Moreover, you may provide the printing apparatus (printer) suitably. Note that the user terminal 2 and the administrator terminal 3 may be portable devices such as smartphones (high function mobile phones) and tablet terminals, for example, but the user terminal 2 is a display device capable of screen display shown in FIG. It is necessary to have.

  The company information creating / editing means 21 executes a process for creating and editing company information necessary for the screen display shown in FIG. 8 to be described later in accordance with an input operation of the system administrator from the administrator terminal 3.

  Specifically, the company information creating / editing means 21 receives the company information inputted by the system administrator and transmitted from the administrator terminal 3 via the network 1, that is, the company identification information (here, as an example) Company name), business attribute identification information of the company (in this example, the business attribute name), and identification information of the company group to which the company belongs (here, the company group as an example) The company information storage means 41 (see FIG. 2) stores data in which the name and the company group name are the same as the parent company name.

  Here, companies stored in the company information storage means 41 (see FIG. 2) include both listed companies and their affiliated companies (unlisted related companies), and a list of all these companies (company master) is stored. Create and edit company information as a basis.

  The business attribute is a word (word) indicating business content related to a company, and a plurality of business attributes can be given to one company. In the example of FIG. 2, business attributes such as department store and mail-order sales are given to the company called Company A2, and business attributes called solar cell and oil refinement are given to the company called Company B.

  For this business attribute, for example, business segment classification labels of each company are acquired from domestic and foreign research companies or sites that provide services that provide company information, and are given to the company master as business attributes. Alternatively, using the Internet curation service, feature words are extracted from the news and financial reports of each company, and given to the company master as business attributes. At this time, for example, a word having a high value of tf-idf (term frequency-inverse document frequency) is extracted and adopted as a business attribute. This tf-idf is an index for identifying how characteristic a specific word that appears in a document is, and is used in fields such as information search and text mining. Thereby, words that appear at a certain frequency and appear only in some news are extracted and adopted as business attributes.

  In addition, the parent-child relationship of the company, that is, the information that associates the parent company with the affiliated company (related company) (linkage information between the company group and the company that belongs to it) Acquired from a survey company, a site, etc., and assigns to which company group each company belongs to the company master according to this association information. In the example of FIG. 2, the company group A is composed of the parent company A, and its affiliated companies A2 and A3, and the company group B is composed of only the company B (however, For convenience, the company B is treated as a parent company.) The company group C is composed of a company C which is a parent company and a company C2 which is an affiliated company.

  The matrix creation means 22 uses the association between the company stored in the company information storage means 41 (see FIG. 2), the business attribute of the company, and the company group to which the company belongs, For the brute force combination, the companies belonging to the same company group and having the same business attributes are extracted, and the number of extracted companies is used as a matrix element. Is created, and the created company group / business attribute co-occurrence matrix X is stored in the matrix storage means 42 (see FIG. 3).

  Specifically, assuming that n = the number of company groups and p = the number of business attributes, the company group / business attribute co-occurrence matrix X (i, j) is a matrix of n rows and p columns, and i = 1 to n. , J = 1 to p. In the example of FIG. 3, the elements in the first row and the first column of the company group / business attribute co-occurrence matrix X are for the company having the business attribute of department store in the first column for the company group A in the first row. Is a number. The example in FIG. 2 corresponds to the example in FIG. 3. In the example in FIG. 2, for company group A, a total of two companies, Company A and Company A2, have business attributes of department stores. In the example of 3, the element of the first row and the first column is X (1,1) = 2.

  The inter-industry similarity calculation means 23 uses the company group / business attribute co-occurrence matrix X stored in the matrix storage means 42 (see FIG. 3) to indicate the similarity between the business attributes by the cosine similarity. The inter-sector similarity S is calculated, and the calculated inter-industry similarity S is stored in the inter-industry similarity storage unit 43 (see FIG. 4).

  Specifically, the cosine similarity is a value calculated by the formula S (l, m) in FIG. 4, and the business attribute of the lth column in the corporate group / business attribute co-occurrence matrix X and the It is a value indicating the degree of similarity with the business attribute in the m column. Here, if p = the number of business attributes, l (el) = 1 to p, m = 1 to p, and when l = m, that is, S (1,1) = S (2,2) = S (3,3) = S (4,4) = ... = 1. Since S (α, β) = S (β, α), the industry similarity S (l, m), which is a square matrix of p rows and p columns, is a symmetric matrix.

  The user input receiving unit 24 executes a process of receiving a user input including a target business attribute (a business attribute to be analyzed) transmitted from the user terminal 2 via the network 1. In addition to the target business attributes, there are selection designation information of company evaluation information (such as sales) used for setting the number f of similar business attribute extractions and the size of the company group node, and the company information shown in FIG. Input data of each item in the input unit 110 of the provision screen 100 is received via the network 1. As for the number f of similar business attributes extracted, in this embodiment, since the user inputs the number of cases including the target business attributes, the number obtained by subtracting 1 from the number actually entered by the user is the number of similar business attributes. The number of extractions f.

  The similar business attribute extracting unit 25 selects business attributes similar to the target business attribute received by the user input receiving unit 24 in descending order of the inter-industry similarity S stored by the inter-industry similarity storage unit 43 (see FIG. 4). The process of extracting as the similar business attributes by the number f of cases received by the user input receiving means 24 is executed. The similar business attribute extraction unit 25 may be configured to extract a predetermined number of cases instead of the number f of cases received by the user input reception unit 24.

  For example, assuming that the target business attribute is τ, S (1, τ), S (2, τ), S (3, τ), S (4, τ) from the inter-industry similarity storage means 43 (see FIG. 4). ),..., S (p, τ) are obtained, sorted in descending order of the similarity S between industries (in descending order), and f industry similarity S is extracted from the largest value, The f business attributes corresponding to the current line number are similar business attributes. At this time, S (τ, τ) = 1 has the largest value. Further, since the similarity S (l, m) between industries is a symmetric matrix, S (τ, 1), S (τ, 2), S (from the industry similarity storage means 43 (see FIG. 4). τ, 3), S (τ, 4),..., S (τ, p) may be acquired. In this case, these are sorted in descending order of the similarity S between industries, and f If the inter-industry similarity S is extracted, f business attributes corresponding to the column number at that time become similar business attributes.

  The similar company extracting unit 26 extracts the company having the target business attribute, the company group to which the company belongs, and the similar business attribute (f pieces) extracted by the similar business attribute extracting unit 25 from the company information storage unit 41 (see FIG. 2). The process of extracting a company having any one of the similar business attributes) and a company group to which the company belongs as a company group to which the similar company and the similar company belong is executed.

  Extracted company list creation means 27 includes target business attributes, set data including similar companies (company having target business attributes) extracted by similar company extraction means 26, and a company group to which the similar companies belong, and similar business attributes, An extracted company list 44A composed of group data including a similar company (company having similar business attributes) extracted by the similar company extraction means 26 and a company group to which the similar company belongs is created, and the extracted extracted company list 44A is extracted as the extracted company. Processing to be stored in the list storage means 44 (see FIG. 5) is executed.

  The extracted company list editing means 28 is a company for each company group included in each set of data (data in each row in FIG. 5) constituting the extracted company list 44A stored in the extracted company list storage means 44 (see FIG. 5). Evaluation information (such as sales) is acquired from the company evaluation information storage means 45, and processing for adding the acquired company evaluation information to each set of data is executed.

  At this time, the extracted company list editing unit 28 converts all types of the plurality of types of company evaluation information stored in the company evaluation information storage unit 45 (for example, sales, market capitalization, sales increase rate, etc.) into the company evaluation information. All types of company evaluation information acquired from the storage unit 45 may be added to each set of data. Among the plurality of types of company evaluation information stored in the company evaluation information storage unit 45, the user input reception unit 24 Only the company evaluation information (for example, only the sales) for which the user's selection designation is received is acquired from the company evaluation information storage means 45, and one type of company evaluation information selected and designated by the user is added to each set of data. Also good.

