Recruitment support system, recruitment support method and program

Abstract

To provide a recruitment support system that can easily modify job postings to fit a format. [Solution] According to one aspect of the present invention, a hiring support system is provided. The hiring support system includes a processor. The processor is configured to execute the following steps: In the acquisition step, an initial job posting is acquired. In the job posting regeneration step, a plurality of job posting components extracted from the initial job posting are input to a classification model, and classification information for each job posting component is output from the classification model, and the plurality of job posting components are arranged in a job posting format having a different configuration from the initial job posting based on the classification information. Here, the job posting components include at least one word or sentence. The classification information is information indicating the item in which the job posting components are arranged in the job posting format. The classification model is a learning model that has been trained so as to be able to input job posting components and output classification information.

Description

[Technical field]

[0001] The present invention relates to a recruitment support system, a recruitment support method, and a program. [Background technology]

[0002] As disclosed in Patent Document 1, a technique is known that enables job recruiters to register job information and job seekers to search for registered job listings (job postings). [Prior art documents] [Patent documents]

[0003] [Patent Document 1] JP 2002-73868 A Summary of the Invention [Problem to be solved by the invention]

[0004] In the above-mentioned recruitment support services, the format of the job posting (for example, required items, the order of items, etc.) differs depending on the service. Therefore, employers need to input the job posting for each recruitment support service.

[0005] In view of the above circumstances, the present invention provides a recruitment support system etc. that can easily modify a job posting form to fit a format. [Means for solving the problem]

[0006] According to one aspect of the present invention, there is provided a recruitment support system. The recruitment support system includes a processor. The processor is configured to execute the following steps: In the acquisition step, an initial job posting is acquired. In the job posting regeneration step, a plurality of job posting components extracted from the initial job posting are input to a classification model, and classification information for each job posting component is output from the classification model, and the plurality of job posting components are arranged in a job posting format having a different configuration from the initial job posting based on the classification information. Here, the job posting components include at least one word or sentence. The classification information is information indicating the item in which the job posting components are arranged in the job posting format. The classification model is a learning model that has been trained so as to be able to input job posting components and output classification information.

[0007] According to this embodiment, the contents of the initial job posting can be arranged to fit the new job posting format, so that the employer can easily modify the job posting to fit the format of the job search service. [Brief description of the drawings]

[0008] [Figure 1] FIG. 1 is a configuration diagram showing a recruitment support system 1. [Diagram 2] 2 is a block diagram showing a hardware configuration of the server device 10. FIG. [Diagram 3] FIG. 2 is a block diagram showing the hardware configuration of a recruiter terminal 20 and a job seeker terminal 30. [Figure 4] FIG. 2 is a block diagram showing functions realized by a server device 10 (control unit 11), a recruiter terminal 20 (control unit 21), and a job seeker terminal 30 (control unit 31). [Diagram 5] FIG. 13 is a diagram showing an example of a job posting JP. [Figure 6] FIG. 13 is a diagram showing an example of a job posting reference website RS. [Figure 7] FIG. 2 is a diagram showing an example of an information acquisition screen AD displayed on the recruiter terminal 20 for acquiring a job advertisement or access information to a job advertisement reference website. [Figure 8] FIG. 13 is a diagram showing an example of a state in which a job advertisement link JL has been acquired from a job advertisement reference website RS entered in the access information input field AF on the information acquisition screen AD. [Figure 9] 13 is a diagram showing an example of a job posting list screen LD displayed on the recruiter terminal 20. FIG. [Figure 10] FIG. 13 is a diagram showing an example of a new job posting creation screen CD. [Figure 11] FIG. 2 is an activity diagram showing the flow of information processing (processing for converting an initial job posting into a new job posting) executed by recruitment support system 1. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0009] DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Hereinafter, embodiments of the present invention will be described with reference to the drawings. Various characteristic features shown in the following embodiments can be combined with each other.

[0010] Incidentally, the program for realizing the software appearing in this embodiment may be provided as a non-transitory computer-readable recording medium, or may be provided so as to be downloadable from an external server, or may be provided so that the program is started on an external computer and its functions are realized on a client terminal (so-called cloud computing).

[0011] In addition, in this embodiment, the term "unit" may include, for example, a combination of hardware resources implemented by a circuit in the broad sense and software information processing that can be specifically realized by these hardware resources. In addition, in this embodiment, various information is handled, and this information is represented, for example, by physical values ​​of signal values ​​representing voltage and current, high and low signal values ​​as a binary bit collection consisting of 0 or 1, or quantum superposition (so-called quantum bits), and communication and calculation can be performed on the circuit in the broad sense.

[0012] In addition, a circuit in the broad sense is a circuit realized by at least appropriately combining a circuit, circuitry, a processor, a memory, etc. In other words, it includes an application specific integrated circuit (ASIC), a programmable logic device (e.g., a simple programmable logic device (SPLD), a complex programmable logic device (CPLD), and a field programmable gate array (FPGA)), etc.

[0013] 1. Hardware Configuration This section describes the hardware configuration.

[0014] <Recruitment Support System 1> Fig. 1 is a configuration diagram showing a hiring support system 1. The hiring support system 1 comprises a communication line 2, a server device 10, a plurality of recruiter terminals 20, and a plurality of job seeker terminals 30. The server device 10, the recruiter terminals 20, and the job seeker terminals 30 are configured to be able to communicate with each other via the communication line 2. The connection between the server device 10, the recruiter terminals 20, and the job seeker terminals 30 may be wired or wireless.

[0015] The hiring support system 1 constitutes part of a recruitment and job search system used by multiple recruiters (a first recruiter U1 and a second recruiter U2) and multiple job seekers (a first job seeker U3 and a third job seeker U4). The hiring support system 1 is primarily responsible for managing job postings. In one embodiment, the hiring support system 1 is comprised of one or more devices or components. These components will be described below.

[0016] <Server device 10> 2 is a block diagram showing a hardware configuration of the server device 10. The server device 10 includes a control unit 11, a storage unit 12, a communication unit 13, and a communication bus 14. The control unit 11, the storage unit 12, and the communication unit 13 are electrically connected to each other inside the server device 10 via the communication bus 14.

[0017] <Control unit 11> The control unit 11 performs processing and control of the overall operation related to the server device 10. The control unit 11 is, for example, a central processing unit (CPU). The control unit 11 realizes various functions related to the server device 10 by reading out a predetermined program stored in the storage unit 12. That is, information processing by software stored in the storage unit 12 can be specifically realized by the control unit 11, which is an example of hardware, and executed as each functional unit included in the control unit 11. These will be described in more detail in the next section. Note that the control unit 11 is not limited to being single, and may be implemented with multiple control units 11 for each function. Also, a combination of these may be used.

