JP7452801B2 - Evaluation support system and evaluation support program - Google Patents

Description

The present invention relates to an evaluation support system and an evaluation support program for supporting evaluation of the value of a subject.

In recent years, financial institutions such as banks have responded to drastic changes in the business environment of lending companies and the shift to DX (Digital Transformation) among their clients, and have been able to quantitatively understand the value of their clients and improve their creditworthiness. Asset analysis for ratings, etc. is calculated using artificial intelligence. As such a technique, for example, a financial information analysis system disclosed in Patent Document 1 has been proposed.

In Patent Document 1, a financial information analysis system inputs values of items selected from corporate accounting information as explanatory variables, and calculates a default rate for each company based on a learning model learned using a recurrent neural network. In addition, a technique is disclosed that refers to the calculated company default rate and company default information and uses a learning model to calculate the default rate of a specific company in a target month.

Japanese Patent Application Publication No. 2018-049400

Here, in the disclosed technology of Patent Document 1, when evaluating the value of a target person, in order to calculate the target's DX conversion, reputational risk, transaction status of the target's group companies, related companies, etc. It is necessary to obtain the default rate of each company and default information of each company, prepare a learning model based on various information other than financial information in advance, and calculate the default rate of each company in the target month. . Therefore, as new information other than financial information is acquired and the amount of information increases, there is a concern that users will spend an enormous amount of time working on the system. Under these circumstances, it is desired to reduce the amount of work time spent by users in assessing the value of a subject.

The present invention has been devised in view of the above-mentioned problems, and its purpose is to reduce the work time spent by users in an evaluation support system for supporting the value evaluation of subjects. The objective is to provide an evaluation support system and evaluation support program that can perform the following tasks.

The evaluation support system according to the first invention is an evaluation support system for supporting the value evaluation of a target company that is the target of a loan or an alliance among companies using financial institutions, and is related to the value evaluation. a setting means for setting identification information of an acquisition source for acquiring information, and circulation information indicating an acquisition order in which data is acquired from a plurality of websites of the acquisition destination; information on the target company ; and a value of the target company. an acquisition unit that acquires target data including value information for determining the value based on identification information of the acquisition source and a plurality of the circulation information; and a value information database that stores the target data acquired by the acquisition unit. , refers to a model definition database storing a plurality of evaluation models for evaluating the value of the target company and the value information database, and determines the value of the target company based on the evaluation model stored in the model definition database. The present invention is characterized by comprising an evaluation means for performing evaluation.

In the evaluation support system according to a second invention, in the first invention, the visiting information set by the setting means includes time information regarding website visiting, the number of times of execution, an acquisition position regarding the acquisition location of the target data, an acquisition paragraph, It is characterized by including at least one of an acquired keyword and an excluded keyword .

In the evaluation support system according to a third invention, in the first invention, the target data acquired by the acquisition means includes public information, press information, etc. published by the target company or a third party other than the target company ; It is characterized by including at least one of word-of-mouth information, website trust information, and DX information regarding digitalization efforts.

In the evaluation support system according to a fourth aspect, in the fourth aspect, the evaluation models stored in the model definition database are classified according to classification information including at least characteristics or types of the acquired data and websites preset by the user. A machine that uses learning data that has an association between the value evaluation results of past value evaluations and the reference information linked to the past value evaluation results. It is characterized by including at least one of an evaluation analysis model that performs evaluation using a learning model constructed through learning, and a comprehensive score model that performs evaluation using a total score calculated using the attention evaluation model and the evaluation analysis model. do.

In the fourth invention, the evaluation support system according to a fifth invention provides an update to reflect the relationship in the association when a relationship between the past value evaluation result and the reference information is newly acquired. The method further comprises means.

In the evaluation support system according to a sixth invention, in the fourth invention, when the evaluation means performs the evaluation using the attention evaluation model, the evaluation means determines that the reliability is equal to or higher than a certain value based on the reliability of each website . The present invention is characterized by further comprising an extraction means for calculating a ranking of keywords used for data acquired from the website, and extracting promising keywords based on a result of a change in the calculated ranking.

The evaluation support system according to a seventh invention is the evaluation support system according to the first invention, further comprising an output means for displaying the evaluation result evaluated by the evaluation means, and the output means is a first display for displaying the evaluation result for the target company. and a second display area that aggregates and displays evaluation results regarding other target companies related to the target company, the first display area and the second display area. each displaying a target company name , transaction information indicating the type of transaction between the target company and the other target company , and evaluation results based on one or more of the evaluation models.

The evaluation support program according to the eighth invention is an evaluation support program for supporting the value evaluation of a target company that is a target of financing or collaboration among companies using financial institutions, and is related to the value evaluation. a setting step of setting identification information of an acquisition source for acquiring information and circulation information indicating an acquisition order in which data is acquired from a plurality of websites of the acquisition destination; information of the target company; an acquisition step of acquiring target data including value information for determining value based on the identification information of the acquisition source and a plurality of the circulation information; and storing the target data acquired in the acquisition step in a value information database. a first storage step of storing a plurality of evaluation models for evaluating the value of the target company in a model definition database; and a second storage step of storing a plurality of evaluation models for evaluating the value of the target company in the model definition database with reference to the value information database. and an evaluation step of evaluating the value of the target company based on the evaluated evaluation model.

According to the first to seventh inventions, the setting means includes identification information of a source from which information about the target company necessary for value evaluation is acquired and a cycle indicating the acquisition order in which data is acquired from the websites of the plurality of acquisition sources. Set information. Therefore, data including information on the target company and value information for determining the value of the target company can be acquired based on the identification information of the acquisition source and a plurality of pieces of circulating information. This makes it possible to reduce the amount of work time spent by the user.

In particular, according to the second invention, the tour information including at least the identification information of the acquisition source and the order of acquisition of the plurality of acquisition sources includes time information regarding the website visit, number of times of execution, acquisition position regarding the data acquisition location, acquisition Contains at least one of a paragraph, an acquired keyword, and an excluded keyword. Therefore, data on the subject whose value is to be evaluated can be acquired at any time, such as immediately or at a scheduled time. This makes it possible to reduce the amount of work time spent by the user.

