JP7474120B2 - Business matching support device and method - Google Patents

Description

The present invention relates to a business matching support device and a business matching support method, and more specifically, to a technology that enables accurate matching of companies that meet the matching needs of a company that has made a matching request.

As the business environment surrounding companies has become tougher in recent years, there has been an increase in cases where companies are working together with other companies to solve problems, such as expanding sales and purchasing destinations and developing new products. Business matching is a service that introduces companies to business partners who can provide the above-mentioned cooperation, and is carried out mainly by financial institutions. When introducing companies, it is important to introduce companies that are likely to close a deal, and for this purpose, methods of business matching support that include company evaluation have been proposed.

As a conventional technology for supporting such business matching, for example, there is a company selection method (Patent Document 1) in which a company search request is received from a company requesting matching, specifying the conditions for the requested company, and a search based on the conditions is requested to local coordinators who regularly exchange information with potential requestee companies. Recommendations for companies that meet the conditions are then received from the local coordinators, and the initial recommendation results are returned to the requesting company. Thereafter, a company is selected by repeating the process of the requester specifying additional conditions and receiving company recommendations from the coordinators several times.

Patent No. 6430151

In conventional technology, recommending a wide range of potential matching companies requires a significant amount of work on the part of the person making the request.

The background to this is that in actual business negotiations and matching, it is also important to match with latent needs and conditions that do not appear beforehand. For example, even if the requesting company wishes to purchase marine products, if the purpose is to serve sushi, there is also latent need for wasabi, ginger, tea, and other items that accompany sushi, and so a match is possible. In this way, there is a relationship between explicit needs and latent needs (hereafter referred to as latent needs relationship).

With conventional technology, the person making the request must first clarify their company's needs, which they input into the system and communicate to the coordinator. Therefore, in order to respond to potential needs, it is necessary to identify potential needs in advance and define each of them as conditions. This is a large amount of work for each requesting company to do.

The object of the present invention is to provide a business matching support device and method that enables business matching that selects a wide range of candidate companies based on the potential needs of the requesting company.

To solve the above problem, an example of a matching support device for supporting business matching between companies according to the present invention includes a data acquisition unit that acquires data describing information about each of a plurality of companies, a text information extraction unit that extracts words that represent the characteristics of each company from the data acquired by the data acquisition unit, a storage unit that stores latent needs structure information that manages the relationship between needs and latent needs as a structure for each type of latent needs-related information, a latent needs-related information structuring unit that reads a structure from the storage unit according to the type of latent needs-related information selected, and structures the latent needs-related information based on the input needs and latent needs for the read structure, and a candidate company evaluation unit that calculates an evaluation value of the ease of matching between companies from the structured latent needs-related information structured by the latent needs-related information structuring unit and the words extracted by the text information extraction unit.

The present invention makes it possible to match businesses with a wide range of potential companies based on their potential needs.

FIG. 1 is a diagram illustrating an example of a schematic configuration of an information processing system. FIG. 2 is a diagram illustrating an example of main functions of the support device. FIG. 2 is a diagram illustrating an example of a hardware block of the support device. FIG. 13 is a diagram illustrating an example of a flowchart for explaining a potential needs related information structuring process. FIG. 11 is a diagram showing an example of potential needs related information. FIG. 13 is a diagram showing an example of input contents regarding potential needs related information. FIG. 13 is a diagram showing an example of structured potential needs related information having a vertical structure. FIG. 13 is a diagram showing an example of structured potential needs related information having a horizontal structure. FIG. 13 is a diagram illustrating an example of a flowchart for explaining an evaluation process. FIG. 13 is a diagram illustrating an example of a flowchart illustrating details of an evaluation value generation process. FIG. 13 is a diagram showing an example of a candidate company evaluation list. FIG. 13 is a diagram illustrating an example of a formula for calculating a positive score.

The following describes an embodiment of the present invention with reference to the drawings. In each drawing for describing the embodiment, the same components are given the same names and symbols as much as possible, and repeated explanations are omitted.