  Further, in this embodiment, the extracted company list editing means 28 uses the company evaluation information about the parent company of each company group as “company evaluation information about each company group”, and sets each group data constituting the extracted company list 44A. Append to In the extracted company list 44A stored in the extracted company list storage means 44 (see FIG. 5), the parent company identification information (here, the parent company name) is used as the company group identification information (here, the company group name). As a result, the parent company can be identified. In the example of FIG. 5, company A, company A2, and company A3 all belong to the same company group A. Therefore, company evaluation information (such as sales) of company A, which is the parent company, is stored in company A, company A2, It is added to each group data of A3 company. The extracted company list editing means 28 does not include the company evaluation information about the parent company of the company group as “company evaluation information about each company group” but includes the parent company and affiliated company of the company group (included in the extracted company list 44A). The company evaluation information obtained by summing up or aggregating each company evaluation information for all affiliated companies included in the company information storage means 41 (see FIG. 2) may be extracted. You may add to each group data which comprises company list 44A.

  The output element creation means 29 uses the extracted company list 44A (see FIG. 5) to which the company evaluation information is added by the extracted company list editing means 28, and uses the extracted company list 44A (see FIG. 5). The size and color setting information of the company group node corresponding to the company group, and the size and color setting information of the business attribute node corresponding to each business attribute included in each set data constituting the extracted company list 44A A node list 46A including the created node list 46A is stored in the output element storage means 46 (see FIG. 6). The node list 46A is created by eliminating duplication so that a plurality of records (rows) cannot be made for the same node.

  In the present embodiment, the extracted company list 44A stored in the extracted company list storage unit 44 (see FIG. 5) also serves as the side list 46B for each side connecting the company group node and the business attribute node. However, for example, when setting the type of each side (different from solid line, dotted line, etc.), thickness (separate from thin line, thick line, etc.), color, etc., the output element creation means 29 uses the extracted company list. Apart from 44A, the side list 46B may be created, and the created side list 46B may be stored in the output element storage means 46 (see FIG. 6).

  Specifically, the output element creation unit 29 converts the size of the company group node (in the case of the company group node flag = 1) after the conversion using the company evaluation information (for example, sales) for the company group. Execute the setting process to set the value. The output element creation means 29 is a company group corresponding to a company group that constitutes the set data including the same business attribute as the target business attribute in the size of the company group node (when the company group node flag = 1). The size of the node (when the target company group flag = 1) is a value after conversion using the company evaluation information about the company group, and other company group nodes (when the target company group flag = 0) A setting process for setting a fixed value may be executed. The fixed value means a value that does not perform conversion using the company evaluation information. For example, adjustment according to the size of the entire screen is allowed.

  Here, the “value after conversion using the company evaluation information” may be a value obtained by performing linear conversion (linear conversion) by a function (y = ax) that multiplies the company evaluation information by a coefficient a. Further, it may be a value obtained by performing conversion using a linear function having a constant term (y = ax + b; a function for adding the constant term b to the proportional component ax based on the company evaluation information). The coefficient a is a coefficient determined in advance according to the numerical value indicating the company evaluation information, and varies depending on the type of the company evaluation information. The “value after conversion using the company evaluation information” may be a value obtained by performing conversion using a more complicated function. In the example of FIG. 5, the sales of the corporate group A are 1,300 (million yen), the sales of the corporate group B are 400 (million yen), and the sales of the corporate group C are 1,600 (million yen) Therefore, in the example of FIG. 6, the size of the corporate group node corresponding to the corporate group A is set to 1.3, and the size of the corporate group node corresponding to the corporate group B is 0. 4 is set, and the size of the company group node corresponding to the company group C is set to 1.6.

  Then, the output element creation means 29, when only the company evaluation information related to the user's selection designation received by the user input receiving means 24 among the plurality of types of company evaluation information is added to the extracted company list 44A, Conversion is performed using the company evaluation information. On the other hand, when a plurality of types of company evaluation information are added to the extracted company list 44A, the output element creation unit 29 selects the company evaluation related to the user's selection designation received by the user input receiving unit 24 from among them. Information is acquired and converted using the acquired company evaluation information.

  Further, the output element creation means 29, for the color of the company group node, the company group node corresponding to the company group that constitutes the set data including the same business attribute as the target business attribute (when the target company group flag = 1) And a process of setting different colors for the other company group nodes (when the target company group flag = 0). For example, the former (when the target company group flag = 1) is highlighted as orange, and the latter (when the target company group flag = 0) is black.

  Furthermore, the output element creation unit 29 executes a setting process for setting a fixed value for the size of the business attribute node (when the business attribute node flag = 1). In the example of FIG. 6, the size of all business attribute nodes is 0.5.

  The output element creation means 29 then sets the color of the business attribute node (when the business attribute node flag = 1), the color of the business attribute node (when the target business attribute flag = 1) corresponding to the target business attribute, A setting process for changing the color of other business attribute nodes (when the target business attribute flag = 0) to a different color is executed. For example, the former (when the target business attribute flag = 1) is conspicuous as red, and the latter (when the target business attribute flag = 0) is blue.

  The output means 30 uses the extracted company list 44A stored in the extracted company list storage means 44 (see FIG. 5) to display a screen showing the relationship between the company group and the business attribute and / or the relationship between the company and the business attribute. , Printing, three-dimensional display, or a combination thereof, or an output process (a process for transmitting these output data to the user terminal 2 via the network 1) by a combination of these and audio output.

  Specifically, the output unit 30 performs network representation according to the setting information of the node list 46A and the side list 46B (or the extracted company list 44A that also serves as the side list 46B) stored in the output element storage unit 46 (see FIG. 6). Screen display processing (display processing of the company information providing screen 100 shown in FIG. 8 to be described later) is executed, screen display processing of the extracted company list 44A is executed. When a print request is received, a print process of the company information provision screen 100 and the extracted company list 44A (a process of transmitting these print data to the user terminal 2 via the network 1) is executed.

  At this time, the output unit 30 executes screen display processing based on network expression using a graph drawing method (force-oriented algorithm) based on a dynamic model. In other words, in order to arrange nodes so that the length of each side is almost equal, and the sides do not intersect as much as possible, a virtual force is assigned with the sides regarded as springs, and stable with low mechanical energy. By searching for the state, the drawing position of the node is determined.

  Further, the output means 30 is a user input operation (for example, the display position of the company group node) that indicates the company group node displayed on the output unit 120 of the company information provision screen 100 shown in FIG. When the user input accepting unit 24 accepts a pointer or cursor to the user, the company belonging to the company group corresponding to the company group node is extracted from the extracted company list storage unit 44 (see FIG. 5). A balloon display process is executed to display the name of the acquired company from the company list 44A in a balloon displayed on the screen in the vicinity of the company group node. The name of the company displayed on the screen in the balloon may be output as a sound, or a sound effect may be generated when a company group node is instructed. Alternatively, the size may be changed according to company evaluation information (for example, sales), or may be changed depending on whether the company group is associated with the target business attribute.

  As shown in FIG. 2, the company information storage means 41 includes company identification information (here, the company name is used as an example) and company group identification information to which the company belongs (here, the company group name is used as an example). Furthermore, the company group name is assumed to be the same as the parent company name) and the business attribute identification information of the company (here, the business attribute name is used as an example) and stored in association with each other. is there.

  As shown in FIG. 3, the matrix storage means 42 uses one of the rows or columns of the matrix as a company group (here, the row is a company group) and the other as a business attribute (here, the column is a business attribute). And the company group / business attribute co-occurrence matrix X having the number of companies having each business attribute for each company group as a matrix element. For example, if n = the number of company groups and p = the number of business attributes, the corporate group / business attribute co-occurrence matrix X (i, j) has n rows and p columns. Since it is a matrix (i = 1 to n, j = 1 to p), each value of X (i, j) is stored in an array, and the row number and company group identification information (here, the company group name and And the correspondence relationship between the column number and the business attribute identification information (in this case, the business attribute name). In addition, a database including a table with rows as records and columns as columns may be prepared.