In particular, according to the third invention, the data includes public information published by the target company or a third party other than the target company , press information, word-of-mouth information, website trust information, and DX conversion regarding digitalization efforts. Contains at least one of the following information. Therefore, it is possible to obtain a wide range of data on the target company whose value is to be evaluated. This makes it possible to easily perform multifaceted evaluations and reduce the amount of work time spent by the user.

In particular, according to the fourth invention, the evaluation model includes at least one of an attention evaluation model, an evaluation analysis model, and a comprehensive score model. Therefore, it is possible to evaluate the value of the subject using an appropriate evaluation model. This makes it possible to easily perform multifaceted evaluations and reduce the amount of work time spent by the user.

In particular, according to the fifth invention, when the updating means newly acquires the relationship between the past value evaluation result and the reference information, the updating means reflects the relationship in the association. Therefore, the database can be updated easily. This makes it possible to reduce the amount of work time spent by the user.

In particular, according to the sixth invention, the evaluation means further includes extraction means. Therefore, when performing evaluation using the attention evaluation model, promising keywords can be extracted from the acquired data. As a result, it is possible to easily confirm and evaluate that the event is not a phenomenon that is currently widely recognized, and it becomes possible to reduce the amount of work time spent by the user.

In particular, according to the seventh invention, the output means includes two or more display areas including at least a first display area and a second display area. Therefore, the first display area and the second display area each display the target company name , transaction information indicating the type of transaction between the target company name and other target companies , and evaluation results based on one or more evaluation models. can be displayed. This makes it easy to compare evaluation results and reduces the amount of work time spent by the user.

According to the eighth invention, the setting step includes setting identification information of a source from which information regarding the target company required for value evaluation is acquired and circulation information indicating the order of acquisition in which data is acquired from the websites of the plurality of acquisition sources. . Therefore, data including information on the target company and value information for determining the value of the target company can be acquired based on the identification information of the acquisition source and a plurality of pieces of circulating information. This makes it possible to reduce the amount of work time spent by the user.

FIG. 1 is a schematic diagram showing an example of the configuration of an evaluation support system in this embodiment. FIG. 2 is a schematic diagram showing an example of the operation of the evaluation support system in this embodiment. FIG. 3(a) is a schematic diagram showing an example of the configuration of the evaluation support system in this embodiment, and FIG. 3(b) is a schematic diagram showing an example of the functions of the evaluation support system in this embodiment. FIG. 4 is a schematic diagram showing an example of a value information database, a model definition database, a circulation information database, and a used keyword ranking database. FIG. 5 is a flowchart showing an example of the operation of the evaluation support system in this embodiment. FIG. 6 is a flowchart showing an example of the operation of the evaluation support system in this embodiment. FIG. 7 is a schematic diagram showing an example of the evaluation results of the evaluation support system in this embodiment. FIG. 8 is a schematic diagram showing an example of an evaluation result display screen of the evaluation support system according to the present embodiment.

An example of an evaluation support system according to an embodiment of the present invention will be described below with reference to the drawings.

(Evaluation support system 100)
An example of the configuration of the evaluation support system 100 in this embodiment will be described with reference to FIGS. 1 to 8. FIG. 1 is a schematic diagram showing an example of the configuration of an evaluation support system 100 in this embodiment.

The evaluation support system 100 is mainly used to support the value evaluation of a target person who is a loan recipient at a financial institution (for example, a bank). The evaluation support system 100 can be used in combination with other technologies as a means for streamlining the evaluation of the value of a target person, such as credit risk management operations, sales, credit management, etc., of a target person with whom a bank or the like conducts a loan, capital alliance, etc. You may.

In particular, the evaluation support system 100 can acquire various data necessary for evaluating the value of a subject by visiting websites, and select an optimal evaluation model for the acquired data, such as the above-mentioned evaluation model. Able to carry out risk management work, etc. As a result, users can spend a huge amount of time and manpower to collect various information about target companies and other targets (e.g., competitors, related companies, etc.), as well as DX information regarding DX initiatives. There is no need to perform work such as confirmation and value evaluation to create new added value, and it is possible to reduce the user's work time.

For example, as shown in FIG. 1, the evaluation support system 100 includes an evaluation support device 1, and a user may use the evaluation support device 1. The evaluation support system 100 includes a terminal 2 connected to the evaluation support apparatus 1 via a communication network 4, for example, and a user may use the evaluation support apparatus 1 via the terminal 2. The evaluation support system 100 includes a server 3 connected to the evaluation support device 1 via, for example, a communication network 4, and allows the user to send and receive various information to and from the server 3 via the evaluation support device 1 or the terminal 2, thereby controlling each means. may be realized.

The evaluation support system 100 can implement processing as shown in FIG. 2, for example. FIG. 2 shows an example in which the user operates the terminal 2 to use the evaluation support device 1. For example, the user may directly operate the evaluation support device 1.

The evaluation support system 100 acquires various information (data) regarding the subject necessary for value evaluation by, for example, scraping a plurality of websites provided from the cloud 5. In addition to the target person's website, the website may be, for example, a news site that posts target person information, an SNS, a word-of-mouth site, a blog of an expert or an analyst, an information vendor, or other information site. . The target company is the target company whose value is to be evaluated, such as affiliated companies, subcontractors, collaborative companies, or competing companies of the target company, as well as government offices, survey/research institutions, educational institutions, etc. It may be an organization, a group, or a specific individual, and the scale, business form, industry, etc. of the company are arbitrary.

The evaluation support system 100 patrols each website based on patrol information that includes at least the identification information of the acquisition source set in advance by the user and the acquisition order of the plurality of acquisition sources. The tour information that includes at least the identification information of the acquisition source and the order of acquisition of the plurality of acquisition sources includes, for example, URL information of the website, as well as time information regarding the website visit, the number of executions, and a display screen displayed on the website. These include the acquisition position, acquisition paragraph, acquisition keyword, exclusion keyword, etc. regarding the data acquisition location in .