The following describes the embodiments with reference to the drawings, but any modifications that can be made to the examples will be described as variations within each example.

1 shows a schematic configuration of an information processing system 1 according to Example 1. The information processing system 1 includes a business matching support device (hereinafter, "support device 10"), one or more Web servers 2, and one or more corporate information servers 3. All of these are information communication devices (computers) and are connected via a communication network 4.
The Web server 2 and the company information server 3 (hereinafter, these devices are collectively referred to as "company information providing devices") provide the support device 10 with various information (company name, industry, business content, research content, owned facilities, owned assets, number of employees, partner companies, suppliers, customers, transaction history, sales performance, advertising, etc.) on companies that may be candidates for company-to-company matching.
The support device 10 provides information regarding the evaluation of companies (business partners) that are potential candidates for matching requests, based on information collected from a company information collection device via the communication network 4.

Figure 2A shows the main functions of the support device 10. As shown in Figure 2A, the support device 10 has the functions of a data acquisition unit 111, a potential needs related information structuring unit 112, a text information extraction unit 113, and a candidate company evaluation unit 114.

The storage device 25 stores structured latent needs related information 151, company text information 152, candidate company evaluation list 153, latent needs structure information T40, and latent needs related information T50.

The structured latent needs related information 151 manages information between words that represent products and services of a company. The structured latent needs related information 151 is information acquired, for example, by operational input from a user. The structured latent needs related information 151 will be described later with reference to Figures 6 and 7.

Company text information 152 is information obtained from a company information providing device and includes text data that describes information about the company.

The candidate company evaluation list 153 includes, for companies that are candidates for matching requests, information indicating the evaluation value of the success rate, the products and services of the companies that affect the evaluation value, and the association between words that represent the products and services. The candidate company evaluation list 153 is information such as that shown in FIG. 10, and is generated by the support device 10.

The latent needs structure information T40 manages the relationship between needs and latent needs as a structure for each type of latent needs-related information, and is explained using Figure 4.

The potential needs-related information T50 is information that corresponds and manages the needs 52, potential needs 53, types of potential needs-related information 54, and conditions 55 that are input by the user to the support device 1, and details are described using FIG. 5.

Of the functions shown in FIG. 2A, the data acquisition unit 111 acquires text data describing information about companies from the company information providing device via the communication network 4. The data acquisition unit 111 acquires the needs words of the requesting company and the seeds words of candidate companies from the acquired information about each of the multiple companies. The data acquisition unit 111 acquires information from the company information providing device via the communication network 4 at any time, for example, by performing crawling or scraping.

The potential needs related information structuring unit 112 supports the user's input format, structures the selection know-how, and stores it in the storage device 25. The potential needs related information structuring unit 112 receives from the user the type of potential needs related information, a word expressing a need, a word expressing a potential need, the type of potential needs related information, and the conditions under which the potential needs related information is valid. The existing needs related information structuring unit 112 then stores each type of potential needs related information in a format suited to that type, thereby generating structured potential needs related information 151 (see Figures 6 and 7).

The text information extraction unit 113 collects the needs of the requesting company and the seeds of candidate companies from the text information, and extracts words that indicate the characteristics of each company from the text data obtained from the company information provider device.

The candidate company evaluation unit 114 evaluates the combination of the requesting company's needs and the candidate company's seeds based on the selection know-how stored as structured latent needs related information 151 to determine whether they match the requesting company's latent needs, and recommends candidate companies based on the scores between the companies. Therefore, the candidate company evaluation unit 114 lists the words indicating the characteristics of the requesting company and the matching candidate companies (words indicating needs, words indicating seeds) and the structured latent needs related information 151 that are extracted from the requesting company and the matching candidate companies, the evaluation value calculated for each matching candidate company, and the latent needs related information that had the greatest impact on the calculated evaluation value, to generate a candidate company evaluation list 153 (see FIG. 10).

Next, the processing performed by the support device 10 will be described. The main processing performed by the support device 10 is broadly divided into processing for structuring latent needs-related information (hereinafter referred to as "latent needs-related information structuring processing S300") and processing for evaluating each company (evaluating their suitability as a matching destination) using latent needs-related information for matching requests received from users (hereinafter referred to as "evaluation processing S800").