[0018] <Storage section 12> The storage unit 12 stores various information defined by the above description. This can be implemented, for example, as a storage device such as a solid state drive (SSD) that stores various programs and the like related to the server device 10 executed by the control unit 11, or as a memory such as a random access memory (RAM) that stores temporarily required information (arguments, arrays, etc.) related to the program calculations. The storage unit 12 stores various programs, variables, etc. related to the server device 10 executed by the control unit 11.

[0019] <Communications Division 13> The communication unit 13 is preferably a wired communication means such as USB, IEEE1394, Thunderbolt (registered trademark), wired LAN network communication, etc., but may also include wireless LAN network communication, mobile communication such as 3G / LTE / 5G, BLUETOOTH (registered trademark) communication, etc. as necessary. In other words, it is more preferable to implement it as a collection of multiple communication means. In other words, the server device 10 may communicate various information from the outside via the communication unit 13 and the network.

[0020] The server device 10 may be in an on-premise form or in a cloud form. As the server device 10 in a cloud form, the above-mentioned functions and processes may be provided in the form of, for example, SaaS (Software as a Service) or cloud computing.

[0021] <Employer Terminal 20> Fig. 3 is a block diagram showing the hardware configuration of the recruiter terminal 20 and the job seeker terminal 30. As shown in Fig. 3A, the recruiter terminal 20 comprises a control unit 21, a memory unit 22, a communication unit 23, an input unit 24, an output unit 25, and a communication bus 26. The control unit 21, the memory unit 22, the communication unit 23, the input unit 24, and the output unit 25 are electrically connected inside the recruiter terminal 20 via the communication bus 26. The explanation of the control unit 21, the memory unit 22, and the communication unit 23 is omitted here as they are the same as the explanation of each unit in the server device 10. Note that the recruiter terminal 20 may be a terminal operated by a human resources agency that communicates with job seekers on behalf of the recruiter.

[0022] <Input section 24> The input unit 24 accepts operation inputs made by the user. The operation inputs are transferred as command signals to the control unit 21 via the communication bus 26. The control unit 21 may execute predetermined control or calculations based on the transferred command signals as necessary. The input unit 24 may be included in the housing of the recruiter terminal 20, or may be externally attached. For example, the input unit 24 may be implemented as a touch panel integrated with the output unit 25. When the input unit 24 is implemented as a touch panel, the user can input tap operations, swipe operations, and the like to the input unit 24. As the input unit 24, a switch button, a mouse, a track pad, a QWERTY keyboard, and the like can be adopted instead of a touch panel.

[0023] <Output section 25> The output unit 25 displays a screen of a graphical user interface (GUI) that can be operated by the user. The output unit 25 may be included in the housing of the recruiter terminal 20, or may be attached externally. Specifically, the output unit 25 may be implemented as a display device such as a CRT display, a liquid crystal display, an organic EL display, or a plasma display. It is preferable that these display devices are implemented by selectively using them according to the type of recruiter terminal 20.

[0024] <Job Seeker Terminal 30> 3B, the job seeker terminal 30 includes a control unit 31, a memory unit 32, a communication unit 33, an input unit 34, an output unit 35, and a communication bus 36. The control unit 31, the memory unit 32, the communication unit 33, the input unit 34, and the output unit 35 are electrically connected within the job seeker terminal 30 via the communication bus 36. Descriptions of the control unit 31, the memory unit 32, the communication unit 33, the input unit 34, and the output unit 35 will be omitted as they are the same as the descriptions of the respective units in the recruiter terminal 20.

[0025] 2. Functional configuration In this section, the functional configuration of the present embodiment will be described. Information processing by the software stored in the storage unit 12 is specifically realized by the control unit 11, which is an example of hardware, and can be executed as each functional unit included in the control unit 11 (a processor included in the recruitment support system 1).

[0026] FIG. 4 is a block diagram showing functions realized by the server device 10 (control unit 11), the recruiter terminal 20 (control unit 21), and the job seeker terminal 30 (control unit 31).

[0027] As shown in Fig. 4A, the server device 10 (control unit 11) includes a basic display control unit 111, an acquisition unit 112, a job posting evaluation unit 113, a list display control unit 114, a job posting regeneration unit 115, an item addition unit 116, a job posting display control unit 117, and an artificial intelligence unit 120. As shown in Fig. 4B, the recruiter terminal 20 (control unit 21) includes a display unit 211 and an operation acquisition unit 212.

[0028] <Basic display control unit 111> The basic display control unit 111 is configured to display various information on the recruiter terminal 20 and the job seeker terminal 30. For example, the basic display control unit 111 displays, on the display unit 211 of the recruiter terminal 20 or the display unit 311 of the job seeker terminal 30, resumes and curriculum vitae created by the job seeker, job advertisements and scouting documents created by the recruiter, etc.

[0029] Recruiters include organizations such as for-profit corporations (e.g., companies), non-profit corporations (e.g., cooperatives, foundations), and public corporations (e.g., local governments). Recruiters also include recruitment agencies that act as agents of organizations to mediate between job seekers and organizations. Recruiters are also called headhunters, agents, etc.

[0030] <Acquisition part 112> The acquiring unit 112 is configured to acquire, as data, an initial job posting that is to be converted into a new job posting (new job posting) that conforms to a specific job posting format.

[0031] FIG. 5 is a diagram showing an example of a job posting JP. In this job posting JP, job information is described for each of a number of item IMs, such as the name of the position being advertised, the job content and working conditions (annual salary, job type, industry, work location, working style, work environment, etc.), application qualifications (skills), the desired personality, and appealing points. In other words, the job posting JP is composed of a number of item IMs divided into information categories. An item IM may further be composed of a number of sub-item SIMs. In the example of FIG. 5, the item IM for "application qualifications" includes sub-item SIMs for "required requirements" and "desired requirements." In addition, the job posting JP may include items such as the job title, heading, and information on the employer (company size (sales, number of employees, etc.), industry, etc.) as contents.

[0032] The job posting database in which the job postings used and referenced by the recruiting support system 1 are registered is not limited to the database used by the recruitment and job search service configured by the recruiting support system 1, but may be a database in which job postings are registered in an external service, or may be a database of a recruitment and job search service other than the recruiting support system 1. In other words, the initial job posting may be a job posting imported from a service external to the recruiting support system 1, or may be a job posting published on the Internet.

[0033] The job posting format of a job posting JP differs depending on the system, service, or other medium that publishes the job posting JP (where the job posting JP is registered). Even within the same medium, there may be job posting JPs with different job posting formats depending on the time the job posting was registered, the type of service used, etc. Furthermore, the job posting format used by recruiting support system 1 is fundamentally different in structure from the job posting format used by media external to recruiting support system 1. Here, the job posting format defines the number, type, order of arrangement, display format, etc. of items that make up the job posting JP.