The evaluation support system 100 acquires various types of data based on circulation information that includes at least the identification information of these acquisition sources and the acquisition order of the plurality of acquisition sources. The evaluation support system 100 may acquire the information using, for example, a known website crawling technique (for example, scraping) when visiting websites. The data acquired by the evaluation support system 100 includes various target data such as the result heading (headline), title, and wording in the body of the evaluation, and includes, for example, the target company and the data excluding the target company. In addition to various information such as public information published by three parties, press information, and word-of-mouth information, for example, website trust information, DX information regarding the subject's digitalization efforts, audio data, video data, image data, It may be an icon, image data, etc.

The evaluation support system 100 stores various acquired data in a value information database. The value information database includes, for example, target person information about the target person (e.g., company name, company profile, business content, affiliated companies, partners, news releases, financial results information, etc.) and information that determines the target person's value as a target person. Data including various value information (for example, management information, transaction information, performance information, credit information, etc.) for the purpose of When storing the acquired data in the value information database, the evaluation support system 100 stores, for example, the website link information (URL) of the acquisition source of each data, the acquired date and time information, as well as the credibility and trustworthiness of the acquisition source website. The information may be acquired together and stored in association with various data.

The evaluation support system 100 assigns classification information to the data stored in the value information database. The classification information may be set in advance by the user, for example, or may be added as appropriate depending on, for example, the keyword or type of the acquired data, or the frequency of appearance of the keyword. Note that a plurality of classification information may be assigned to one piece of acquired data, for example.

The evaluation support system 100 evaluates the value of a subject by evaluating data based on an evaluation model. The evaluation model is an evaluation model for evaluating the value of a subject, and a plurality of evaluation models are stored in a model definition database. The model definition database stores, for example, an "attention evaluation model", an "evaluation analysis model", and a "total score model" which will be described later.

<Data>
The data indicates, for example, the target company's target company information and the content of value information for determining the target company's value, and specifically, the target company or a third party other than the target company. Shows public information published by people, press information, word-of-mouth information, website trust information, DX information regarding digitalization efforts, etc. The data may be, for example, data regarding each target person (in addition to companies, for example, government offices, survey/research institutions, educational institutions, organizations, groups, or specific individuals). For example, when the target data is target data, the target data may be character data acquired by visiting websites, data newly added or registered by the user, various types of information, etc., for example.

The data may be divided into keywords, words, or morphemes, and may be extracted from the target data using known language recognition technology. The data may be, for example, audio data, video data, image data, icons, image data, etc. in addition to target data.

<Database>
The evaluation support system 100 includes a database and data table as shown in FIG. 4, for example. The database includes at least a value information database and a model definition database, and may further include, for example, a circulation information database for setting circulation information for acquiring information regarding value evaluation, a used keyword ranking database, and the like. In addition to being stored in the evaluation support system 100, each database may be stored in a storage part different from the evaluation support apparatus 1, such as the server 3, and the storage location and storage timing for each database are arbitrary.

<<Value information database>>
The value information database stores target data including, for example, target person information of a target person and value information for determining target person value. The value information database includes, for example, a "target person ID" that identifies a target person, a "company name" that indicates the name of a target person, a "website" that indicates URL information of a website to be visited, and information on websites and targets. "Trust information" that indicates the reliability of data, "Target information" that shows various information about the target person, "Value information" that shows various information to judge the target person's value, The information is stored as "public information" indicating information published by a third party, etc. other than the company of the target person, and "word of mouth" indicating posted information such as word of mouth regarding the target person. Note that the value information database may store acquired target data together with date and time information.

<<Model definition database>>
The model definition database stores, for example, an "attention evaluation model," an "evaluation analysis model," and a "total score model" as a plurality of evaluation models for evaluating the value of a subject. The model definition database includes, for example, a "model ID" that identifies an evaluation model, a "model name" that indicates the name of the evaluation model, an "evaluation formula/parameter" that indicates conditions related to the target data to be evaluated, and a task schedule for each evaluation model. , is stored as "evaluation date/time information/update information" indicating update information.

The "attention evaluation model" evaluates the value of the target person based on, for example, the characteristics of the acquired data or classification information including at least the type and the reliability that indicates the credibility of each website set in advance by the user. It's a model. The reliability varies depending on the type of website, and is set as a threshold value such as "publication (company, public) = 2", "reporting = 1.5", "word of mouth (post, write) = 1", etc. . The evaluation support system 100 calculates the number of individual target data for the target person and the reliability of the website, and creates an attention evaluation model as a quantitative ratio, average value, etc. for the target data for each target person, for example. Output the results of the evaluation.

An "evaluation analysis model" is a learning model that has a relationship between the value evaluation results of past value evaluations (e.g., target data, etc.) and reference information (e.g., positive/neutral/negative) linked to the past value evaluation results. This is a model in which data is created in advance and evaluation is performed using a learning model constructed by machine learning using the created training data. Machine learning may be, for example, "supervised learning" or "semi-supervised learning".

This makes it possible for the "evaluation analysis model" to score, for example, the ratio of positive/neutral/negative to newly acquired target data for each subject. The evaluation support system 100 outputs the results of the evaluation using the evaluation analysis model as a qualitative evaluation of the target data for each subject, for example, using the learning model.

The "Comprehensive score model" is based on the evaluation results of the "Attention evaluation model" and the "Evaluation analysis model". is defined as "reputation", and the evaluation is performed by multiplying them ("attention level" x "reputation"). "Reputation" may be set as a threshold such as "(Positive ratio x 1 + Neutral ratio x 0.1 + Negative ratio x (-2)"). The evaluation support system 100 sets a threshold for reputation. This makes it possible to evaluate all target data.The evaluation support system 100 performs a comprehensive evaluation for each target person based on the evaluation results of the "attention evaluation model" and the "evaluation analysis model." As an evaluation, the results of the evaluation using the comprehensive score model are output.

The evaluation support system 100 refers to the model definition database, selects, for example, all or any of the "attention evaluation model," "evaluation analysis model," and "overall score model," and evaluates the value of the target of interest. Output evaluation results to support. The evaluation result is displayed, for example, on the display unit 109 of the evaluation support device 1 as an evaluation result 109a (for example, "value evaluation result of Corporation A").