FIG. 2B is a diagram showing a hardware configuration of the support device 10 of FIG.
The support device 10 has a processor 21, a memory 22, a communication device 23, an input/output device 24, a storage device 25, and an internal signal line 26. The input/output device 24 is connected to a monitor 27, a keyboard 28, and a mouse 29.

The memory 22 is composed of a volatile storage medium such as a DRAM, and stores a data acquisition program 111, a latent needs-related information structuring program 112, a text information extraction program 113, and a candidate company evaluation program 114. The programs stored in the memory 22 are executed by the processor 21, which is a processing unit, to realize the functions of the data acquisition unit 111, the latent needs-related information structuring unit 112, the text information extraction unit 113, and the candidate company evaluation unit 114 described in FIG. 2A.

The storage device 25 is composed of a non-volatile storage medium such as a hard disk drive, and stores structured latent needs related information 151, company text information 152, candidate company evaluation list 153, latent needs structure information T40, and latent needs related information T50.

Figure 3 is a flowchart explaining the latent needs-related information structuring process S300 executed by the latent needs-related information structuring unit 112. The support device 10 executes process S300 to generate structured latent needs-related information 151.

First, the user inputs four pieces of information (S311) as latent needs-related information (T50) shown in FIG. 5: a word expressing a need (need 52), a word expressing a latent need (latent need 53), a type of latent needs-related information 54, and a condition 55 for which the latent needs-related information is satisfied. Hereinafter, the word expressing a need 52 may be referred to simply as need 52, and the word expressing a latent need 53 may be referred to as latent needs 53.

The type 54 of potential needs related information is selected by the user from the types 42 of the potential needs structure information T40 (Figure 4) that have been registered in advance.

Figure 4 shows an example of latent needs structure information T40.

The potential needs structure information T40 manages the relationship of potential needs to needs as a structure for each type of potential needs-related information. The potential needs structure information T40 is input in advance by the user and stored in the storage device 25.

The potential needs structure information T40 manages the correspondence between the identifier 41 and the type 42, structure 43, and interactivity 44 of the potential needs related information.

Type 42 of latent needs related information indicates the relationship between needs 52 and latent needs 53, and is selected when the user inputs latent needs related information T50 in FIG. 5. For example, when "Relationship between products and raw materials" of type 42 of latent needs related information is selected as the latent needs related information T50 input by the user, "Relationship between products and raw materials" is registered in type 54 of latent needs related information, as shown by identifier #2-1 in FIG. 5.

The structure 43 in Figure 4 indicates whether the structure is a vertical structure or a horizontal structure depending on the type 42 of the potential needs related information. In a vertical structure, the relationship between needs and potential needs is one-way, while in a horizontal structure, the relationship between needs and potential needs is two-way and can be interchanged. The vertical and horizontal structures themselves are stored in the storage device 25, although not shown. The vertical structure will be described later using Figure 6, and the horizontal structure will be described later using Figure 7.

Figure 5 shows an example of potential needs-related information T50, which is the input content in S311.

The identifier 51 corresponds to each piece of latent needs-related information input by the user. For ease of understanding, the identifier 51 in FIG. 5 is written corresponding to the identifier 41 in the latent needs structure information T40, but there does not necessarily have to be a correspondence.

In each row corresponding to an identifier 51, needs 52, potential needs 53, type of potential needs-related information 54, and conditions 55 are managed as potential needs-related information.

Needs 52, potential needs 53, type of potential needs-related information 54, and conditions 55 are information input by the user. Here, type of potential needs-related information 54 is selected from type of potential needs-related information 42 in FIG. 4.

Conditions 55 specify the conditions required for a company to be a candidate for matching, such as the condition for the potential needs-related information to be valid, for example, when identifier 51 is #4-1, the company operates a sushi restaurant as a service, and when identifier 51 is #2-1, the company manufactures its own products.