[0034] Note that the job posting JP in Fig. 5 is an example of a job posting that can be used by the recruiting support system 1 (i.e., created according to a job posting format in a service that the recruiting support system 1 configures). Therefore, the structure of an initial job posting obtained from outside the recruiting support system 1 does not necessarily match the structure of the job posting JP in Fig. 5.

[0035] The acquisition unit 112, for example, displays a list of available initial job listings on the employer terminal 20 and accepts an input of a selection of an initial job listing from the employer terminal 20. Furthermore, the acquisition unit 112 may automatically acquire information on all initial job listings registered in a job listing database designated by the user, or any job listing designated by the user, as an initial job listing.

[0036] Furthermore, the acquiring unit 112 may acquire an initial job listing from a website on which the job listing is published. The acquiring unit 112 may also acquire website access information for accessing a job listing reference website, including job listing access information for accessing the initial job listing, and present job listing access information that can be acquired from the website access information. This allows multiple initial job listings to be acquired based on a single piece of website access information input by a user. This facilitates the task of converting multiple initial job listings into new job listings.

[0037] FIG. 6 is a diagram showing an example of a job posting reference website RS. The job posting reference website RS lists multiple job posting links JL (job posting access information) each including a URL. On the job posting reference website RS, an individual job posting can be accessed by selecting a job posting link JL. An example of the job posting reference website RS is a job posting list page prepared for each organization and provided by an applicant tracking system (ATS: Applicant Tracking System) used by the organization. The job posting list page is a website on which a list of jobs is listed together with the job posting link JL for each job posting. The job posting list page may be a website that can be viewed (with viewing restrictions) only by organizations associated with the posted job postings.

[0038] FIG. 7 is a diagram showing an example of an information acquisition screen AD displayed on the recruiter terminal 20 for acquiring access information to a job posting or a job posting reference website. The information acquisition screen AD includes a URL input field UF and an access information input field AF. The URL input field UF accepts input from the recruiter terminal 20 of the URL of a website (job detail page) on which an individual job posting is published. The acquisition unit 112 acquires the job posting information published at the URL entered in the URL input field UF as an initial job posting. Here, the acquisition unit 112 may accept input of the URLs of multiple job detail pages in the URL input field UF.

[0039] The access information input field AF accepts input of the URL of the job posting reference website RS from the recruiter terminal 20. The URL of the job posting reference website RS may be input directly by the user from the recruiter terminal 20, or may be acquired and displayed by the acquisition unit 112 from the job posting reference website RS viewed (accessed) by the user. The acquisition unit 112 acquires the URL entered in the access information input field AF as website access information.

[0040] After acquiring the website access information, the acquiring unit 112 acquires a plurality of job posting links JL included in the job posting reference website RS using the website access information. The job posting reference website RS may include links other than the job posting links JL (e.g., links to a privacy policy display page, a help page, etc.). Therefore, the acquiring unit 112 extracts only the job posting links JL from among the links included in the job posting reference website RS, for example, by URL pattern recognition.

[0041] Specifically, the acquisition unit 112 extracts the job posting link JL by, for example, referring to a database in which URL patterns of the job posting link JL are registered. URL patterns may be registered on a per job posting reference website RS (per ATS) basis. Examples of URL patterns for determining the job posting link JL include patterns that include specific keywords such as "career" or "job," and patterns in which a string of a predetermined number of digits is added following the URL of the job posting reference website RS. The acquisition unit 112 extracts URLs that fit these patterns as the URL of the job posting link JL.

[0042] In addition, if the page of the URL entered in the access information input field AF cannot be found, or if the page of the URL entered in the access information input field AF is not the job posting reference website RS (in other words, if the page of the URL entered in the access information input field AF does not contain the job posting link JL), the acquisition unit 112 displays an error on the information acquisition screen AD.

[0043] The acquisition unit 112 may also display one input field (for example, only the access information input field AF) on the information acquisition screen AD. The acquisition unit 112 may be configured to determine whether the URL entered in the input field is the URL of a job details page or the URL of a job posting reference website, to acquire an initial job posting in the former case, and to acquire a job posting link JL in the latter case. The acquisition unit 112 may determine whether the entered URL is the URL of a job posting reference website, for example, by referring to a database in which patterns of URLs of job posting reference websites (for example, patterns including keywords such as "list") are registered. The acquisition unit 112 may also determine whether the entered URL is the URL of a job posting reference website, depending on whether the page of the entered URL includes a link to a job details page (job posting link).

[0044] Fig. 8 is a diagram showing an example of a state in which a job posting link JL has been acquired from a job posting reference website RS entered in the access information input field AF of the information acquisition screen AD. As shown in Fig. 8, the acquisition unit 112 displays a list of the acquired job posting links JL (job posting URLs) on the information acquisition screen AD. The user selects the job posting link JL from which the job posting is to be acquired (i.e., the initial job posting) by, for example, checking a checkbox.

[0045] Also, as shown in FIG. 8, the acquisition unit 112 may group and display a plurality of job posting links JL. That is, the acquisition unit 112 may create an access information group by grouping a plurality of job posting access information by analyzing the URL of each job posting access information. The acquisition unit 112 creates an access information group by, for example, pattern recognition of characters (alphabets, numbers, symbols, etc.) included in the URL. That is, the acquisition unit 112 groups URLs with high similarity in their patterns. The similarity of the URLs is calculated, for example, according to a rule that defines a criterion for determining the similarity. The similarity may be calculated by a known method such as a feature vector or Levenshtein distance. In addition, as a criterion for determining the similarity, for example, the number of overlapping characters in the two URLs, the replaceability of characters when the two URLs are partially compared (for example, character strings separated by delimiters such as " / " and "."), etc. are used. Furthermore, the URLs (job listing access information) may be grouped based on a correspondence table that records the relationship between the contents of initial job listings previously acquired by the acquisition unit 112 and the URLs of those initial job listings. This correspondence table is prepared for each medium, such as a system or service, that publishes job listings. The correspondence table also describes URL patterns in association with conditions in the job listing, such as job type, industry, and annual salary. The acquisition unit 112 refers to this correspondence table to classify and group job listing links JL according to conditions, such as job type, industry, and annual salary.

[0046] When the Import button B1 is pressed while one or more job posting links JL are selected, the acquisition unit 112 executes a web scraping process on the selected job posting links JL to acquire job posting data from the web pages of the respective URLs. In other words, the acquisition unit 112 acquires initial job postings from each of the selected job posting access information.

[0047] Furthermore, the user can collectively select multiple job posting links JL included in a URL group by selecting the URL group by checking a check box attached to each URL group, for example. In this manner, the acquisition unit 112 may accept a selection on an access information group basis on the information acquisition screen AD, and acquire an initial job posting from the job posting access information included in the selected access information group. This allows the user to collectively convert multiple initial job postings that are highly similar or related to each other into a new job posting, thereby improving the efficiency of the job posting conversion work.