＜＜Travel information database＞＞
The tour information database stores various kinds of tour information related to, for example, a tour (scraping) for acquiring information regarding value evaluation. The patrol information database includes, for example, a "visit destination ID" that identifies the conditions of the patrol, a "website" that indicates URL information of the website to be patrolled, a "subject ID" that identifies the target of the patrol destination, and a "target person ID" that identifies the target of the patrol. "Visit time" that indicates the time to start the patrol, "Execution information" that indicates the execution of the patrol, "Location information" that indicates the location of the target data to be obtained from the website, "Body/Paragraph" that indicates the structure of the text to be obtained, It is also stored as a "used KW (keyword)" or the like indicating a title or characteristic keyword in the target data. Note that the tour information database may also store, for example, information on the date and time when the website tour started, information on the date and time when it ended, and the like.

The evaluation support system 100 refers to a patrol information database, for example, based on a task scheduler (not shown) in which execution information is registered as a daily process, and if a task to be patrolled is registered, a match is found. A task schedule may also be started.

<<Used keyword ranking database>>
The used keyword (used KW) ranking database is created by the evaluation support system 100, for example, based on the calculated results of ranking the appearance frequency of the used keywords for the acquired target data, and the results of ranking changes, for example, in a certain period of time. Data on which keywords that have increased rapidly are extracted as promising keywords is stored. As shown in Figure 4, the used keyword ranking database includes, for example, a "keyword ID" that identifies extracted keywords, and a "used keyword" that indicates used keywords obtained from the target data of the title and text posted on the website. "Period 1/Previous", "Period 2/This" which indicates the ranking (rank) for the period, "Website" which indicates the website from which the used keyword was obtained, "Visiting destination ID" which indicates the visited destination ID, extracted It is stored as a "usage rate" indicating the rate at which a used keyword is used, a "promising candidate" indicating the presence or absence of a promising keyword candidate, etc.

<Evaluation support device 1>
FIG. 3(a) is a schematic diagram showing an example of the configuration of the evaluation support device 1. As the evaluation support device 1, a known electronic device such as a personal computer (PC), a smartphone, or a tablet terminal is used. The evaluation support device 1 includes, for example, a housing 10, a CPU (Central Processing Unit) 101, a ROM (Read Only Memory) 102, a RAM (Random Access Memory) 103, a storage unit 104, and I/Fs 105 to 107. , an input section 108, and a display section 109. Each configuration 101-107 is connected by an internal bus 110.

The CPU 101 controls the entire evaluation support device 1 . The ROM 102 stores operation codes for the CPU 101. The RAM 103 is a work area used when the CPU 101 operates. The storage unit 104 stores various information such as the above-mentioned databases. As the storage unit 104, for example, an HDD (Hard Disk Drive), an SSD (Solid State Drive), or the like is used.

The I/F 105 is an interface for transmitting and receiving various information to and from the terminal 2, server 3, communication network 4, and the like. I/F 106 is an interface for transmitting and receiving various information to and from input unit 108 . I/F 107 is an interface for transmitting and receiving various information to and from display unit 109 .

In addition to using a keyboard as the input unit 108, a sound pickup device such as a microphone may also be used. A user using the evaluation support device 1 can input various information such as target data and audio via the input unit 108, for example.

The display unit 109 displays various information such as display data stored in the storage unit 104, the processing status of the evaluation support device 1, and the like. In addition to using a display as the display unit 109, for example, a speaker may be used.

Note that, for example, the same I/F 105 to I/F 107 may be used, or a plurality of I/Fs 105 to I/F 107 may be used, respectively. Further, when a touch panel display is used as the display unit 109, the display unit 109 may include the input unit 108.

FIG. 3(b) is a schematic diagram showing an example of the functions of the evaluation support device 1. The evaluation support device 1 includes a setting section 11, an acquisition section 12, an evaluation section 13, an updating section 14, and an extraction section 15, and may also include, for example, an output section 16 and a storage section 17. Note that each function shown in FIG. 3(b) is realized by the CPU 101 executing a program stored in the storage unit 104 or the like using the RAM 103 as a work area. Also, a portion of each function may be controlled by artificial intelligence. Here, "artificial intelligence" may be based on any known artificial intelligence technology.

<Setting section 11>
The setting unit 11 sets circulation information for acquiring information regarding the title and text for which value evaluation is to be performed. The setting unit 11 obtains, for example, via the terminal 2, various kinds of tour information regarding the subject whose value evaluation is to be performed, input in advance from the user, and stores the obtained various kinds of tour information in the tour information database. The setting unit 11 can, for example, set various kinds of tour information to acquire target data via the terminal 2, and can also set a task schedule for selecting any one of the tour information and acquiring the tour information, for example. You can also do this. The setting unit 11 may set, for example, the timing of visiting websites, as well as the positions, paragraphs, and keywords of target data of various titles and texts displayed on the website.

<Acquisition unit 12>
Based on the circulation information set by the setting unit 11, the acquisition unit 12 acquires target data (title, text, etc.) including target person information of the target person for value evaluation and value information for determining the target person value. get. The acquisition unit 12 refers to, for example, the tour time and execution information set as each tour destination set in the tour information database. If the acquisition unit 12 has the matching date and time visit information, it sequentially transmits requests regarding the acquisition of content data to the URLs indicated as websites. The acquisition unit 12 receives a reply from a website, and registers various content data to be evaluated in a value information database according to the result of the received reply.

The acquisition unit 12 may classify, for example, target person information and value information included in the target data. In this case, classification can be performed using, for example, a known technique of natural language analysis (morphological analysis). Similarly to the above, the acquisition unit 12 may add characteristic information to the acquired content data, if any, in addition to the target person information and value information included in the target data, for example. Further, the acquisition unit 12 may register the acquired target data by sorting it into category information for evaluating its value. The classification information may be, for example, a company name of the acquired content data, a title given to the acquired content data, or a category indicating the characteristics or type of the content data.

The acquisition unit 12 acquires public information published by the target person or a third party other than the target person, press information, word of mouth information, website trust information, DX information regarding digitalization efforts, etc. , the data may be stored in the value information database in association with the target person ID identified by the website from which it was obtained.