Returning to FIG. 3, the support device 10 generates structured latent needs-related information 151 based on the latent needs-related information T50 received from the user (S312).

Specifically, the needs and potential needs are added to the data structure 43 corresponding to the type 42 of potential needs-related information selected by the user. The potential needs-related information structuring unit 112 reads the structure 43 from the storage device 25 according to the type of potential needs-related information selected from the potential needs structure information, and structures the potential needs-related information by inputting the needs 52 and potential needs 53 into the read structure 43.

When the relationship between needs and potential needs is one-way, the latent needs related information structuring unit 112 reads out a vertical structure as shown in FIG. 6 from the storage device 25, and when the relationship between needs and potential needs is two-way, the latent needs related information structuring unit 112 reads out a horizontal structure as shown in FIG. 7 from the storage device 25. The latent needs related information structuring unit 112 then inputs the input needs and potential needs into a vertical structure or horizontal structure depending on the type of latent needs related information.

Details will be explained using Figures 6 and 7.

FIG. 6 shows an example of the vertical structure of the structured potential needs related information 151. As shown in FIG.
Here, we use as an example the type 54 of latent needs related information "relationship between category and specific object" managed corresponding to identifiers #1-1 to #1-6 in Fig. 5. Each arrow corresponds to one piece of stored latent needs related information, with the starting point being the needs side and the end point being the latent needs side.

First, when the user inputs "relationship between categories and specific objects" as the type 54 of latent needs-related information, the latent needs structure information T40 in FIG. 4 identifies it as a vertical structure 43, and the vertical structure as shown in FIG. 6 is read from the storage device 25.

For the vertical structure in Figure 6, the need 52 is entered as "beverages" 60 and the potential need 53 is entered as "tea" 61 for identifier #1-1 in Figure 5.

Furthermore, for identifier #1-2, the need 52 is entered as "beverages" 60 and the potential need 53 is entered as "alcohol" 62.

Furthermore, for identifier #1-3, need 52 is entered as "tea" 61 and potential need 53 is entered as "green tea" 63.

Furthermore, for identifier #1-4, need 52 is entered as "green tea" 63 and potential need 53 is entered as "matcha" 64.

Furthermore, for identifier #1-5, the need 52 is entered as "alcohol" 62 and the potential need 53 is entered as "distilled spirits" 65.

Furthermore, for identifier #1-6, the need 52 is entered as "beverage" 60 and the potential need 53 is entered as "coffee" 66.

In this way, structured potential needs related information 151 as shown in Figure 6 is created using the words entered as needs 52 and the words entered as potential needs 53 in the vertical structure.

Here, in the vertical structure, words that are not potential needs in any potential need-related information are set to "depth 0," and potential needs in potential need-related information for which a word with depth 0 is a need are set to "depth 1," and so on to determine "depth." For example, in FIG. 6, "beverage" 60 is not a potential need, so it has a "depth 0," and potential needs "tea" 61, "alcohol" 62, and "coffee" 66 for which "beverage" 60 is a need have a "depth 1." Furthermore, potential needs for "distilled spirits" 65 and "wine" for "alcohol" 62 have a "depth 2," and potential need "vodka" for distilled spirits 65 has a "depth 3."

Figure 7 shows an example of the horizontal structure of structured potential needs related information 151. For example, a structure storing "relationships provided by the same service" and a structure storing "spelling variations" of identifier #4 in Figure 4 are shown. As shown in Figure 4, "relationships provided by the same service" of type 42 of potential needs related information has a horizontal structure 43. "Spelling variations" use a known spelling variation dictionary to show spelling variations for needs 52 and potential needs 53 entered by the user.

As latent needs related information T50 in FIG. 5, for identifier #4-1, "Relationships provided by the same service" is selected as type 54 of latent needs related information, and "tuna" 73 is input as need 52 and "ginger" 72 is input as latent need 53. Furthermore, for identifier #4-2, "tuna" 73 is input as need 52 and "green tea" 74 is input as latent need 53.

In addition, the structured latent need related information 151 manages "shouga" 75 in a horizontal structure as a spelling variation of "shouga" 72 input as a latent need 53 by referring to a spelling variation dictionary.