  As shown in FIG. 4, the inter-industry similarity storage unit 43 stores an inter-industry similarity S indicating the similarity between business attributes, calculated by the cosine similarity. For example, if p = the number of business attributes, the similarity S (l, m) between industries is a square matrix of p rows and p columns (l (el) = 1 to 1). Since p, m = 1 to p), each value of S (l, m) is stored in an array, and the row number or the column number (S is a symmetric matrix, so both are the same). The correspondence with the business attribute identification information (here, the business attribute name) is stored. In addition, a database including a table with rows as records and columns as columns may be prepared.

  As shown in FIG. 5, the extracted company list storage means 44 includes, as the extracted company list 44A, company identification information (here, the company name is used as an example) and identification information of the company group to which the company belongs (here. Then, the company group name is used as an example, and the company group name is the same as the parent company name.) And the business attribute identification information of the company (here, the business attribute name is used as an example). The similarity between company groups and business attributes (in the case of the second embodiment described later) and company evaluation information (for example, sales of the company group) are stored in association with each other. In the first embodiment, each record of the extracted company list 44A includes a target business attribute, a similar company extracted by the similar company extraction unit 26 (a company having the target business attribute), and a company group to which the similar company belongs. It is group data, or group data including a similar business attribute, a similar company extracted by the similar company extraction means 26 (a company having a similar business attribute), and a company group to which the similar company belongs.

  The company evaluation information storage means 45 uses a plurality of types of company evaluation information, such as financial information and management indicators about the company, as information for evaluating the size and status of the company, as identification information (here, an example) As a company name). Specifically, the company evaluation information storage means 45 includes, for example, sales, market capitalization, cash and deposits and equivalents, company value (EV), EBITDA margin, sales increase rate, cash deposit increase rate as company evaluation information. Are stored in association with the identification information of the company (here, the company name). The company evaluation information stored in the company evaluation information storage unit 45 may be data stored from another system via the network 1 and is input from the administrator terminal 3 by the system administrator and is stored in the company. The data received by the information creating / editing means 21 may be stored. Further, the company evaluation information storage unit 45 may constitute a part of the company information storage unit 41 (see FIG. 2) or may be integrated with the company information storage unit 41.

  The output element storage means 46 connects a node list 46A in which drawing data related to nodes (company group nodes and business attribute nodes) as output elements are collected, and an edge (company group node and business attribute nodes) as output elements. A side list 46B in which drawing data related to (element) is collected is stored.

  As shown in FIG. 6, the node list 46A indicates node identification information (here, the label of the node, which is the name of the company group name or the business attribute), and whether it is a company group node. Company group node flag, business attribute node flag indicating whether or not a business attribute node, and target business attribute indicating whether or not it is a business attribute node for the target business attribute (the business attribute input by the user as an analysis target) This is data in which an attribute flag, a target company group flag indicating whether or not a company group is associated with a target business attribute, and a node size.

  The side list 46B associates the identification information of two nodes (company group node and business attribute node) connected by the side, and sets detailed information about the type, thickness, color, etc. of the side. Is also associated data. At this time, the identification information of the company group node may be the same as the identification information of the company group, and the identification information of the business attribute node may be the same as the identification information of the business attribute. However, the edge list 46B can also be used as the extracted company list 44A stored in the extracted company list storage means 44 (see FIG. 5). In this embodiment, the extracted company list 44A also serves as the edge list 46B. ing.

  In the above, each means 21-30 which performs various processes of the company information provision server 20 is the central processing unit (CPU) provided in the computer main body which comprises the company information provision server 20, and operation | movement of this CPU. Implemented by one or more programs that define the procedure.

  The storage units 41 to 46 of the company information providing server 20 are preferably realized by, for example, a hard disk, a solid state drive (SSD), or the like, as long as there is no problem in storage capacity or access speed. For example, other recording media such as a flash memory and a DVD may be adopted.

  Further, the company information providing server 20 is not limited to one realized by one computer or one CPU, and may be realized by performing distributed processing with a plurality of computers or a plurality of CPUs. .

  <Flow of processing by the company information providing system 10>

  In the first embodiment, the company information providing system 10 performs the company information providing process as follows.

  In FIG. 7, the system administrator operates the administrator terminal 3 to create and edit company information. In the company information providing server 20, the company information creation / editing means 21 receives data accompanying the input operation of the system administrator transmitted from the administrator terminal 3 via the network 1, and the received data is converted into the company information. It memorize | stores in the memory | storage means 41 (refer FIG. 2) (step S1). At this time, the system administrator may input data created / edited by another computer from the administrator terminal 3, or all creation / editing operations may be performed by the administrator terminal 3. Since the contents of creation and editing of company information (including setting of business attributes for companies and setting of company groups to which companies belong) have already been described in detail, description thereof is omitted here.

  As for company evaluation information (for example, sales) stored in the company evaluation information storage unit 45, data transmitted from another system via the network 1 is automatically stored in the company evaluation information storage unit 45. Alternatively, the company information creation / editing means 21 may accept an input operation of the system administrator from the administrator terminal 3 and store it in the company evaluation information storage means 45. When data is received from another system as in the former, a data acquisition request signal is transmitted to the other system by company evaluation information acquisition means (not shown) provided in the company information providing server 20. Alternatively, data transmission may be received at a timing managed by another system.

  Subsequently, in the company information providing server 20, the matrix creation unit 22 creates and creates a company group / business attribute co-occurrence matrix X using the data stored in the company information storage unit 41 (see FIG. 2). The company group / business attribute co-occurrence matrix X is stored in the matrix storage means 42 (see FIG. 3) (step S2). The creation timing may be given by the system administrator from the administrator terminal 3 or may be created periodically, or the update of data stored in the company information storage means 41 (see FIG. 2) is detected. You may make it. Since the contents of creation of the company group / business attribute co-occurrence matrix X have already been described in detail, description thereof is omitted here.

  Further, the company information providing server 20 uses the company group / business attribute co-occurrence matrix X stored in the matrix storage unit 42 (see FIG. 3) by the inter-industry similarity calculation unit 23 to calculate each cosine similarity according to the cosine similarity. The inter-industry similarity S indicating the similarity between business attributes is calculated, and the calculated inter-industry similarity S is stored in the inter-industry similarity storage means 43 (see FIG. 4) (step S3). The calculation processing of the similarity S between industries is executed when the company group / business attribute co-occurrence matrix X is created or updated. The calculation timing is given from the administrator terminal 3 by the system administrator. Alternatively, it may be linked with the processing (step S2) by the matrix creating means 22 described above. Since the contents of the calculation process of the similarity S between industries are already described in detail, the description is omitted here.

  After that, in the state where the above pre-processing is completed, in the company information providing server 20, the target business attribute (business to be analyzed) transmitted from the user terminal 2 via the network 1 by the user input receiving unit 24. The user input including (attribute) is received (step S4).

  Specifically, when the user accesses from the user terminal 2, display data (for example, WEB data) of the company information provision screen 100 is transmitted to the user terminal 2 via the network 1 by the user input receiving unit 24. Is done. Then, a company information provision screen 100 as shown in FIG. 8 is displayed on the screen of the user terminal 2. The display process of the company information provision screen 100 (the display process of the part excluding the drawing part by network expression) may be performed by a program preinstalled in the user terminal 2, and in that case, the company information provision server The user input receiving unit 24 on the 20 side is configured to perform processing for receiving user input data transmitted from the user terminal 2 via the network 1 but not to perform screen display processing.