<Evaluation section 13>
The evaluation unit 13 evaluates the value of the target person by evaluating the target data described in the value information database based on the evaluation model stored in the model definition database. The evaluation unit 13 refers to the model definition database and evaluates the attention evaluation model based on classification information including at least the characteristics or type of the acquired data and reliability indicating the credibility of each website set in advance by the user. , an evaluation analysis model that performs evaluation using a learning model constructed by machine learning using learning data that has a relationship between value evaluation results of past value evaluations and reference information linked to past value evaluation results, or An evaluation is performed that includes at least one of the comprehensive score models that perform evaluation based on the comprehensive score calculated using the attention evaluation model and the evaluation analysis model. The evaluation unit 13 may evaluate these evaluation models individually or in combination.

The evaluation unit 13 includes a provisional evaluation unit that evaluates the acquired title based on the title acquired by the acquisition unit 12 as a preliminary evaluation before the main evaluation.

<<Temporary Evaluation Department>>
The provisional evaluation unit may evaluate negative/positive value as a value evaluation for each piece of information (for each piece of news) based on the circulating information, and may provide an evaluation result (judgment result/result value). The provisional evaluation unit performs negative/positive evaluation using a "result headline" (title) based on the circulation information, for example. The provisional evaluation unit may be executed, for example, after the cyclic processing is completed. The temporary evaluation unit may start the job in a time period set by the task scheduler after the completion of the patrol.

The provisional evaluation section refers to the circulation information database and evaluates various content texts based on the circulation information acquired by the acquisition section 12. The tentative evaluation unit evaluates, for example, scraping execution date and time information, target site type information, result headline information, and a deletion flag that is a flag indicating whether logical deletion has occurred. Furthermore, by referring to the model definition database, for example, the evaluation model name indicating the model to be evaluated, the evaluation formula used to calculate the results for each information, the evaluation weight indicating the parameters, the evaluation start date and time, the evaluation end date and time, and the evaluation indicating update information. Perform preliminary evaluation tasks based on date and time information.

The provisional evaluation unit further sets the "end date and time" of the previous provisional evaluation execution in the "start date and time" of the "evaluation execution conditions", sets the "end date and time" to unset (NULL), and further sets the "end date and time" of the previous provisional evaluation execution to the "start date and time" of the "evaluation execution conditions", sets the "end date and time" to unset (NULL), and further sets the "end date and time" of the previous provisional evaluation execution to the "start date and time" of the "evaluation execution condition", sets the "end date and time" to not set (NULL), and further sets the "end date and time" of the previous provisional evaluation execution to the "end date and time" of the "evaluation execution condition". Results whose dates and times are between the "start date and time" and "end date and time" of the "evaluation execution conditions" (if not set, the time of processing execution) may be evaluated. For example, the temporary evaluation unit may analyze only the search results whose target site type is "target site type=3 (news)". Further, the provisional evaluation unit may exclude from the evaluation target cyclic results that have been logically deleted by a "deletion flag."

The provisional evaluation section refers to the model definition database and uses the "result headline" (title) based on the patrol results using various defined models, for example, to determine the evaluation judgment results for each model.・Calculate the result value (positive, neutral, negative). Calculations by the provisional evaluation unit may be performed by aggregating positive, neutral, and negative results using, for example, the sum of weights for each model identified by model ID and model name, weighted average, etc., and calculating evaluation results for each information. good. The provisional evaluation unit stores, for example, the evaluation results of the provisional evaluation for each company in the value evaluation database in association with the target person ID of each company.

Furthermore, the evaluation section 13 may instruct the acquisition section 12 to acquire the corresponding text based on the result of the preliminary evaluation evaluated by the preliminary evaluation section. The acquisition unit 12 acquires the text to be subjected to the main evaluation based on the text acquisition instruction from the evaluation unit 13 .

The evaluation unit 13 refers to the main text acquired by the acquisition unit 12 and performs the main evaluation of value evaluation (main evaluation unit).

<<This evaluation department>>
This evaluation section performs the same processing as the preliminary evaluation section, but whereas the preliminary evaluation section uses the "result headline" (title) of the patrol results, the main evaluation section uses the "result headline" (title) of the patrol results. The difference is that the evaluation is performed using the main text. For example, after the acquisition unit 12 obtains the text, the evaluation unit uses the text to perform evaluation. The main evaluation unit stores, for example, the evaluation results of the main evaluation for each company in the value evaluation database in association with the target person ID of each company.

The evaluation unit 13 evaluates the value of the target person based on the evaluation model input by the user via the terminal 2, for example, and also evaluates the value of the target person based on information such as the amount and characteristics of the aggregated target data. An evaluation model may be appropriately determined and the value of each subject may be evaluated individually (individual evaluation section).

<<Individual Evaluation Department>>
The individual evaluation unit refers to, for example, a value evaluation database, aggregates the results of the main evaluation, and calculates results for each subject. The individual evaluation section refers to a database that stores the results of the main evaluation by the main evaluation section for each company, for example, and totalizes the positive, neutral, and negative information for each piece of information from the results of the main evaluation by the main evaluation section. The individual evaluation unit calculates positive, neutral, and negative evaluations as evaluation determination results for each company, for example, based on the aggregated results. For example, the individual evaluation unit adds the reliability of each site to the attention score of the main text for each news, and calculates the individual evaluation for each news. For example, the individual evaluation unit may calculate the rank of attention for each news group nationwide, in the same industry, and in the same scale. Furthermore, the individual evaluation unit may calculate, for example, the overall score ranking for each country, the same industry, and the same size based on the result of the attention ranking.

The individual evaluation section calculates the attention score, evaluates the reputation ratio (positive, neutral, negative), and evaluates the overall score. The attention score is calculated, for example, as the sum of reliability by site (total of analysis results by information linked to the target person). For example, the reputation ratio (positive, neutral, negative) is determined on a company-by-company basis. At this time, the reputation ratio may be a positive, neutral, or negative ratio, for example, depending on the evaluation result for each information associated with the target person.