[0048] The acquisition unit 112 may directly acquire the initial job listing from a system that provides a job listing, such as an ATS. In this case, the acquisition unit 112 accepts an input of a system name (or a service name) or a selection from options. The acquisition unit 112 may further accept a login operation (input of an ID, a password, etc.) required by the system from the user, and acquire multiple initial job listings from the system to which the user has logged in by API linkage, crawling processing, scraping processing, etc. The acquisition unit 112 may also accept an input of a link to a storage (or a folder in the storage) in which the initial job listing is saved, and acquire multiple initial job listings by API linkage, crawling processing, scraping processing, etc. in the storage. In other words, the initial job listing may be acquired not only from a job listing reference website, but also from a job listing reference system or a job listing reference storage. The job listing reference storage includes a storage of a system (job listing reference system) that includes job listing access information, a storage that includes a folder (job listing reference folder) in which the job listing is saved, etc.

[0049] The acquisition unit 112 may accept direct input of data of an initial job posting. For example, the acquisition unit 112 may accept upload of document data (e.g., PDF data) of an initial job posting, upload of a folder including data of an initial job posting, input of a link (e.g., URL) to data of an initial job posting stored in cloud storage, etc., and acquire the initial job posting from these data.

[0050] The acquisition unit 112 may acquire the URL or data of the initial job posting from the recruitment web page of the organization to which the user (employer) belongs. For example, the acquisition unit 112 may identify the organization to which the user belongs from the user's login information (registration information of the employer) to the recruitment support system 1, acquire the URL of the recruitment web page of the organization, and acquire the job posting link included in the recruitment web page. This allows the user to omit inputting the URL of the initial job posting or the job posting reference website.

[0051] The initial job posting acquired by the acquisition unit 112 is registered, for example, in a database for initial job postings stored in the storage unit 12. The initial job posting acquired by the acquisition unit 112 may include only the main text of the job posting (information related to the job content), or may include additional information other than the main text. Examples of additional information include control information (e.g., HTML tags) that controls the configuration, display format, etc. of the initial job posting. When acquiring the initial job posting by web scraping processing on the job posting link JL, the acquisition unit 112 may acquire the initial job posting as text data including HTML tags.

[0052] <Job Posting Evaluation Department 113> The job posting evaluation unit 113 is configured to calculate scores for a plurality of initial job postings based on reference information including a correlation between the contents of the job postings and the scores of the job postings.

[0053] For example, the job posting evaluation unit 113 calculates an expected value regarding the occurrence rate or the number of occurrences of actions on the initial job posting as a score, based on the content of the initial job posting acquired by the acquisition unit 112 and the reference information. "Actions on the initial job posting" include employer actions, which are actions taken by the employer based on the initial job posting, and job seeker actions, which are actions taken by the job seeker on the initial job posting.

[0054] Employer actions include direct actions that originate from an initial job posting (e.g., sending a scouting document based on the initial job posting) and indirect actions that originate from another action (e.g., sending a scouting document and making a job offer for a job posting for which a job seeker has responded, or closing a job posting for which a scouting document has been sent). Job seeker actions also include direct actions that originate from an initial job posting (e.g., applying for a job posting) and indirect actions that originate from another action (e.g., replying to a scouting document sent by the employer, accepting a job offer from the employer, or closing a job posting for which a job seeker has applied).

[0055] Specifically, the employer actions for which the job advertisement evaluation unit 113 calculates the expected value may include at least one of sending a scouting document to a job seeker based on an initial job advertisement, a reply from the job seeker to the scouting document, and a job offer to the job seeker. Also, the job seeker actions for which the job advertisement evaluation unit 113 calculates the expected value may include at least one of an application from a job seeker in response to the job advertisement, and a job offer to a job seeker who has applied. Furthermore, the job advertisement evaluation unit 113 may calculate the expected value for actions based on the initial job advertisement, such as passing a document screening, passing a job interview (first interview, second interview, etc.), and receiving a job offer.

[0056] The reference information includes a correlation between the contents of the job posting and the occurrence or occurrence number of actions in the job posting. The reference information is stored, for example, in the storage unit 12. The reference information is an estimator constructed so as to be able to output expected value information of the initial job posting (a predicted occurrence rate or a predicted occurrence number of actions) by inputting an initial job posting.

[0057] The reference information may be an expectation value calculation model that has been trained to be able to input the initial job advertisement and output the expectation value information. In this case, the job advertisement evaluation unit 113 inputs the initial job advertisement into the expectation value calculation model of the artificial intelligence unit 120 and causes the expectation value calculation model to output the expectation value information.

[0058] The expected value calculation model is a learning model that predicts whether or not an action will occur in an initial job posting, or the number of occurrences. The expected value calculation model is a learning model that is trained using a reference job posting for learning and data on the occurrence rate or number of occurrences of the corresponding action as training data. A "reference job posting" is a job posting that is registered or has been registered in a job posting database, and that has actually recruited in the past or is currently recruiting. In this case, the parameters of the expected value calculation model calculated or tuned by learning correspond to the correlation between the contents of the job posting and whether or not an action will occur in the job posting, or the number of occurrences. Here, the reference job posting for learning may include text (text data) related to the work content, etc., and attribute information that is selected and input from predetermined categories or options, such as job type, industry, work location, and annual income range.

[0059] The expected value calculation model may be a generation AI including a large-scale language model. In this case, the job posting evaluation unit 113 inputs an initial job posting, inputs a prompt including an instruction to predict and output the occurrence rate or occurrence number of the action of the initial job posting, and causes the expected value calculation model to output the occurrence rate or occurrence number of the action. In addition, the job posting evaluation unit 113 may input, in addition to the initial job posting and the instruction to predict and output the occurrence rate or occurrence number of the action, a prompt in which, for example, one or more samples of the job posting and one or more samples of the occurrence rate or occurrence number of the action corresponding thereto are inserted as input and output samples, to the expected value calculation model.

[0060] The job advertisement evaluation unit 113 may, for example, input a feature amount including a first vector obtained by vectorizing the text included in the initial job advertisement into the expected value calculation model, and cause the expected value calculation model to output expected value information. The job advertisement evaluation unit 113 may also input a feature amount including a second vector obtained by vectorizing the attribute information included in the initial job advertisement into the expected value calculation model, and cause the expected value calculation model to output expected value information. The job advertisement evaluation unit 113 may also input a feature amount including both the first vector and the second vector into the expected value calculation model, and cause the expected value calculation model to output expected value information. The expected value calculation model is trained to receive a feature amount including at least one (preferably both) of the first vector obtained by vectorizing the text included in the initial job advertisement and the second vector obtained by vectorizing the attribute information included in the job advertisement as input, and output expected value information.