  In FIG. 8, the company information provision screen 100 includes an input unit 110 for inputting each item, and an output unit 120 for performing drawing output by network expression. However, in the step S4, the input unit 110 is displayed, but the output unit 120 is not displayed (drawing by network expression).

  The input unit 110 includes a user ID input unit 111 and major classification, middle classification, and minor classification input sections 112, 113, and 114 for selectively inputting target business attributes (business attributes to be analyzed). Is provided. The number of business attributes p (the number of rows and columns of the similarity S between industries in FIG. 4 which is a symmetric matrix) is very large, for example, about 500, so that it is easy to select a large classification (for example, 13 It is possible to narrow down in stages, such as about an individual), medium classification (for example, about 69), and small classification (for example, about 500). Accordingly, a word (word) displayed as an option in the bottom subcategory is a business attribute name in the present invention. For example, “Retail” is selected as the major category, “Mail order / door-sales” is selected as the middle category below, and “Mail order (stationery / office supplies)” is selected as the minor category below. Or, you can select “Mechanical / Electrical products” in the major category, select “Electronic components / device manufacturing” in the middle category below, and then select “Sensor” in the minor category below. .

  The input unit 110 is provided with an input unit 115 for the number of extracted similar business attributes. In the present embodiment, since the number of extractions that the user inputs here is the number including the target business attributes, the number of similar business attribute extractions f used in the processing by the similar business attribute extraction unit 25 is input by the user. This is the number obtained by subtracting 1 from the number entered in the section 115. Note that the input unit 115 may be configured to input the number of similar business attribute extractions f itself (number of target business attributes not included).

  Further, the input unit 110 is provided with an input unit 116 for setting a threshold value for filtering the sales target company in the balloon by sales amount. Increasing the threshold setting in the input unit 116 increases the number of companies excluded from the display target. Even if the company group node displayed on the output unit 120 is indicated (by placing a pointer or cursor on the company group node) Also, there may be no company displayed in the balloon, or no balloon may be displayed.

  Further, the input unit 110 is provided with an input unit 117 for inputting a set value for adjusting the size of the company group node (company group name) as a whole. The setting value input by the input unit 117 is uniformly reflected in the size of all the company group nodes, and the size reflecting this is further converted using company evaluation information (for example, sales). In other words, there is a large or small difference in the size of each company group node.

  Furthermore, in the input unit 110, the company evaluation information used when the conversion process for changing the size of the company group node (company group name) according to the numerical value of the company evaluation information (for example, sales) is performed. An input unit 118 for selecting and specifying is provided.

  The user inputs necessary information to the input units 111 to 118 of each item provided in the input unit 110 of the company information provision screen 100 in FIG. 8, and then clicks a transmission button (not shown) to input data. Is transmitted to the company information providing server 20 via the network 1. In the company information providing server 20, this input data is received by the user input receiving means 24. The user ID input to the input unit 111 may be transmitted at the login stage.

  Then, in the company information providing server 20, the business attribute similar to the target business attribute received by the user input receiving unit 24 is stored by the similar business attribute extracting unit 25 by the inter-industry similarity storage unit 43 (see FIG. 4). Only the number f received by the user input accepting means 24 is extracted as the similar business attribute in descending order of the similarity S between the industries (step S5). Since the content of the similar business attribute extraction process has already been described in detail, the description thereof is omitted here.

  Subsequently, in the company information providing server 20, the similar company extracting unit 26 extracts the company having the target business attribute, the company group to which the company belongs, and the similar business attribute extracting unit from the company information storage unit 41 (see FIG. 2). A company having a similar business attribute (any one of f similar business attributes) extracted by 25 and a company group to which this company belongs are extracted as a similar company and a company group to which this similar company belongs (step S6).

  Further, the company information providing server 20 includes the target business attributes by the extracted company list creation means 27, the similar companies extracted by the similar company extraction means 26 (the companies having the target business attributes), and the company group to which the similar companies belong. Create an extracted company list 44A composed of set data including set data, similar business attributes, similar companies extracted by the similar company extraction means 26 (company having similar business attributes), and a company group to which the similar companies belong, The created extracted company list 44A is stored in the extracted company list storage means 44 (see FIG. 5) (step S7).

  Subsequently, in the company information providing server 20, the extracted company list editing unit 28 causes each set of data (data in each row in FIG. 5) constituting the extracted company list 44A stored in the extracted company list storage unit 44 (see FIG. 5). ) Is acquired from the company evaluation information storage means 45, and the acquired company evaluation information is added to each set of data (step S8). Since the contents of the process of acquiring the company evaluation information for each company group have already been described in detail, the description is omitted here.

  Thereafter, the company information providing server 20 uses the extracted company list 44A (see FIG. 5) with the company evaluation information added by the extracted company list editing means 28 by the output element creating means 29, and extracts the extracted company list 44A. The size and color setting information of the company group node corresponding to each company group included in each set group data, and the business attribute node corresponding to each business attribute included in each set data constituting the extracted company list 44A The node list 46A including the size and color setting information is created, and the created node list 46A is stored in the output element storage means 46 (see FIG. 6) (step S9). Since the contents of the creation of the node list 46A have already been described in detail, the description thereof is omitted here.

  In this embodiment, the extracted company list 44A (see FIG. 5) also serves as the side list 46B. However, when detailed settings are made for the side, the output element creation unit 29 creates the side list 46B. The output element storage means 46 (see FIG. 6) may be stored.

  In the company information providing server 20, the output means 30 uses the extracted company list 44A stored in the extracted company list storage means 44 (see FIG. 5), and more specifically, uses the extracted company list 44A. The relationship between the company group and the business attribute according to the setting information of the node list 46A and the edge list 46B (or the extracted company list 44A that also serves as the edge list 46B) that is created and stored in the output element storage means 46 (see FIG. 6). And / or screen display data in a network expression indicating the relationship between the company and the business attribute is created, and the created screen display data is transmitted to the user terminal 2 via the network 1 (step S10). The screen display data may be stereoscopic display data or may include audio data.

  Then, on the screen of the user terminal 2, as shown in FIG. 8, drawing output by network expression is performed on the output unit 120 of the company information provision screen 100. The display example shown in the output unit 120 in FIG. 8 corresponds to the extracted company list 44A in FIG. 5 and the node list 46A in FIG.

  Therefore, “department store” is the target business attribute (target business attribute flag = 1 in FIG. 6), and the corporate groups A and D are associated with the target business attribute “department store”. That is, as shown in FIG. 5, the company groups A and D include companies having “department store” as a business attribute. For this reason, as shown in FIG. 8, the business attribute node corresponding to “department store” has a different color from other business attribute nodes so as to stand out. In addition, the business attribute node corresponding to “Department store” and the company group nodes corresponding to the company groups A and D linked to “Department store” are connected to each other and connected to the company groups A and D. Each corresponding company group node has a different color from the other company group nodes so as to stand out. Further, as shown in FIG. 5, since the sales of the corporate groups A and D are relatively large, as shown in FIG. 6, the size of each corporate group node corresponding to the corporate groups A and D is set to be relatively large. Then, as shown in FIG. 8, it is drawn large so as to stand out in the network expression.

  Further, as shown in FIG. 5, since the company group D is composed of the company D and the company D2, when the company group node corresponding to the company group D is indicated on the screen (when a pointer or cursor is placed), As shown in FIG. 8, Company D and Company D2 are displayed in the balloon. Furthermore, as shown in FIG. 5, companies D and D2, which belong to company group D, have business attributes such as “display planning / operation” and “apparel specialty store” in addition to “department store” as the target business attribute. The company group node corresponding to the company group D is not only connected to the business attribute node corresponding to the “department store” by the side, but also “display planning / operation” and “apparel specialty store”. Each business attribute node corresponding to is connected with each side.