When actually stored in an information communication device, the structured latent needs-related information 151 can be realized by, for example, associating each word with a list of latent needs-related information for which that word is a need, and collecting multiple lists of associated words and latent needs-related information. The structured latent needs-related information 151 is stored as a separate structure for each type of latent needs-related information.

The latent needs-related information structuring unit 112 constructs structured latent needs-related information 151 based on the latent needs-related information of FIG. 5 input by the user. For example, when type 54 of latent needs-related information is input, structure 43 (FIG. 4) corresponding to type 42 of latent needs-related information is read out, and the latent needs-related information of FIG. 5 (needs 52, latent needs 53, etc.) input into the structure portion to construct structured latent needs-related information 151 of FIGS. 6 and 7.

This concludes the explanation of the potential needs-related information structuring process S300 in Figure 3.

Here, a method for structuring latent needs-related information based on user input has been shown, but other methods may also be used, such as automatically acquiring information from menus, language resources, or text data, or automatically acquiring information from past matching logs. The date and time when the needs 52, latent needs 53, type of latent needs-related information 54, and conditions 55 were input for each identifier 51 of the latent needs-related information T50 may also be recorded.

Next, we will explain the evaluation process S800.

Figure 8 is a flowchart showing the evaluation process S800. The support device 10 inputs the structured potential needs related information 151 and the company text information 152, performs this process, and generates a candidate company evaluation list 153.

First, the text information extraction unit 113 extracts words that represent information on the products and services required by the matching requesting company (hereinafter, "needs words") and words that represent information on the products and services that can be provided by the candidate companies (hereinafter, "seeds words") from the company text information 152 by text processing (S811).

As a method for extracting needs words and seeds words in S811, for example, a list of words representing products and services may be created in advance and extracted based on the list, or rules may be established regarding how words representing products and services appear in sentences and extraction may be performed based on the rules.

The candidate company evaluation unit 114 then uses the needs words and seed words extracted in S811 and the structured latent needs related information 151 to create an evaluation of the likelihood of matching for the pair of the requesting company and the candidate company, and obtains the latent needs related information that influenced the evaluation (S812). The above-mentioned S811 and S812 are performed as a series of steps for all candidate companies. Details of the processing of step S812 are shown in Figure 9. The evaluation processing S800 is performed for all candidate companies, as shown in Figure 8.

As shown in FIG. 9, first, the candidate company evaluation unit 114 searches for needs words and seed words from the structured potential needs related information 151 (S911). S911 and subsequent steps are performed for each pair of needs words and seed words.

For example, in the case of a vertical structure, a search can be performed by recording a list of words at "depth 0" in advance, retrieving words at "depth 1" that are potential needs (potential need-related information) for which those words are needs, and repeating the process.

In the case of a horizontal structure, a group of words connected by latent needs-related information can be regarded as one group, and a representative word for that group can be stored in advance, allowing latent needs to be similarly obtained from the needs in the latent needs-related information.

Through the above search, the candidate company evaluation unit 114 obtains in which structure of the type of latent needs-related information in the structured latent needs-related information 151 the needs words and seeds words exist, and also the depth of the vertical structure.

Then, the candidate company evaluation unit 114 determines whether or not both a needs word and a seeds word were present in S911 (S912). If not, the process proceeds to S917, which will be described later.

Here, we will first describe the case where both a needs word and a seeds word are present in S912. The variable N is initialized to N=1 (S913). In addition, the "zeroth-order latent needs word" is defined as a needs word.

Then, the candidate company evaluation unit 114 acquires all latent needs-related information from all types of structured latent needs-related information 151 that has the "(N-1)th latent needs word" as a need and has a word that has not been acquired before as a latent need, and acquires the latent needs in that latent needs-related information as the "Nth latent needs word" (S914).