[0061] The job advertisement evaluation unit 113 may generate at least two of the following: first expected value information, which is the probability of a job offer being concluded; second expected value information, which is the probability of a reply from a job seeker; and third expected value information, which is a predicted value of the number of scout documents to be sent. Furthermore, the job advertisement evaluation unit 113 may calculate a score for the initial job advertisement based on at least two of the first expected value information, the second expected value information, and the third expected value information.

[0062] Furthermore, for example, the job posting evaluation unit 113 may output a score of the evaluation element included in the initial job posting acquired by the acquisition unit 112, using reference information in which the relationship between the evaluation element and the score is defined. The evaluation element may include one or more elements selected from the amount of information included in the initial job posting and the sentence expression included in the initial job posting. The evaluation element may include an element calculated based on whether the initial job posting contains a term expressing the job content of the human resources desired by the employer.

[0063] The job posting evaluation unit 113 may calculate a score for the amount of information included in the initial job posting based on the amount of information included in the items constituting the initial job posting. For example, a score for each item may be calculated based on a comparison result between the number of characters included in each item and a predetermined range of appropriate character numbers for each item, and a score for the amount of information in the entire initial job posting may be calculated based on the score for each item. In addition, the job posting evaluation unit 113 may calculate a score for the expression of the text included in the initial job posting based on the consistency between the items constituting the initial job posting and the text included in the item. For example, if the "job description" item contains text other than the job description (for example, content related to application qualifications), the score for the expression may be calculated to be low.

[0064] The job posting evaluation unit 113 may calculate a score for the expression of the sentences included in the initial job posting based on the consistency between the sentences included in each of two or more items constituting the initial job posting. For example, if a sentence included in the "job description" item is inconsistent with a sentence included in the "appealing points" item, the score for the expression may be calculated to be low. Furthermore, the job posting evaluation unit 113 may calculate a score for the expertise of the human resources desired by the employer by referring to a database in which terms expressing the job description of the human resources desired by the employer are registered. Such analysis of the sentences and calculation of the score may be performed by natural language processing including morphological analysis using a learning model of the artificial intelligence unit 120.

[0065] The job advertisement evaluation unit 113 may be configured to calculate a score for a new job advertisement created by the job advertisement regeneration unit 115 described below.

[0066] <List display control unit 114> The list display control unit 114 is configured to display a list of information on multiple initial job listings acquired by the acquisition unit 112, together with the scores calculated by the job listing evaluation unit 113, on the employer terminal 20. This makes it easy to understand and manage the status of the imported initial job listings (e.g., the status of conversion to new job listings). In addition, by displaying the scores of the initial job listings, the user can consider whether or not to convert to a new job listing depending on the scores.

[0067] 9 is a diagram showing an example of a job posting list screen LD displayed on recruiter terminal 20. The job posting list screen LD includes a job posting list LT that displays a plurality of job postings stored in memory unit 12 in list form, an add import button B2, and a new creation button B3. The job postings displayed on the job posting list screen LD include initial job postings acquired (imported) from outside the recruiting support system 1, and regular job postings created and registered within the recruiting support system 1. The job posting list LT displays the job title JT, job posting creator JN, job posting type JK, job posting score JS, service registration status JR, copy button B4, and delete button B5 for each job posting.

[0068] The job title JT is the title or position name of the job posting. The job posting type JK is a label indicating whether it is an initial job posting acquired from outside or a normal job posting created and registered internally. In the example of FIG. 9, the initial job posting is labeled "imported", and the normal job posting is labeled "created by AI", "created manually", etc. The label "created by AI" is attached to a normal job posting created by, for example, the artificial intelligence (learning model) of the artificial intelligence unit 120, and the label "created manually" is attached to a normal job posting created by the user manually entering items. The employability displayed as the job posting score JS is a score calculated by the job posting evaluation unit 113 and represents the expected value (probability) of the conclusion of the job posting. The list display control unit 114 may display a score other than the expected value of the conclusion of the job posting as the job posting score JS.

[0069] The service registration status JR is information indicating whether or not a job posting has been registered in the recruiting support system 1. "Registered in the recruiting support system 1" means that in the services made up of the recruiting support system 1, a scouting document can be sent to a job seeker based on the job posting, and / or the job seeker can search for and refer to the job posting. In the case of an initial job posting, if it has been converted into a new job posting, it will be in a "registered" state, and if it has not been converted into a new job posting, it may be in a "not registered" state.

[0070] When the Add Import button B2 is operated, a screen that accepts input to acquire a new initial job posting (for example, the information acquisition screen AD in FIG. 7) is displayed on the recruiter terminal 20. When the New Create button B3 is operated, a screen that accepts input to create a new job posting is displayed on the recruiter terminal 20. When the Duplicate button B4 is operated, a duplicate job posting of the corresponding job posting is created. When the Delete button B5 is operated, the corresponding job posting is deleted from the job posting list screen LD and the database in the memory unit 12.

[0071] By selecting a job posting displayed in the job posting list LT, the user can register the job posting (convert it into a new job posting), modify it, etc.

[0072] The list display control unit 114 may recommend initial job postings to be converted into new job postings based on the items displayed on the job posting list screen LD. For example, the list display control unit 114 may highlight job postings whose job posting scores JS are higher than a predetermined value (i.e., there is a high possibility that the job seeker can be hired) and that are not registered in the hiring support system 1. This can help the user decide whether or not to convert.

[0073] <Recruitment form regeneration unit 115> The job listing regeneration unit 115 is configured to regenerate a new job listing from an initial job listing (convert the initial job listing into a new job listing). Specifically, the job listing regeneration unit 115 inputs a plurality of job listing components extracted from the initial job listing into a classification model, causes the classification model to output classification information for each job listing component, and arranges the plurality of job listing components into a job listing format having a different configuration from the initial job listing based on the classification information.

[0074] A job posting component includes at least one word or sentence. A job posting component may be composed of only a keyword such as a tag word, may be a sentence composed of one or more sentences, or may be composed of a keyword such as a tag word and a sentence. Furthermore, job posting components are extracted regardless of the configuration of the items, paragraphs, etc. of the initial job posting. Therefore, multiple sentences written in the same item or paragraph of the initial job posting may be extracted as separate job posting components. Note that job posting components are words or sentences extracted as they are from the contents of the initial job posting (the body of the job posting). In other words, job posting components are words or sentences extracted from the contents of the initial job posting without editing, such as replacement, addition, or deletion.