  Further, in the example of FIG. 8, the company group node and the business attribute node are composed of characters (labels) indicating the name of the company group and the business attribute name, and a round figure. When variably setting the conversion process using evaluation information (such as sales) or setting the color of a node, consider both the character and the figure as one, and change the size and color of both the character and the figure. Alternatively, the character and the graphic may be considered separately, and the size or color of either the character or the graphic may be changed. In the latter case, the size and color of the character (label) may be changed, but the size and color of the figure may not be changed. Conversely, the size and color of the figure are changed, but the size of the character (label) is changed. The color may not be changed. The graphic is not limited to a simple graphic such as a circle or square that can be prepared as text data, but may be prepared as image data such as a company logo.

  <Effects of First Embodiment>

  According to such 1st Embodiment, there exist the following effects. That is, the company information providing system 10 uses the company group / business attribute co-occurrence matrix X (see FIG. 3) and uses the cosine similarity to indicate the similarity between business attributes S (see FIG. 4). ), And using the similarity S between industries, f similar business attributes are extracted from the target business attributes input by the user, and an extracted company list 44A (see FIG. 5) is created based on the extracted business business attributes. In addition, since the extracted company list 44A is used to output the company information, even a user who has little search know-how or knowledge of the industry structure can easily search for company information.

  Further, since the company information providing system 10 includes the matrix creating means 22, even when the data content stored in the company information storage means 41 (see FIG. 2) is updated, a new company group / business attribute is shared. The starting matrix X (see FIG. 3) can be easily prepared.

  Furthermore, since the company information providing system 10 includes the inter-industry similarity calculation means 23, the data content stored in the company information storage means 41 (see FIG. 2) is updated, and the company group / business attribute co-occurrence matrix is updated. Even when there is a change in the contents of X (see FIG. 3), it is possible to easily prepare a new inter-industry similarity S (see FIG. 4) according to the changed corporate group / business attribute co-occurrence matrix X it can.

  And since the output means 30 performs the output by the network expression which connected the enterprise group node and the business attribute node by the edge, the relationship between the enterprise group and the business attribute is visualized as a network plot. Gain an overview of the industry structure of the field. For this reason, for example, it is possible to support the direction of the advisory of the company research work for the extracted company.

  The company information providing system 10 includes a company evaluation information storage unit 45, an extracted company list editing unit 28, and an output element creating unit 29. The output element creating unit 29 is added to the extracted company list 44A (see FIG. 5). Since the size of the company group node is set using the received company evaluation information (for example, sales etc.), the output means 30 has the node list 46A (see FIG. 6) created by the output element creation means 29. ), The size of the company group node can be changed according to the numerical value of the company evaluation information. For this reason, in addition to an overview of the target business and the surrounding industry structure, it is also possible to visually grasp the size and status of the corporate group.

  Further, the output element creating unit 29 receives the company evaluation information (for example, sales) related to the user designation received by the user input receiving unit 24, that is, the company evaluation information selected and specified by the user from a plurality of types of company evaluation information. It is configured to set the size of the company group node, so that the user can visually grasp the company group from various viewpoints, making it easier to grasp the company information from various perspectives. Can be realized.

  The output element creation unit 29 is configured to make settings so that the color of the enterprise group node and the business attribute node can be easily distinguished from the analysis target according to the user's designation and the others. Therefore, the user can more easily view the industry structure of the target business and its peripheral fields.

  Further, since the output unit 30 performs the balloon display process, the user can easily grasp the information of the companies belonging to the company group when necessary, while overlooking the industry structure of the target business and its peripheral fields. it can.

  [Second Embodiment]

  FIG. 9 shows the overall configuration of a company information providing system 200 according to the second embodiment of the present invention. FIG. 10 is an explanatory diagram of an appearance probability calculation process by a potential Dirichlet allocation method (LDA: Latency Dirichlet Allocation). FIG. 11 is a flowchart showing the flow of the company information providing process performed by the company information providing system 200.

  <Configuration of company information providing system 200>

  The company information providing server 200 according to the second embodiment has many parts having the same configuration and functions as those of the company information providing server 10 according to the first embodiment. The description will be omitted or the corresponding names will be given for simplification, and the following description will focus on the different parts.

  In FIG. 9, a company information providing system 200 includes a company information providing server 220 that executes various processes relating to the provision of company information and stores various data necessary for these processes. A user terminal 2 operated by a user and an administrator terminal 3 operated by a system administrator are connected to the company information providing server 220 via a network 1.

  The company information providing server 220 includes one or a plurality of computers. The company information creating / editing means 221, the matrix creating means 222, the appearance probability calculating means 223, the user input receiving means 224, the similar company group Extraction means 225, similar business attribute extraction means 226, extraction company list creation means 227, extraction company list editing means 228, output element creation means 229, output means 230, company information storage means 241, and matrix storage Means 242, appearance probability storage means 243, extracted company list storage means 244, company evaluation information storage means 245, and output element storage means 246 are included.

  The company information creation / editing means 221, matrix creation means 222, company information storage means 241, and matrix storage means 242 are the company information creation / editing means 21, matrix creation means 22, company information storage means of the first embodiment. 41 and matrix storage means 42. Therefore, the company group / business attribute co-occurrence matrix X (see FIG. 3) and the company information (see FIG. 2) for creating it are the same as those in the first embodiment.

  Appearance probability calculation means 223 uses the business group / business attribute co-occurrence matrix X (similar to FIG. 3) stored in matrix storage means 242, and uses the Dirichlet allocation method (LDA) for each business in each business group. A process of calculating the appearance probability Q of the attribute and storing the calculated appearance probability Q in the appearance probability storage unit 243 (see FIG. 10) is executed.

  Here, the use of the latent Dirichlet distribution method (LDA) by the appearance probability calculation means 223 is performed by the following association. LDA is a technique mainly used in the field of text mining, and is a model that expresses the appearance probability of a “word” in a certain “document” by a mixed distribution of word frequency tendencies of “topics”. Therefore, in the second embodiment, since the appearance probability of “business attribute” in a certain “company group” is calculated, when using LDA, “document” is replaced with “company group” and “word” is replaced with “word”. Replace with “business attribute” and replace “topic” with “industry trend (business attribute trend)”. For example, when a topic is classified into a sports system, a social system, an economic system, or the like when classifying news, the topic indicates this classification, and the number of classifications is the number of topics.

  More specifically, in FIG. 10, β (k, j) is a matrix of K rows and p columns, and the appearance probability of the word j in a certain topic k is replaced with the appearance probability of the business attribute j in a certain industry trend k. It is a thing. k = 1 to K, and K = number of industry trends, which corresponds to the number of topics. Also, j = 1 to p and p = the number of business attributes, which corresponds to the number of words. Note that in the second embodiment, the system administrator determines an appropriate number of K = the number of industry trends (corresponding to the number of topics).

  Π (i) is a K-dimensional vector (K = number of industry trends: corresponding to the number of topics), and the appearance probability of each topic in the document i is changed to the appearance probability of each industry trend in the company group i. It is a replacement. i = 1 to n and n = the number of company groups, which corresponds to the number of documents.

  Further, N (i) is obtained by replacing the number of words in the document i with the number of business attributes that the company group i has.

  And t in the upper right of β means transposition, and an operation of multiplying the K-dimensional vector π and the transposition matrix of β is performed. By multiplying the appearance probability of the business attribute j in the tendency k, the appearance probability Q of each business attribute in each company group is obtained.

  The system administrator uses Gibbs sampling or the like to calculate the K-dimensional vector π (i) indicating the appearance probability of each industry trend in the company group i and the appearance probability of the business attribute j in a certain industry trend k. Β (k, j) which is a matrix of K rows and p columns shown is estimated from the data, and the estimated value is stored in estimated value storage means (not shown) provided in the company information providing server 220. . Further, N (i) indicating the number of business attributes of the company group i is also stored in the estimated value storage means. Then, the appearance probability calculation means 223 includes the company group / business attribute co-occurrence matrix X (similar to FIG. 3) stored in the matrix storage means 242, and π (i) and β (k) stored in the estimated value storage means. , J), N (i), the LDA is used to calculate the appearance probability Q of each business attribute in each company group, and the calculated appearance probability Q is stored in the appearance probability storage means 243 (see FIG. 10). . The appearance probability Q of each business attribute in each business group can be rephrased as the appearance probability of each business group in each business attribute.