To explain based on the structured latent needs related information in FIG. 6, for example, in step S912, "beverage" is defined as the needs word for the "0th order latent needs word." "Tea," "alcohol," and "coffee" are obtained as "Nth order latent needs words." If there are conditions for the latent needs related information, the information of the requesting company in the company text information 152 is searched for those conditions, and the information is used only if the conditions are met.

Then, the candidate company evaluation unit 114 determines whether the "Nth potential needs word" has no words that have not been obtained as potential needs words up to the (N-1)th order, or has a seed word (S915). If either of the above is true, the process proceeds to S917, which will be described later.

Here, we first describe the case where none of the conditions are met in S915. In this case, the candidate company evaluation unit 114 updates N to N+1 (S916) and returns to S913. That is, the next potential need is acquired. For example, the potential needs for "distilled spirits" and "wine" for "alcohol" acquired in S914 of the first loop are acquired in step S914 of the next loop.

In step S917, if the determination in step S912 is NO, i.e., if neither the needs word nor the seeds word is found in the search, if a new word cannot be obtained in step S915, or if a seeds word is obtained, the candidate company evaluation unit 114 calculates a score for the pair of needs word and seeds word (S917). Depending on the process leading up to S917, it is divided into the following three cases. This score is information for calculating an evaluation value between the requesting company and the candidate company.

The first is when the needs word and seed word are each present in some type of structured latent needs-related information 151 in S912, and the seed word can be obtained in S914. In this case, latent needs-related information linking the needs and seeds has been found, so a positive score is given. Furthermore, the type of latent needs-related information in which the needs word and seed word exist, and in the case of a vertical structure, its depth, as well as some latent needs and passed latent needs-related information from the primary latent needs are obtained.

The second case is when the needs word and the seeds word are both present in some type of structured latent needs-related information 151 in S912, but the seeds word cannot be obtained in S914. For example, this is the case when all the vertically structured latent needs-related information is processed downwards, and no word expressing a seed is found up to the deepest latent need. In this case, a score of 0 is assigned because no latent needs-related information linking the needs and the seeds has been found.

The third case is when, in S911, either the needs word or the seeds word, or both, are not present in any type of structured latent needs-related information 151. In this case as well, a score of 0 is assigned because no latent needs-related information linking the needs and seeds has been found.

In this way, the candidate company evaluation unit 114 sets the score to 0 if both the needs word and the seed word are not present in the structured latent needs related information 151 structured by the latent needs related information structuring unit 112. Also, if both the needs word and the seed word are present in the structured latent needs related information 151 and the seed word is obtained from the needs word in the structured latent needs related information 151, the score is set to positive. Also, the candidate company evaluation unit 114 sets the score to 0 if both the needs word and the seed word are present in the structured latent needs related information 151 and the seed word is not obtained from the needs word in the structured latent needs related information 151.

When a positive score is given, for example, if the needs word is in the structured latent needs-related information 151 of the "relationship between categories and specific objects," latent needs in the latent needs-related information are more specific than needs, so by giving a high score when the depth of the needs word is deep, more specific needs can be highly evaluated. Also, since the further one traces the latent needs-related information, the further one gets from the original needs word, the score can be reduced according to the number of latent needs that have passed.

It is also possible to change the weighting for each piece of latent need-related information and reflect this in the score. For example, based on the date and time information stored for each piece of latent need-related information, the score can be lowered if old latent need-related information is used.

For example, the calculation of the positive score can be performed based on the formula 1 shown in FIG. 11.

Here, M is the value of N when the transition occurs from S915 to S9175, that is, M at which the seed word is acquired as the (M+1)th potential need.

WN is a weight according to the type of structured needs-related information from which the Nth-order latent need was obtained. In other words, the evaluation formula is the sum of the weights WN for the series of latent needs up to the point where the seed word is obtained as the (M+1)th-order latent need, and then subtracted from 1.

An execution example will be shown using the examples of Figures 6 and 7. Assume that the needs word is "tuna" and the seeds word is "matcha." First, as a result of the search in S911, it is found that "tuna" exists in the structured latent needs-related information in the "relationship provided by the same service" and "matcha" exists in the "relationship between category and concrete object."