Figure 7 shows the calculation of the attention score by the individual evaluation section, the evaluation of the reputation ratio (positive, neutral, negative), and the evaluation of the overall score. FIG. 7 shows, for example, the evaluation results of the current, previous, and difference evaluations by the individual evaluation unit, and the evaluation results of the degree of attention and evaluation ratio for each subject are calculated. The overall score is determined by adding, for example, "attention level x 100 for positive", "attention level x 10 for neutral", "attention level x -200 for negative", etc. for each information analysis result, and - A cap floor of 1000 to 500 may be provided. As a result, the attention ranking and overall score ranking are determined based on, for example, the attention ranking (nationwide, same industry, same size) and overall score ranking (nationwide, same industry, same size). The individual evaluation unit stores, for example, the evaluation results of the main evaluation for each company in the value evaluation database in association with the target person ID of each company.

The individual evaluation unit refers to the model definition database and calculates a score for each model. The score for each model is referenced in the evaluation of the "result heading (headline)" (title) in the preliminary evaluation department, for example, and in the main evaluation department, the main text is used to judge the evaluation result, for example, in the model definition database. With reference to each definition model, the "positive degree threshold" and "negative degree threshold" are obtained based on the evaluation formula and parameters. The individual evaluation unit calculates a score for each model based on, for example, a score for each model (a degree of positivity, a degree of neutrality, and a degree of negativity).

For example, if "(positive degree - negative degree)>positive degree threshold", the individual evaluation unit determines that it is "positive". Further, if "(degree of negativity - degree of positivity)> threshold of degree of negativity", it is determined to be "negative". Furthermore, in cases other than these, it may be determined as "neutral". The individual evaluation unit calculates scores for each model, using, for example, known scoring methods (provided by third-party organizations, unsupervised learning models, supervised learning models, etc.), general polarity dictionaries, etc. A ring service may be used, and any scoring method may be used. The individual evaluation unit may store, for example, model/information-specific evaluation results for each company in the value evaluation database in association with the target person ID of each company.

<Update section 14>
The updating unit 14 updates the evaluation model stored in the model definition database, for example. For example, when the update unit 14 newly acquires a relationship between past value evaluation results and reference information, the update unit 14 reflects the relationship in the association. For example, when an evaluator or the like determines the accuracy of the content of the evaluation result based on the evaluation result generated by the evaluation section 13, and the evaluation support device 1 acquires the determination result, the updater 14 updates the evaluation result. Update the association of evaluation models included in the model definition database based on the results.

<Extraction part 15>
The extraction unit 15 extracts keywords that will be promising in the future, for example, when evaluating the value of a target person. For example, when performing an evaluation using an attention evaluation model, the extraction unit 15 calculates the ranking of used keywords based on the reliability of each website for target data obtained from websites whose reliability is above a certain value. do. The extraction unit 15 extracts promising keywords based on the results of the calculated ranking changes.

For example, the extraction unit 15 determines highly reliable websites based on trust information registered in the value information database, and uses target data obtained from websites whose trust information is registered as "high". Extract keywords. The extraction unit 15 extracts used keywords as promising keywords based on, for example, the ranking of the extracted used keywords for each period and the usage rate of each website in the period. Thereby, for example, promising keywords can be efficiently discovered from the acquired target data, which is useful for evaluating the value of the target person.

<Output section 16>
The output unit 16 outputs, for example, the result of the value evaluation of the subject evaluated by the evaluation unit 13. The output unit 16 reports various information via the display unit 109, the terminal 2, etc. so that the user can recognize it. The output unit 16 outputs various information to the terminal 2 etc. via the I/F 105, and outputs the evaluation result 109a to the display unit 109 via the I/F 107.

FIG. 8 shows the value evaluation results output by the output unit 16. The output unit 16 outputs the evaluation result of the subject's value evaluation to the first display area 109b with the content shown in the evaluation result 109a. The output unit 16 includes two or more display areas including at least a second display area 109c that aggregates and displays the evaluation results of other subjects together with the evaluation results of the subject. The output unit 16 displays, in each of the first display area 109b and the second display area 109c, the target person's name and transaction information indicating the type of transaction between the target person and another target person, Evaluation results based on one or more evaluation models are displayed.

<Storage unit 17>
The storage unit 17 retrieves various data such as a database stored in the storage unit 104 as necessary. The storage unit 17 stores various data acquired or generated by each of the components 11 to 17 in the storage unit 104 as necessary. The storage unit 17 includes a first storage unit that stores the target data acquired by the acquisition unit 12 in a value information database, and stores a plurality of evaluation models for evaluating the value of the target person in a model definition database. A second storage unit is included.

<Evaluation support program>
The evaluation support program is an evaluation support program in the evaluation support system 100 (evaluation support device 1) that supports evaluation of the value of the subject. The evaluation support program includes a setting step of setting, for each target person or identification information, circulation information including at least identification information of an acquisition source for acquiring information related to value evaluation, and the order of acquisition of a plurality of acquisition sources, for example; an acquisition step of acquiring target data including target person information of the target person and value information for determining target person value based on circulation information including at least identification information of the acquisition source and acquisition order of the plurality of acquisition sources; has.

Furthermore, the evaluation support program includes a first storage step of storing target data in a value information database, a second storage step of storing a plurality of evaluation models for evaluating the value of the target person in a model definition database, and an evaluation step of referring to the value information database and evaluating the value of the subject based on the evaluation model stored in the model definition database, and is executed by, for example, the evaluation support system 100 (evaluation support device 1).

<Terminal 2>
As the terminal 2, a known electronic device such as a personal computer, a smartphone, a tablet terminal, etc. is used. The terminal 2 may have, for example, at least part of the same configuration and functions as the evaluation support device 1 described above. For example, a plurality of terminals 2 may be provided, and each terminal 2 may be connected to the evaluation support device 1 via the communication network 4.

<Server 3>
The server 3 stores various information such as the above-mentioned database. The server 3 stores various information sent from the evaluation support device 1 and the like via the communication network 4, for example. The server 3 may store, for example, the same information as the storage unit 104, and may exchange various information with the evaluation support device 1 and the like via the communication network 4. That is, in the evaluation support system 100, the server 3 may be used instead of the evaluation support device 1 or the storage unit 104 of the evaluation support device 1.