  Specifically, if n = the number of company groups and p = the number of business attributes, the appearance probability Q (i, j) is a matrix of n rows and p columns, i = 1 to n, j = 1 to p. It is. If the example of FIG. 2 and the example of FIG. 3 of the first embodiment are used as they are, in the example of FIG. 10, the element of the first row and the first column of Q (1,1) = appearance probability Q is the first row. Is the appearance probability of the business attribute of department store in the first column in the corporate group A.

  The user input receiving unit 224 has a configuration substantially similar to that of the user input receiving unit 24 of the first embodiment, and executes a process of receiving a user input including a target business attribute. The difference is that the user input reception unit 24 of the first embodiment uses the number f of similar business attribute extractions used in the processing by the similar business attribute extraction unit 25 (more precisely, the number f of similar business attribute extractions) On the other hand, the user input receiving unit 224 of the second embodiment is similar to that used in the processing by the similar business attribute extracting unit 226. This is a configuration that accepts input of the number of business attribute extractions f2 (details will be described later). Note that the user input receiving unit 224 may be configured to receive an input of the number f1 of extracted similar company groups (details will be described later) used in the processing by the similar company group extracting unit 225.

  The similar company group extracting unit 225 selects a company group having a high appearance probability with respect to the target business attribute received by the user input receiving unit 224 in descending order of the appearance probability Q stored in the appearance probability storage unit 243 (see FIG. 10). A process of extracting as a similar company group by a predetermined number f1 is executed. The number f1 of similar company group extractions may be the number of user inputs received by the user input receiving unit 224.

  For example, if the target business attribute is τ, Q (1, τ), Q (2, τ), Q (3, τ), Q (4, τ), Q, from the appearance probability storage means 243 (see FIG. 10). Q (5, τ), Q (6, τ),..., Q (n, τ) are obtained and sorted in descending order of appearance probability Q, and f1 occurrences from the largest value If the probability Q is extracted, f1 company groups corresponding to the row number at that time become similar company groups. In the example of FIG. 10, Q (1,3), Q (2,3), Q (3,3), Q (4,3), Q (5,3), which are columns of the target business attribute τ = 3. , Q (6,3),..., Q (n, 3), sort them in descending order, and the third row of corporate groups corresponding to the row number of Q (3,3), and Q (5 , 3) and the like in the fifth row corresponding to the row number are extracted as f1 similar enterprise groups.

  The similar business attribute extraction unit 226 uses the appearance probability Q stored in the appearance probability storage unit 243 (see FIG. 10) as the business attribute having a high appearance probability for each similar company group extracted by the similar company group extraction unit 225. In order from the highest, the number of cases f2 received by the user input receiving means 224 is extracted as similar business attributes. Note that the number f2 of similar business attribute extractions may be a predetermined number.

  In the second embodiment, unlike the first embodiment, the target business attributes designated by the user are referred to as similar business attributes. However, depending on which business attribute the target business attribute designated by the user is and the size of the number f2 of extracted similar business attributes designated by the user, the appearance probability Q of the extracted top f2 cases may be There may be a situation in which the target business attribute is not included in the similar business attributes, but in this case, the user increases the input value of the number of extractions f2 to increase the similarity among the top f2 similar business attributes. Can include target business attributes. If the target business attribute is not included in the similar business attributes even if the input value of the number of extractions f2 is increased, the target business attributes are rendered with a useful network expression in the first place. It was a business attribute that could not be.

  For example, if the similar company group extracted by the similar company group extraction unit 225 is σ, Q (σ, 1), Q (σ, 2), Q (σ, 3) from the appearance probability storage unit 243 (see FIG. 10). ), Q (σ, 4),..., Q (σ, p) are sorted in descending order of appearance probability Q, and f2 appearance probabilities Q are extracted in descending order of value. For example, f2 business attributes corresponding to the column number at that time are similar business attributes. Since f1 similar company groups are extracted, such f2 similar business attribute extraction processing is performed for all f1 similar company groups. In the example of FIG. 10, Q (3,1), Q (3,2), Q (3,3), Q (3,4),. p) are obtained, sorted in descending order, f2 appearance probabilities Q are extracted in descending order, and f2 business attributes corresponding to the column number at that time are used as similar business attributes. Similarly, Q (5,1), Q (5,2), Q (5,3), Q (5,4),. Acquired and sorted in descending order, f2 appearance probabilities Q are extracted in descending order, and f2 business attributes corresponding to the column number at that time are used as similar business attributes. If similar company groups are extracted in addition to similar company groups σ = 3 and 5, similar processing is performed, and f2 similar business attributes are extracted for each.

  The extracted company list creation means 227 is a similar business extracted by the similar business attribute extraction means 226 among the companies belonging to the similar company group extracted by the similar company group extraction means 225 from the company information storage means 241 (similar to FIG. 2). A company having an attribute is extracted as a similar company, an extracted company list 244A is created that includes the extracted similar company, a similar company group to which the similar company belongs, and group data including similar business attributes. The extracted company list storage means 244 (similar to FIG. 5) is stored. Further, in the second embodiment, the extracted company list creation unit 227 uses the appearance probability Q when the similar business attribute is extracted by the similar business attribute extraction unit 226 as the similarity between the company group and the business attribute in the set data. Append. Therefore, in FIG. 5, the similarity between company groups and business attributes included in each set of data is data unique to the second embodiment.

  The extracted company list editing means 228, the output element creating means 229 for creating the node list 246A (or the side list 246B in addition to the node list 246A), the output means 230, the company evaluation information storage means 245, and the output element storage means 246 include: This is the same as the extracted company list editing means 28, output element creation means 29, output means 30, company evaluation information storage means 45, and output element storage means 46 (see FIG. 6) of the first embodiment. Therefore, the company evaluation information (for example, sales) used when setting the size of the company group node in the conversion process is the same as that in the first embodiment.

  Regarding the output means 230, when drawing an edge in the network expression, the similarity between company groups and business attributes included in the extracted company list 244A (similar to FIG. 5) (the edge list 246B of the extracted company list 244A). It is also possible to adopt a configuration in which the type, thickness, color, etc. of the sides are determined using a portion that also serves as the

  As shown in FIG. 10, the appearance probability storage means 243 stores the appearance probability Q of each business attribute in each company group calculated by LDA. The specific data holding form is arbitrary, and assuming that n = the number of company groups and p = the number of business attributes, the appearance probability Q (i, j) is an n-by-p matrix (i = 1 to n, j = 1 to p), each value of Q (i, j) is stored in an array, and the correspondence between the row number and the identification information of the corporate group (here, the corporate group name), And the correspondence between the column number and the identification information of the business attribute (in this case, the name of the business attribute). In addition, a database including a table with rows as records and columns as columns may be prepared.

  Similarly to the case of FIG. 5, the extracted company list storage unit 244 includes, as the extracted company list 244 </ b> A, company identification information (here, the company name is used as an example) and identification information of the company group to which the company belongs ( Here, the company group name is used as an example, and the company group name is the same as the parent company name.) And the business attribute identification information of the company (here, the business attribute name is used as an example.) And the similarity between company groups and business attributes (which is specific to the case of the second embodiment) and the company evaluation information (for example, sales of the company group) are stored in association with each other. is there. In the second embodiment, each record of the extracted company list 244A is similar to the extracted similar company (the similar business attribute extraction unit 226 extracted from the similar company group extracted by the similar company group extraction unit 225). A company having a business attribute), a similar company group to which the similar company belongs, a similar business attribute (in the second embodiment, the business attribute including the target business attribute), and a similar business attribute extracting means 226 This is set data including the similarity between company groups and business attributes, which is the appearance probability Q when the similar business attributes are extracted.