[0075] Classification information is information that indicates the item in which a job posting component is placed in the job posting format. In the example of the job posting JP in FIG. 5, the classification information is information that indicates whether a job posting component is placed in the "position name" item IM, the "job content and working conditions" item IM, the "application qualifications" item IM (more specifically, the sub-item SIM of "required requirements" and the sub-item SIM of "welcome requirements") item IM, or the "desired talent" item IM. For example, a job posting component to which a code indicating "job content and working conditions" has been assigned as classification information by the classification model is placed as text written in the "job content and working conditions" item IM. The classification model will be described later.

[0076] In extracting job posting components from the initial job posting, the job posting regeneration unit 115 may execute a first division process to divide the contents of the initial job posting (the main text of the initial job posting) into at least one intermediate element, and a second division process to divide the intermediate element into multiple job posting components. This allows the job posting components to be extracted by dividing the contents of the initial job posting in two stages, improving the accuracy of adding classification information to the job posting components (i.e., classifying them according to the items of the new job posting).

[0077] In the first division process, the job posting regeneration unit 115 may divide the contents of the initial job posting into at least one intermediate element based on a delimiter included in the initial job posting. Delimiters include, for example, control symbols such as HTML tags, and symbols indicating the division of sentences or paragraphs, such as periods, colons, semicolons, slashes, and line feed symbols. This makes it possible to extract intermediate elements suitable for the second division while reducing the processing load in the first division process.

[0078] Furthermore, in the first division process, the job advertisement regeneration unit 115 may input the contents of the initial job advertisement into the first division model and cause the first division model to output at least one intermediate element. The first division model is a learning model that is trained so that it can input the contents of the initial job advertisement and output multiple intermediate elements based on predetermined intermediate element classification information. This allows intermediate elements to be extracted according to the contents of the contents of the initial job advertisement, thereby improving the accuracy of extraction of job advertisement components.

[0079] The intermediate element classification information is information that defines categories when the contents of the initial job posting are divided into multiple intermediate elements, and the first division model outputs the contents divided into categories included in the intermediate element classification information as intermediate elements. The categories defined by the intermediate element classification information are the same as the items (including sub-items) in the job posting format of the new job posting (i.e., the converted job posting format), the items that are conceptualized as higher-level items (multiple items grouped together), or the items that are subdivided (divided). In addition, the categories defined by the intermediate element classification information may include items that are not included in the new job posting. Specifically, the intermediate element classification information includes information that divides and classifies the contents into "title" (an example of a category that is the same as the items in the job posting format), "job content" (an example of a category that subdivides the item "job content and working conditions" in the job posting format), "working conditions" (same as above), "application qualifications" (an example of a category that conceptualizes the items "required requirements" and "welcome requirements" in the job posting format), "company information" (an example of a category that is not included in the job posting format), etc. The intermediate element classification information is stored in the storage unit 12, for example, as a table in which the names of a plurality of categories are described.

[0080] The first division model is a learning model that learns using data on the contents of the description of the initial job posting and the corresponding data on a plurality of intermediate elements divided based on the intermediate element classification information as training data. The first division model may also be a generation AI including a large-scale language model. In this case, the job posting regeneration unit 115 inputs the contents of the initial job posting, inputs a prompt including an instruction to generate an intermediate element based on the contents and the intermediate element classification information to the first division model, and causes the first division model to output the intermediate element. In addition, the job posting regeneration unit 115 may input a prompt into the first division model in which, in addition to the instructions for generating and outputting the intermediate element and the contents of the initial job posting, a sample of the description of one or more initial job postings and a sample of one or more corresponding intermediate elements are inserted as input and output samples.

[0081] The first division model may assign category information based on the intermediate element classification information to each of the divided intermediate elements. For example, the category information (category label) of "work content" may be assigned to an intermediate element divided into the category of "work content".

[0082] The job posting regeneration unit 115 may divide a part of the content of the initial job posting into intermediate elements based on the delimiter, and may divide the remaining part into intermediate elements by the first division model. For example, the job posting regeneration unit 115 may divide the HTML "title tag" ( <title> and< / title> Based on the combination of the above, extract the title of the initial job posting (for example, the HTML "title tag" <title> The character string following the is extracted as the title, and the content excluding the title may be input to the first division model. The extraction of intermediate elements including the title of the initial job posting is not limited to the method based on the "title tag", and intermediate elements may be extracted based on other tags or elements of HTML.< / title>

[0083] If the intermediate element is divided by a delimiter in the first division process, the job posting regeneration unit 115 may divide the intermediate element into multiple job posting components in the second division process based on a delimiter included in the initial job posting that is different from the delimiter used in the first division process. This reduces the processing load in the second division process and makes it possible to extract job posting components suitable for creating a new job posting.

[0084] The delimiter used in the second division process is a delimiter included in the intermediate element divided by the first delimiter. The delimiter used in the second division process may be a delimiter with a smaller division unit than the delimiter used in the first division process. For example, if the delimiter used in the first division process is a control symbol such as an HTML tag, the intermediate element may be divided in the second division process by a symbol indicating a sentence or paragraph separator, such as a full stop, period, colon, semicolon, slash, or line feed symbol. In particular, the intermediate element may be divided in the second division process by a symbol indicating a paragraph separator (for example, two consecutive line feed symbols forming a blank line). This ensures that the job posting component elements suitable for adding classification information are divided.

[0085] In the second division process, the job posting regeneration unit 115 may input the intermediate elements to a second division model and cause the second division model to output a plurality of job posting components. The second division model is a learning model that is trained so that it can input the intermediate elements and output a plurality of job posting components in which the sentences included in the intermediate elements are grouped together based on a common theme. This generates job posting components that are divided by the subject of the sentences (i.e., divided according to the context), thereby improving the accuracy of assigning classification information by the classification model described below.

[0086] The second division model is a learning model that learns using data of intermediate elements and corresponding data of job posting components in which sentences are organized by a common theme as training data. The second division model may also be a generation AI including a large-scale language model. In this case, the job posting regeneration unit 115 inputs the intermediate elements, inputs a prompt including an instruction to generate a job posting component based on the intermediate elements to the second division model, and causes the second division model to output the job posting component. The job posting regeneration unit 115 may also input a prompt into the second division model in which, in addition to the instruction to generate and output the job posting component and the intermediate elements, one or more samples of intermediate elements and one or more samples of the corresponding job posting component are inserted as input and output samples.

[0087] The second division model may divide the intermediate elements by referring to the category information that the first division model has assigned to the intermediate elements. For example, for an intermediate element to which category information of "job content" has been assigned, the second division model may divide the intermediate element into job posting components by extracting a theme related to the "job content". The second division model may also assign theme information to each of the divided job posting components. For example, a job posting component divided from an intermediate element having category information of "job content" may be assigned theme information (theme label) such as "business overview" and "attractiveness of the job". Similarly, for example, a job posting component divided from an intermediate element having category information of "application qualifications" may be assigned theme information such as "required requirements" and "welcome requirements".