<Communication network 4>
The communication network 4 is an Internet network or the like to which the evaluation support device 1 is connected via a communication circuit. The communication network 4 may be constituted by a so-called optical fiber communication network. Further, the communication network 4 may be realized by a known communication network such as a wireless communication network in addition to a wired communication network.

(Example of operation of evaluation support system 100)

Next, an example of the operation of the evaluation support system 100 in this embodiment will be described. 5 and 6 are flowcharts showing an example of the operation of the evaluation support system 100 in this embodiment.

The evaluation support system 100 includes, for example, as shown in FIGS. 5 and 6, a setting means S110, an acquisition means S120, an evaluation means S130, and an output means S140, and includes, for example, at least an extraction means S150, an updating means, and a storage means. Either one may be provided. Furthermore, the evaluation means S130 may include at least any one of the provisional evaluation means S210, the main evaluation means S220, and the individual evaluation means S230.

<Setting means S110>
Setting means S110 sets tour information. For example, the setting unit 11 sets tour information for acquiring information regarding value evaluation in the tour information database based on text input by the user via the terminal 2 or the input unit 108. The setting unit 11 may set, for example, the destination ID, website, person ID, visit time, execution information, position, paragraph, KW, etc. for each target person as the information to be set.

For example, the setting unit 11 sets the patrol information such that the information regarding the target person is at least one of the following: time information regarding the visit of the website, execution information, acquisition position regarding the acquisition location of the target data, acquisition paragraph, acquisition keyword, and exclusion keyword. Set as information.

<Obtaining means S120>
The acquisition means S120 acquires target data based on the circulation information. The acquisition unit 12 acquires target data including target person information of the target person and value information for determining target person value based on the circulation information, for example, from a title and a body text. Further, the acquisition unit 12 acquires target data from the main text, for example, based on the result of the preliminary evaluation by the preliminary evaluation means. For example, the acquisition unit 12 stores the acquired target data in the storage unit 104. The acquisition unit 12 may acquire target data, for example, at each timing of a user's input, or may acquire, for example, a plurality of target data accumulated over a certain period of time at once. Note that a known technique can be used as a method for generating target data from non-text data input via the terminal 2 or the input unit 108.

The target data acquired by the acquisition unit 12 includes, for example, public information published by the target person or a third party other than the target person, press information, word-of-mouth information, website trust information, and DX related to digitalization initiatives. Contains at least one of the conversion information.

<Evaluation means S130>
The evaluation means S130 evaluates the value of the subject. The evaluation unit 13 evaluates the value of the target person by evaluating the target data described in the value information database based on the evaluation model stored in the model definition database. The evaluation unit 13 uses an attention evaluation model that performs evaluation based on classification information that includes at least the characteristics or type of the acquired data and the reliability that indicates the credibility of each website set in advance by the user, and the value of past value evaluations. An evaluation analysis model that performs evaluation using a learning model constructed by machine learning using learning data that has a relationship between evaluation results and reference information linked to past value evaluation results, and an attention evaluation model and evaluation analysis. The value of the subject is evaluated by making an evaluation based on at least one of the overall score models that evaluate based on the overall score calculated using the model.

The evaluation means S130 may include at least one of the provisional evaluation means S210, the main evaluation means S220, and the individual evaluation means S230.

<<Temporary evaluation means S210>>
The provisional evaluation means S210 performs provisional evaluation based on each title of the search results. After the patrol processing is completed, the provisional evaluation section performs a positive, neutral, or negative evaluation for each piece of information (for each news item) based on the patrol results obtained by the acquisition section 12, and calculates the judgment result and result value. , an addition is given based on the weight for each evaluation model. The provisional evaluation unit makes a positive determination, a neutral determination, and a negative determination, for example, based on the total sum of positive, neutral, and negative (maximum votes, etc.).

The temporary evaluation unit further determines, for example, the target site type, and evaluates only the patrol results with "target site type = 3 (news)", and, for example, the patrol results that have been logically deleted by the "deletion flag". may be excluded from the evaluation. The provisional evaluation unit instructs the acquisition unit 12 to acquire the corresponding text based on the evaluation of the provisional evaluation. The acquisition unit 12 acquires the main text to be subjected to the main evaluation based on the instruction to obtain the main text from the temporary evaluation unit.

<<This evaluation means S220>>
The main evaluation means S220 performs the main evaluation based on the text. This evaluation department. After the acquisition unit 12 acquires the text, the text is used for evaluation. This evaluation section performs the same processing as the provisional evaluation section, and based on the patrol results obtained by the acquisition section 12, evaluates the main text as positive, neutral, or negative, for example, by information (by news). , calculate the judgment result and result value, and add weights for each evaluation model. The evaluation unit makes a positive determination, a neutral determination, and a negative determination, for example, based on the total sum of positive, neutral, and negative (maximum votes, etc.).

<<Individual evaluation means S230>>
The individual evaluation means S230 performs various individual evaluations. The individual evaluation unit refers to, for example, a value evaluation database and aggregates the results of the main evaluation. Based on the results of the main evaluation performed by the main evaluation section, the individual evaluation section adds, for example, the reliability of each site to the attention score of the main text for each news, and calculates the individual evaluation for each news. For example, the individual evaluation department aggregates the numbers based on positive, neutral, and negative values for each piece of information, and calculates positive, neutral, and negative evaluations as evaluation judgment results for each company.

For example, the individual evaluation unit may calculate the rank of attention for each news group nationwide, in the same industry, and in the same scale. Furthermore, the individual evaluation unit may calculate, for example, the overall score ranking for each country, the same industry, and the same size based on the result of the attention ranking.

<Output means S140>
The output means S140 outputs the result of the value evaluation of the subject as the evaluation result 109a. The output unit 16 includes, for example, a first display area 109b that displays evaluation results for the target person, and a second display area 109c that aggregates and displays evaluation results for other targets related to the target person. At least two or more display areas are provided. The output unit 16 displays, for example, a transaction message indicating the target person's name and the type of transaction between the target person and the other target person in each of the first display area 109b and the second display area 109c. Information and evaluation results based on one or more of the evaluation models are displayed. The output unit 16 may switch the display order and display the evaluation results based on an instruction to switch the display of the evaluation results inputted via the terminal 2, for example.