  <Flow of processing by company information providing system 200>

  In the second embodiment, corporate information provision processing is performed by the corporate information provision system 200 as follows.

  In FIG. 11, the processes of steps S21 and S22 are the same as the processes of steps S1 and S2 of the first embodiment (see FIG. 7).

  Subsequently, in the company information providing server 220, the occurrence probability calculation unit 223 uses the company group / business attribute co-occurrence matrix X (similar to FIG. 3) stored in the matrix storage unit 242, and the potential Dirichlet distribution method ( LDA) calculates the appearance probability Q of each business attribute in each company group, and stores the calculated appearance probability Q in the appearance probability storage means 243 (see FIG. 10) (step S23). The appearance probability Q is calculated when the company group / business attribute co-occurrence matrix X is created or updated. The calculation timing may be given from the administrator terminal 3 by the system administrator. In addition, the processing by the matrix creating means 222 described above (step S22) may be linked. Since the content of the calculation process of the appearance probability Q has already been described in detail, the description is omitted here.

  Thereafter, in the state where the above pre-processing is completed, in the company information providing server 220, the target business attribute (business to be analyzed) transmitted from the user terminal 2 via the network 1 by the user input receiving unit 224. The user input including (attribute) is received (step S24). The process of step S24 is substantially the same as the process of step S4 of the first embodiment (see FIG. 7). However, in the first embodiment, the user input receiving unit 24 uses the similar business attribute extracting unit 25. In contrast to accepting the input of the number of similar business attribute extractions f used in processing (more precisely, the number of similar business attribute extractions f plus one-tenth of the target business attribute) In the second embodiment, the user input accepting unit 224 extracts the number f2 of similar business attributes used in the processing by the similar business attribute extracting unit 226 (in the second embodiment, the target business attributes are included and referred to as similar business attributes). Is different in that it accepts input. Further, the user input receiving unit 224 may receive an input of the number f1 of extracted similar company groups used in the processing by the similar company group extracting unit 225.

  Subsequently, in the enterprise information providing server 220, the enterprise group having a high appearance probability for the target business attribute received by the user input receiving unit 224 by the similar company group extracting unit 225 is displayed in the appearance probability storage unit 243 (see FIG. 10). A predetermined number f1 is extracted as a similar company group in descending order of the stored appearance probabilities Q (step S25). Since the content of the extraction process of the similar company group has already been described in detail, the description is omitted here. The number f1 of similar company group extractions may be the number of user inputs received by the user input receiving unit 224.

  Further, the company information providing server 220 uses the similar business attribute extracting unit 226 to display the business attribute having a high appearance probability for each similar company group extracted by the similar business group extracting unit 225 as the appearance probability storage unit 243 (see FIG. 10). ), The number f2 of items received by the user input receiving unit 224 is extracted as similar business attributes in descending order of appearance probability Q (step S26). The target business attributes are also included in the f2 similar business attributes extracted here. Note that the number f2 of similar business attribute extractions may be a predetermined number. Since the content of the similar business attribute extraction process has already been described in detail, the description thereof is omitted here.

  Then, the company information providing server 220 uses the extracted company list creating means 227 to extract the similar business among the companies belonging to the similar company group extracted by the similar company group extracting means 225 from the company information storage means 241 (similar to FIG. 2). A company having a similar business attribute extracted by the attribute extraction unit 226 is extracted as a similar company, and an extracted company list 244A including the extracted similar company, a similar company group to which this similar company belongs, and set data including the similar business attribute is created. Then, the created extracted company list 244A is stored in the extracted company list storage unit 244 (similar to FIG. 5) (step S27). The similar business attributes referred to in the second embodiment include target business attributes.

  The subsequent steps S28, S29, and S30 are the same as the steps S8, S9, and S10 of the first embodiment (see FIG. 7).

  <Effects of Second Embodiment>

  According to such 2nd Embodiment, there exist the following effects. That is, the company information provision system 200 uses the company group / business attribute co-occurrence matrix X (similar to FIG. 3) and the occurrence probability Q () of each business attribute in each company group by the latent Dirichlet distribution method (LDA). (See FIG. 10), and using the appearance probability Q, f2 similar business attributes are extracted for the target business attributes input by the user, and based on this, the extracted company list 244A (similar to FIG. 5) Is created, and the extracted company list 244A is used to output the company information. Therefore, even a user who lacks search know-how and industry structure knowledge can easily search for company information.

  Further, since the company information providing system 200 includes the matrix creating unit 222, even when the data content stored in the company information storage unit 241 (similar to FIG. 2) is updated, a new company group / business attribute is created. A co-occurrence matrix X (similar to FIG. 3) can be easily prepared.

  Furthermore, since the company information providing system 200 includes the appearance probability calculating means 223, the data content stored in the company information storage means 241 (similar to FIG. 2) is updated, and the company group / business attribute co-occurrence matrix X is updated. Even when a change occurs in the content of (as in FIG. 3), a new appearance probability Q (see FIG. 10) can be easily prepared according to the changed company group / business attribute co-occurrence matrix X.

  And since the output means 230 performs the output by the network expression which connected the enterprise group node and the business attribute node by the edge, the relationship between the enterprise group and the business attribute is visualized as a network plot. Gain an overview of the industry structure of the field. For this reason, for example, it is possible to support the direction of the advisory of the company research work for the extracted company.

  Further, the company information providing system 200 includes a company evaluation information storage unit 245, an extracted company list editing unit 228, and an output element creating unit 229. The output element creating unit 229 includes an extracted company list 244A (similar to FIG. 5). Since the size of the company group node is set using the added company evaluation information (for example, sales amount), the output unit 230 has the node list 246A (see FIG. 6) created by the output element creation unit 229. The size of the company group node can be changed in accordance with the numerical value of the company evaluation information when executing the output processing by the network expression according to the setting information of For this reason, in addition to an overview of the target business and the surrounding industry structure, it is also possible to visually grasp the size and status of the corporate group.

  Furthermore, the output element creation unit 229 receives the company evaluation information (for example, sales amount) related to the user designation received by the user input receiving unit 224, that is, the company evaluation information selected and specified by the user from a plurality of types of company evaluation information. It is configured to set the size of the company group node, so that the user can visually grasp the company group from various viewpoints, making it easier to grasp the company information from various perspectives. Can be realized.

  Then, the output element creation unit 229 is configured to make a setting for easily distinguishing the analysis target related to the user's designation and the other color schemes of the company group node and the business attribute node. Therefore, the user can more easily view the industry structure of the target business and its peripheral fields.

  Further, since the output unit 230 performs the balloon display process, the user can easily grasp the information of the companies belonging to the company group when necessary, while overlooking the industry structure of the target business and its peripheral fields. it can.

  [Deformation form]

  Note that the present invention is not limited to the above-described embodiments, and modifications and the like within a range in which the object of the present invention can be achieved are included in the present invention.

  For example, in each of the above-described embodiments, the company information providing systems 10 and 200 are server / client type systems that communicate via the network 1, but the company information providing system of the present invention is a stand-alone system. It is good also as a system.

  As described above, the company information providing system and program according to the present invention uses, for example, business segment classification labels of company information data for the purpose of multilaterally analyzing the similarity between companies in the company research business. By doing so, it is suitable for use when scoring the similarity of each company from the viewpoint of business composition.