[0088] The second division model may divide the intermediate elements into job posting components by extracting a theme that is not included in the category of the intermediate elements. For example, a job posting component including a theme of "required requirements" may be extracted from an intermediate element having category information of "job content."

[0089] The job posting regeneration unit 115 may divide some of the intermediate elements generated in the first division process into job posting constituent elements based on the delimiter, and may divide the remaining intermediate elements into job posting constituent elements by the second division model. In addition, the job posting regeneration unit 115 may directly treat some of the intermediate elements (for example, the intermediate element extracted as "title") as job posting constituent elements without performing the second division process.

[0090] Furthermore, the job posting regeneration unit 115 may execute the first division process and the second division process using one component extraction model in which the first division model and the second division model are integrated. In this case, the job posting regeneration unit 115 inputs the contents of the initial job posting to the component extraction model, and causes the component extraction model to output a plurality of job posting components. For example, if the body of the job posting is composed of only one item (if it is divided by only one HTML tag), the job posting regeneration unit 115 may extract the job posting components from the contents (body) of the initial job posting using the component extraction model, without performing two-stage division.

[0091] The component extraction model is a learning model that is trained to be able to input the contents of the initial job posting and output a plurality of job posting components in which the sentences included in the contents are grouped by a commonality of the subject. The component extraction model is a learning model that is trained using the data of the contents of the initial job posting and the corresponding data of the job posting components in which the sentences are grouped by a commonality of the subject as training data. The component extraction model may also be a generation AI including a large-scale language model. In this case, the job posting regeneration unit 115 inputs the contents of the initial job posting, inputs a prompt including an instruction to generate a job posting component based on the contents of the posting to the component extraction model, and causes the component extraction model to output the job posting component. In addition, the job posting regeneration unit 115 may input a prompt into the component extraction model in which, for example, a sample of the contents of the initial job posting and a sample of one or more corresponding job posting components are inserted as input and output samples in addition to the instructions to generate and output the job posting components and the contents of the initial job posting.

[0092] After extracting the job listing components, the job listing regeneration unit 115 inputs the multiple job listing components extracted from one initial job listing together into a classification model, or inputs the multiple job listing components one by one into a classification model. The classification model outputs the job listing components with classification information attached, for example.

[0093] The classification model is a learning model trained to be able to input job posting components and output classification information. The classification model is a learning model trained using data of the job posting components and data of the corresponding classification information as teacher data. The classification model may also be a generation AI including a large-scale language model. In this case, the job posting regeneration unit 115 inputs the job posting components, inputs a prompt including an instruction to output classification information corresponding to the job posting components to the classification model, and causes the classification model to output the classification information. In addition, the job posting regeneration unit 115 may input a prompt into the classification model in which, for example, samples of one or more job posting components and samples of one or more corresponding classification information are inserted as input and output samples in addition to the instruction to output the classification information and the job posting components.

[0094] The classification model may generate classification information for the job posting component by referring to the subject information that the second division model has assigned to the job posting component. For example, for a job posting component to which subject information of "job summary" has been assigned, the classification model may assign classification information of "job content and working conditions" to the job posting component by referring to the subject information.

[0095] When multiple job posting components are input together into a classification model, which is a large-scale language model, the job posting regeneration unit 115 may input an instruction to generate classification information for each job posting component to the classification model. This improves the accuracy of the classification information assigned to the multiple job posting components. Specifically, the job posting regeneration unit 115 inputs multiple job posting components, inputs a prompt including an instruction to output classification information corresponding to each of the multiple job posting components to the classification model, and causes the classification model to output the classification information for each of the multiple job posting components.

[0096] The job posting regeneration unit 115 may cause the classification model to output identifiers individually assigned to the job posting components and classification information associated with the identifiers. This reduces the amount of information output from the classification model compared to when the classification model outputs job posting components (words or sentences) to which classification information is assigned, thereby reducing the processing load of the control unit 11. In addition, when there is a restriction on the upper limit of the amount of information that can be output from the classification model, this restriction can be avoided. The identifiers (IDs) assigned to the job posting components are composed of individually identifiable (non-overlapping) codes (e.g., "ID001" and the like). The classification information may be associated with individual identifiers, or may be generated in association with multiple identifiers. The classification information associated with multiple identifiers includes a correspondence between the items in the job posting format and the identifiers corresponding to (placed on) the items, such as, for example, "attractiveness of the job" - "ID001·ID003", "essential conditions" - "ID005", and the like.

[0097] When causing the classification model to output an identifier, the job posting regeneration unit 115 inputs job posting components to the classification model, which is a large-scale language model, inputs a prompt to the classification model including an instruction to output identifiers and classification information corresponding to each of the job posting components, and causes the classification model to output the identifiers and classification information.

[0098] The job listing regeneration unit 115 may assign an identifier to each of a plurality of job listing components, and then input the job listing components and the identifiers to the classification model, and cause the classification model to output classification information associated with the identifiers. The job listing regeneration unit 115 may also input pairs of job listing components and identifiers to the classification model one by one, and cause the classification model to output pairs of identifiers and classification information one by one.

[0099] The job posting regeneration unit 115 may extract job posting components and output classification information using a first lump-sum classification model in which the first division model, the second division model, and the classification model are integrated. In this case, the job posting regeneration unit 115 inputs the contents of the initial job posting to the first lump-sum classification model, and causes the first lump-sum classification model to output multiple job posting components and classification information for each job posting component.

[0100] The first batch classification model is a learning model that has been trained using data on the contents of the initial job posting and data on the corresponding job posting components and classification information as training data. The first batch classification model may also be a generation AI including a large-scale language model. In this case, the job posting regeneration unit 115 inputs the contents of the initial job posting, inputs a prompt including an instruction to output the job posting components and classification information corresponding to the contents of the initial job posting to the first batch classification model, and causes the first batch classification model to output the job posting components and classification information. In addition, the job posting regeneration unit 115 may input a prompt into the first batch classification model that has, for example, a sample of the contents of the initial job posting and a sample of each of the corresponding one or more job posting components and classification information, as input and output samples, in addition to the output instruction for the job posting components and classification information and the contents of the initial job posting.

[0101] Furthermore, the job posting regeneration unit 115 may extract job posting components and output classification information using a second lump classification model in which the second division model and the classification model are integrated. In this case, the job posting regeneration unit 115 inputs the intermediate elements to the second lump classification model and causes the second lump classification model to output a plurality of job posting components and classification information for each job posting component. In other words, the second lump classification model is a learning model trained by replacing the input of the first lump classification model with the intermediate elements, or a large-scale language model to which a prompt including an instruction to input the intermediate elements and output the job posting components and classification information is input.