<Extraction means S150>
Extraction means S150 extracts promising keywords. For example, when performing an evaluation using an attention evaluation model, the extraction unit 15 calculates the ranking of used keywords based on the reliability of each website for target data obtained from websites whose reliability is above a certain value. do. The extraction unit 15 extracts promising keywords based on the results of the calculated ranking changes.

Thereby, the optimal process can be selected for the target data input by the user, and, for example, the operation of the evaluation support system 100 in this embodiment ends.

Further, according to the present embodiment, the evaluation support system 100 refers to the value information database and obtains, for example, target person information of the target person included in the target data, characteristics of acquired data for specifying value information, or Based on classification information including at least the type, it may be determined whether or not there is any overlap in the target data, and if there is any overlap in the target data, the duplicate target data may be deleted. The evaluation support system 100 stores the aggregated target data after deletion in the value information database. Thereby, it is possible to reduce the work time spent by the user in evaluating the value of the subject.

Further, according to the present embodiment, the evaluation support program for supporting the value evaluation of the target person includes a setting step S110 for setting tour information, and an acquisition step S120 for acquiring target data based on the tour information. step, a first storage step of storing target data in a value information database, a second storage step of storing a plurality of evaluation models in a model definition database, an evaluation step of evaluating the value of the target person S130, and an evaluation step of storing the evaluation results. The information can be provided by causing a computer to execute the evaluation step S140 for display.

Although several embodiments of the invention have been described, these embodiments are presented by way of example and are not intended to limit the scope of the invention. These novel embodiments can be implemented in various other forms, and various omissions, substitutions, and changes can be made without departing from the gist of the invention. These embodiments and their modifications are included within the scope and gist of the invention, as well as within the scope of the invention described in the claims and its equivalents.

1: Evaluation support device 2: Terminal 3: Server 4: Communication network 5: Cloud 10: Housing 11: Setting unit 12: Acquisition unit 13: Evaluation unit 14: Updating unit 15: Extracting unit 16: Output unit 17: Storage unit 100: Evaluation support system 101: CPU
102: ROM
103: RAM
104: Storage section 105: I/F
106: I/F
107: I/F
108: Input section 109: Display section 109a: Evaluation result 109b: First display area 109c: Second display area 110: Internal bus S110: Setting means S120: Acquisition means S130: Evaluation means S140: Output means S150: Extraction means S210: Temporary evaluation means S220: Main evaluation means S230: Individual evaluation means

Claims (8)

An evaluation support system for supporting the value evaluation of target companies for financing or collaboration among companies using financial institutions ,
a setting means for setting identification information of an acquisition source for acquiring information regarding the value evaluation, and circulation information indicating an acquisition order in which data is acquired from a plurality of websites of the acquisition destination;
acquisition means for acquiring target data including information on the target company and value information for determining the value of the target company based on the identification information of the acquisition source and the plurality of pieces of patrol information;
a value information database that stores the target data acquired by the acquisition means;
a model definition database storing a plurality of evaluation models for evaluating the value of the target company ;
an evaluation means that refers to the value information database and evaluates the value of the target company based on the evaluation model stored in the model definition database;
An evaluation support system comprising: The tour information set by the setting means is:
The evaluation support system according to claim 1, further comprising at least one of time information regarding a website visit, the number of times of execution, an acquisition position regarding an acquisition location of the target data, an acquisition paragraph, an acquisition keyword, and an exclusion keyword. The target data acquired by the acquisition means is:
It is characterized by containing at least one of public information published by the target company or a third party other than the target company , press information, word-of-mouth information, website trust information, and DX information regarding digitalization efforts. The evaluation support system according to claim 1, wherein: The evaluation model stored in the model definition database is:
An attention evaluation model that performs evaluation based on the classification information including at least the characteristics or type of the acquired data and the reliability that indicates the credibility of each website set in advance by the user, the value evaluation results of past value evaluations, and the past value evaluations. An evaluation analysis model that performs evaluation using a learning model constructed by machine learning using learning data that has a relationship with reference information linked to value evaluation results, and the attention evaluation model and the evaluation analysis model. The evaluation support system according to claim 1, further comprising at least one of a total score model that performs evaluation based on a total score calculated using a total score model. The evaluation according to claim 4, further comprising updating means for reflecting the relationship in the association when a relationship between the past value evaluation result and the reference information is newly acquired. support system. The evaluation means is
When performing an evaluation using the attention evaluation model, based on the reliability of each website, calculate the ranking of the keywords used for target data obtained from websites whose reliability is higher than a certain value, The evaluation support system according to claim 4, further comprising an extraction means for extracting promising keywords based on the calculated result of the ranking change. Further comprising an output means for displaying the evaluation result evaluated by the evaluation means,
The output means includes at least two or more display areas, including at least a first display area that displays evaluation results for the target company , and a second display area that aggregates and displays evaluation results for other target companies related to the target company . The first display area and the second display area each display a target company name , transaction information indicating a transaction type between the target company and the other target company , and one or more display areas. The evaluation support system according to claim 1 , further comprising displaying an evaluation result based on the evaluation model. An evaluation support program for supporting the value evaluation of target companies for financing or collaboration among companies using financial institutions ,
a setting step of setting identification information of an acquisition source for acquiring information regarding the value evaluation, and circulation information indicating an acquisition order in which data is acquired from a plurality of websites of the acquisition destination;
an acquisition step of acquiring target data including information on the target company and value information for determining the value of the target company based on the identification information of the acquisition source and the plurality of pieces of circulating information;
a first storage step of storing the target data acquired in the acquisition step in a value information database;
a second storage step of storing a plurality of evaluation models for evaluating the value of the target company in a model definition database;
an evaluation step of referring to the value information database and evaluating the value of the target company based on the evaluation model stored in the model definition database;
An evaluation support program characterized by causing a computer to execute.

Images