10,200 Company information providing system 22,222 Matrix creating means 23 Industry similarity calculation means 24,224 User input accepting means 25 Similar business attribute extracting means 26 Similar company extracting means 27,227 Extracted company list creating means 28,228 Extraction Company list editing means 29,229 Output element creation means 30,230 Output means 41,241 Company information storage means 42,242 Matrix storage means 43 Industry-specific similarity storage means 44,244 Extracted company list storage means 44A, 244A Extracted company list 45,245 Company evaluation information storage means 46,246 Output element storage means 46A, 246A Node list 223 Appearance probability calculation means 225 Similar enterprise group extraction means 226 Similar business attribute extraction means 243 Appearance probability storage means X Company group / business affiliation The probability of occurrence of each business attribute in the co-occurrence matrix S industry similarity between Q each group of companies that show the degree of similarity between each business attribute

Claims (11)

A company information providing system configured by a computer that provides company information,
A company information storage means for storing a company, a business attribute of the company, and a company group to which the company belongs;
Using a corporate group / business attribute co-occurrence matrix in which one of the rows or columns of the matrix is a corporate group, the other is a business attribute, and the number of companies having each business attribute for each corporate group is an element of the matrix, An inter-industry similarity storage means for storing inter-industry similarity indicating the similarity between business attributes, calculated by cosine similarity,
User input receiving means for executing processing for receiving user input including the target business attribute;
The business attributes similar to the target business attribute received by the user input receiving means are the number of cases received by the user input receiving means or in advance in descending order of the similarity between the industries stored by the inter-industry similarity storage means. A similar business attribute extracting means for executing a process of extracting a predetermined number of similar business attributes as a similar business attribute,
The company having the target business attribute and the company group to which the company belongs, and the company having the similar business attribute extracted by the similar business attribute extraction unit and the company group to which the company belongs are similar to the company information storage unit. A similar company extraction means for executing a process of extracting as a company and a company group to which the similar company belongs;
The target business attribute, the group data including the similar company extracted by the similar company extracting unit and the company group to which the similar company belongs, the similar business attribute, the similar company extracted by the similar company extracting unit, and An extracted company list creating means for executing a process of creating an extracted company list including set data including a company group to which the similar company belongs;
Using the extracted company list created by the extracted company list creating means, screen display, print, three-dimensional display, or a combination thereof showing the relationship between the company group and the business attribute and / or the relationship between the company and the business attribute Or an output means for executing an output process based on a combination of these and voice output. Matrix storage means for storing the company group / business attribute co-occurrence matrix;
By using the company group / business attribute co-occurrence matrix stored in the matrix storage means, the similarity between industries indicating the similarity between each business attribute is calculated by the cosine similarity, and the calculated similarity between the industries The business information providing system according to claim 1, further comprising: an inter-industry similarity calculation unit that executes a process of storing the inter-industry similarity storage unit. A company information providing system configured by a computer that provides company information,
A company information storage means for storing a company, a business attribute of the company, and a company group to which the company belongs;
Using a corporate group / business attribute co-occurrence matrix in which one of the rows or columns of the matrix is a corporate group, the other is a business attribute, and the number of companies having each business attribute for each corporate group is an element of the matrix, Appearance probability storage means for storing the appearance probability of each business attribute in each company group, calculated by the potential Dirichlet distribution method,
User input receiving means for executing processing for receiving user input including the target business attribute;
With respect to the target business attributes received by the user input receiving unit, the number of cases received by the user input receiving unit or a predetermined number of company groups having a high appearance probability in descending order of the appearance probabilities stored in the appearance probability storage unit. A similar company group extraction means for executing a process of extracting as many similar company groups as a similar company group,
For each of the similar company groups extracted by the similar company group extracting means, the number of business attributes having a high appearance probability received by the user input receiving means in descending order of the appearance probabilities stored in the appearance probability storage means Alternatively, similar business attribute extraction means for executing a process of extracting a predetermined number of similar business attributes as similar business attributes,
Of the companies belonging to the similar company group extracted by the similar company group extracting unit, the company having the similar business attribute extracted by the similar business attribute extracting unit is extracted as a similar company from the company information storage unit, and extracted. An extracted company list creating means for executing a process of creating an extracted company list including set data including the similar company, the similar company group to which the similar company belongs, and the similar business attribute;
Using the extracted company list created by the extracted company list creating means, screen display, print, three-dimensional display, or a combination thereof showing the relationship between the company group and the business attribute and / or the relationship between the company and the business attribute Or an output means for executing an output process based on a combination of these and voice output. Matrix storage means for storing the company group / business attribute co-occurrence matrix;
Using the company group / business attribute co-occurrence matrix stored in the matrix storage means, the occurrence probability of each business attribute in each company group is calculated by the potential Dirichlet distribution method, and the calculated appearance probability is calculated as the occurrence probability. The company information providing system according to claim 3, further comprising appearance probability calculating means for executing processing to be stored in the probability storage means. The output means includes
Using the extracted company list created by the extracted company list creating means, the company group node corresponding to the company group and the business attribute node corresponding to the business attribute are represented by the existence of the relationship between the company group and the business attribute. The company information providing system according to any one of claims 1 to 4, wherein an output process is performed by screen display, printing, or stereoscopic display using a network representation connected by the indicated side. A company evaluation information storage means for storing a company and financial information, management indicators, or other company evaluation information about the company in association with each other;
About the parent company of the company group acquired from the company evaluation information storage means as the company evaluation information for each company group included in each set of data constituting the extracted company list created by the extracted company list creation means Company evaluation information, or the company evaluation information obtained by totaling or aggregating the company evaluation information of the parent company and affiliated company of the company group, and adopting the adopted company evaluation information for each group Extracted company list editing means for executing processing to be added to data,
Using the extracted company list with the company evaluation information added by the extracted company list editing means, the size of the company group node corresponding to each company group included in each set of data constituting the extracted company list And processing for creating a node list including setting information of size and color of the business attribute node corresponding to each business attribute included in each set of data constituting the extracted company list Output element creating means for
The output element creation means includes
For the size of the company group node, execute a setting process to be a value after conversion using the company evaluation information for the company group,
Alternatively, regarding the size of the company group node corresponding to the company group that constitutes the set data including the same business attribute as the target business attribute among the size of the company group node, the company for the company group The value after conversion using the evaluation information, and the size of the other company group node is configured to execute a setting process to be a fixed value,
The output means includes
The company information providing system according to claim 5, wherein the output processing based on the network expression is executed according to the setting information of the node list created by the output element creating means. The user input receiving means
When a plurality of types of company evaluation information is stored in the company evaluation information storage means, the process is configured to execute a process of receiving user input for designating which company evaluation information is used,
The output element creation means includes
The company information provision according to claim 6, wherein a setting process for calculating a value after conversion using the company evaluation information specified by the user received by the user input receiving unit is executed. system. The output element creation means includes
Regarding the color of the company group node, the color of the company group node corresponding to the company group constituting the set data including the same business attribute as the target business attribute is different from the color of the other company group node. Color and
The size of the business attribute node is a fixed value,
The color of the business attribute node is configured to execute a setting process in which the color of the business attribute node corresponding to the target business attribute is different from the colors of the other business attribute nodes. The system for providing corporate information according to claim 6 or 7, wherein: The user input receiving means
It is configured to execute a process of accepting a user input indicating the company group node in a screen display or stereoscopic display state,
The output means includes
The company belonging to the company group corresponding to the company group node that received the user instruction by the user input receiving means is acquired from the extracted company list, and the acquired company name is displayed on the screen in the vicinity of the company group node. Or it is set as the structure which performs the balloon display process which displays a screen or three-dimensionally in the balloon displayed in three dimensions. The company information provision system in any one of Claims 5-8 characterized by the above-mentioned. Using the association between the company stored in the company information storage means, the business attribute of the company and the company group to which the company belongs, the same company group for the brute force combination of each company group and each business attribute Matrix creation means for executing the process of creating the company group / business attribute co-occurrence matrix by extracting companies having the same business attribute and having the number of extracted companies as an element of the matrix The company information providing system according to any one of claims 1 to 9, wherein the system is provided.   The program for functioning a computer as a company information provision system in any one of Claims 1-10.

Images