Therefore, in S912, since both the needs word and the seeds word are hit, the process proceeds to step S913. Next, in S913, "ginger" and "green tea" are obtained as primary latent needs from the "relationship of being provided by the same service". In S915, the condition is not met and the process loops. S913 is performed again, and "ginger" is obtained from the primary latent need "ginger" through "spelling variation", and "matcha" is obtained from the primary latent need "green tea" through the "relationship between category and concrete object". In S915, since "matcha" is a seeds word, the process ends and the process transitions to S917.

Next, an example of the calculation of the evaluation value in S917 is shown. Here, the weight W for "relationship between being provided in the same service" is 0.1, and the weight W for "relationship between category and concrete object" is 0.05. In the above execution example, a positive score is given based on the conditions. Furthermore, when obtaining the seed words, the primary latent need was "green tea" obtained from "relationship between being provided in the same service", and the secondary latent need was "matcha" obtained from "relationship between category and concrete object". Therefore, the evaluation value in this case is 1 - (0.1 + 0.05) = 0.85.

The score in S915 is not limited to the above policy, but can be calculated based on the results obtained through S912 and S914.

The above steps up to S915 are carried out for each pair of needs words and seeds words.

Next, the candidate company evaluation unit 114 uses the scores for the pairs of needs words and seeds words to calculate an evaluation of the ease of matching between the requesting company and the candidate companies (S919).

In S919, the candidate company evaluation unit 114 calculates an evaluation between the requesting company and the candidate company based on the score calculated in step S917. For example, the evaluation of the ease of matching can be calculated based on the maximum score or total score for the pair of needs words and seed words, or some other formula. At this time, the pair of needs words and seed words that influenced the evaluation of the ease of matching between the requesting company and the candidate company is selected, and the "latent needs words" and latent needs-related information obtained when S914 was performed on the needs words and seed words are obtained.

As a result of the above, the candidate company evaluation unit 114 obtains, for each candidate company, an evaluation of the ease of matching between the requesting company and the candidate company pair, as well as information related to potential needs that influenced the evaluation.

Returning to FIG. 8, the candidate company evaluation unit 114 generates a candidate company evaluation list 153 using the evaluation of the ease of matching obtained for each candidate company and the potential needs-related information that influenced the evaluation (S813).

Figure 10 is an example of a candidate company evaluation list 153.

101 indicates the ranking of ease of matching. Company name 102 indicates the name of the candidate company, evaluation value 103 is the value calculated in step S919 of FIG. 9, and needs 104 and seeds 105 are registered words searched for in step S911 of FIG. 9. Latent needs related information 106 indicates structured latent needs related information having needs words and seeds words. Each row indicates an evaluation of the ease of matching for each candidate company. The list of candidate company evaluations is displayed on monitor 27 as an evaluation between the requesting company and the candidate companies.

As described above, the support device 10 performs the process using the structured latent needs-related information 151 and company text information 152 as inputs to generate a candidate company evaluation list 153. By generating the latent needs-related information, which is the relationship between needs and latent needs, from the structured latent needs-related information 151, evaluation is performed by reflecting not only textual matches but also various latent needs-related information.

This concludes the explanation of evaluation process S800 in Figure 8.

According to the above embodiment, the latent needs-related information structuring process S300 uses structured latent needs-related information that has been previously structured, and the evaluation process S800 evaluates each company (evaluates their suitability as a matching destination) based on the structured latent needs-related information and the information on each company provided by the company information providing device in response to a matching request received from a user, thereby making it possible to perform business matching with a wide range of candidate companies based on latent needs.

10: Support device 111: Data acquisition unit 112: Latent needs related information structuring unit 113: Text information extraction unit 114: Candidate company evaluation unit 151: Structured latent needs related information 152: Company text information 153: Candidate company evaluation list 21: Processing unit 22: Memory 24: Input/output device 25: Storage device

Claims (5)

A business matching support device that supports business matching between companies,
A data acquisition unit that acquires data describing information regarding each of a plurality of companies;
a text information extraction unit that extracts words that represent characteristics of each company from the data acquired by the data acquisition unit;
A storage unit that stores latent needs structure information that manages the relationship between needs and latent needs as a structure for each type of latent needs related information;
a latent needs related information structuring unit that reads out a structure from the storage unit in accordance with the selected type of latent needs related information, and structures the latent needs related information based on the input needs and latent needs for the read out structure;
a candidate company evaluation unit that calculates an evaluation value of the ease of matching between companies from the structured latent needs-related information structured by the latent needs-related information structuring unit and the words extracted by the text information extraction unit,
The potential needs related information structuring unit
When the relationship between the needs and the potential needs is one-way, the vertical structure is read out from the storage unit;
When the relationship between the needs and the potential needs is two-way, the horizontal structure is read out from the storage unit;
The input needs and potential needs are input into the vertical structure or the horizontal structure according to the type of potential needs-related information;
The data acquisition unit is
From the information on each of the multiple companies obtained, we obtain the needs words of the requesting company and the seeds words of the candidate companies,
The candidate company evaluation unit:
If neither the needs word nor the seed word exists in the structured latent needs related information structured by the latent needs related information structuring unit, the score for calculating the evaluation value is set to 0;
When both the needs word and the seed word exist in the structured latent needs-related information and the seed word is obtained from the needs word in the structured latent needs-related information, the score is set to positive;
The sum of the scores for each pair of needs words and seeds words is calculated and used as an evaluation value for the ease of matching between companies.
A business matching support device comprising: 2. The business matching support device according to claim 1 ,
The candidate company evaluation unit:
A business matching support device characterized in that when both the needs word and the seed word exist in the structured latent needs related information and the seed word is not obtained from the needs word in the structured latent needs related information, the score is set to 0. 3. The business matching support device according to claim 2 ,
A business matching support device comprising a display unit that displays an evaluation value between the requesting company and the candidate companies. 3. The business matching support device according to claim 2 ,
the latent needs-related information structuring unit records the date and time when each latent needs-related information was stored;
the candidate company evaluation unit determines the score depending on the date and time when the used potential needs-related information was stored;
A business matching support device comprising: A business matching support method for supporting business matching between companies using a business matching support device having a data acquisition unit, a text information extraction unit, a potential needs related information structuring unit, a candidate company evaluation unit, and a storage unit, comprising:
the storage unit stores latent needs structure information that manages the relationship of latent needs to needs as a structure for each type of latent needs related information;
The data acquisition unit acquires data describing information about each of a plurality of companies,
The text information extraction unit extracts words that represent characteristics of each company from the data acquired by the data acquisition unit;
the potential needs-related information structuring unit reads a structure from the storage unit in accordance with the selected type of potential needs-related information, and structures the potential needs-related information based on the input needs and potential needs for the read structure;
the candidate enterprise evaluation unit calculates an evaluation value of the ease of matching between enterprises from the structured latent-needs-related information structured by the latent-needs-related information structuring unit and the words extracted by the text information extraction unit ;
The potential needs related information structuring unit
When the relationship between the needs and the potential needs is one-way, the vertical structure is read out from the storage unit;
When the relationship between the needs and the potential needs is two-way, the horizontal structure is read out from the storage unit;
The input needs and potential needs are input into the vertical structure or the horizontal structure according to the type of potential needs-related information;
The data acquisition unit is
From the information on each of the multiple companies obtained, we obtain the needs words of the requesting company and the seeds words of the candidate companies,
The candidate company evaluation unit:
If neither the needs word nor the seed word exists in the structured latent needs related information structured by the latent needs related information structuring unit, the score for calculating the evaluation value is set to 0;
When both the needs word and the seed word exist in the structured latent needs-related information and the seed word is obtained from the needs word in the structured latent needs-related information, the score is set to positive;
When both the needs word and the seed word exist in the structured latent needs-related information, and the seed word is not obtained from the needs word in the structured latent needs-related information, the score is set to 0;
The sum of the scores for each pair of needs words and seeds words is calculated and used as an evaluation value for the ease of matching between companies.
A business matching support method comprising:

Images