[0102] With classification information assigned to the job posting components or identifiers, the job posting regeneration unit 115 places the multiple job posting components in the items of the job posting format indicated by the classification information associated with the job posting components. This generates a new job posting. For example, a job posting component associated with classification information indicating the "Job Description / Working Conditions" item of the job posting format is placed in the item as a word, sentence, or paragraph that constitutes the content of the item.

[0103] Furthermore, when the classification model is caused to output the identifier of a job posting component (i.e., when classification information is associated with the identifier of a job posting component), the job posting regeneration unit 115 places the job posting component corresponding to the identifier associated with the classification information in the item indicated by the classification information in the job posting format. This makes it possible to rearrange the job posting component in the job posting format while avoiding the content of the job posting component being converted by the classification model.

[0104] A combination of multiple job posting components may be placed in one item of the job posting format. When multiple job posting components are placed in one item, the order in which the multiple job posting components are placed in the item (e.g., the order of the sentences) is determined according to a predetermined rule. The predetermined rule is, for example, the order in which they were written in the initial job posting (the order according to the location of the item in which they were written), the order in which they move from abstract to concrete content, etc. In addition, the job posting regeneration unit 115 may assign identifiers to the job posting components based on the order in which they were written in the initial job posting (i.e., assign identifiers including numbers indicating the order), and place the multiple job posting components in one item in the order indicated by the identifiers (starting with the lowest number). This allows the job posting components to be placed without changing the order in which they were written in the initial job posting.

[0105] <Item addition part 116> The item adding unit 116 presents options to be added to predetermined items of the job listing format based on the contents of the initial job listing. The options are extracted from preset options. This allows information that is not directly included in the initial job listing or information that was not individually extracted from the initial job listing to be individually added to predetermined items of the new job listing without changing the contents of the job listing components extracted from the initial job listing.

[0106] An option is a keyword to be placed in a specific item. The item adding unit 116 presents multiple keywords as option candidates. An option may be a keyword included in the initial job posting, or may be a keyword not included in the initial job posting. When an option is a keyword included in the initial job posting, the option is placed in an item indicated in the classification information of the job posting format as a part of the job posting component, and is also placed as a keyword (e.g., a tag) in another item specified as an option.

[0107] The item adding unit 116 may present a plurality of types of options. Examples of information indicated by each option include job type, work location, annual salary, working style (e.g., whether remote work is possible), and work environment (e.g., whether passive smoking measures are in place). These options are added to the "Job Description and Working Conditions" item of the new job posting, for example.

[0108] The option candidates presented by the item adding unit 116 are extracted, for example, from a keyword list for options in which candidate keywords are described for each type of option. The keyword list is stored, for example, in the storage unit 12. The item adding unit 116 extracts, as option candidates, keywords related to the contents of the initial job posting from among the keywords described in the keyword list. For example, the item adding unit 116 compares, for each type of option, the keywords described in the keyword list with the contents of the initial job posting, and presents keywords that are the same as or similar to the words included in the contents of the initial job posting as option candidates. In addition, for a numerical option such as annual salary, the item adding unit 116 may calculate the value of the annual salary to be presented as an option from the numerical values ​​of monthly salary, bonus, etc. described in the initial job posting.

[0109] The item adding unit 116 may present option candidates using an option candidate extraction model. In this case, the item adding unit 116 inputs the description content of the initial job advertisement into the option candidate extraction model, and causes the option candidate extraction model to output at least one option candidate. The option candidate extraction model is a learning model that learns using data of the description content of the initial job advertisement and data of the corresponding option candidate as teacher data. The option candidate extraction model may also be a generation AI including a large-scale language model. In this case, the item adding unit 116 inputs the description content of the initial job advertisement, refers to a keyword table (or also inputs the keyword table), inputs a prompt including an instruction to output option candidates corresponding to the description content of the initial job advertisement into the option candidate extraction model, and causes the option candidate to be output by the option candidate extraction model. In addition, the item adding unit 116 may input a prompt into the option candidate extraction model in which, for example, a sample of the description content of one or more initial job advertisements and a sample of one or more option candidates corresponding thereto are inserted as input and output samples in addition to the output instruction of the option candidate and the description content of the initial job advertisement.

[0110] Option candidates are presented on the recruiter terminal 20, for example, as tag-like objects that can be added and deleted, as list items such as a pull-down list, or as selectable items such as check boxes. The item adding unit 116 accepts the selection of one or more options, and adds the selected option to a predetermined item in the job posting format according to the type of option. Note that if no option is selected by the user, the item adding unit 116 does not add an option.

[0111] 10 is a diagram showing an example of a new job posting creation screen CD. The new job posting creation screen CD includes a first option selection area OS1, a second option selection area OS2, a third option selection area OS3, a fourth option selection area OS4, and a new job posting display area DA.

[0112] The new job listing display area DA displays initial job listing information IJ and the job listing body MB. The initial job listing information IJ displays, for example, job listing access information for the initial job listing (the URL from which the initial job listing was obtained). The initial job listing information IJ may also be a link for accessing the initial job listing. The job listing body MB is job listing components extracted from the initial job listing, arranged in each item based on classification information.

[0113] In the example of FIG. 10, the first option selection area OS1 presents job type options as job type tags TG1. In the initial state, keywords corresponding to the contents of the initial job posting from the job type keyword list (job type tag list) are displayed as job type tags TG1. When the delete button BD attached to each job type tag TG1 is operated, the job type tag TG1 is deleted. When the pull-down button BP is operated, job type tags that were not presented (job type keywords included in the keyword list) are displayed in a list. When a job type tag displayed in the list is selected, the job type tag is added to the first option selection area OS1. The user selects keywords to be added as jobs to the new job posting by deleting and adding job type tags TG1.

[0114] Similarly, in the second option selection area OS2, the option of work location is presented as work location tag TG2. In the initial state, keywords from the work location keyword list (work location tag list) that correspond to the contents of the initial job posting are displayed as work location tag TG2. The user selects keywords to be added as work location to the new job posting by deleting and adding the work location tag TG2 in the same procedure as for the job type tag TG1.

[0115] The third option selection area OS3 presents the minimum and maximum annual income values. In the initial state, the minimum and maximum annual income values ​​according to the contents of the initial job advertisement are displayed in the first pull-down box PL1 and the second pull-down box PL2, respectively. The minimum and maximum annual income values ​​are updated by selecting the numerical values ​​displayed in the lists in the first pull-down box PL1 and the second pull-down box PL2. The item addition unit 116 may also accept direct input of numerical values ​​from the user into the first pull-down box PL1 and the second pull-down box PL2.

[0116] The fourth option selection area OS4 shows whether remote work is available. In the initial state, the check box CB is set to on (the "remote work available" option) or off (the "remote work not available" option) according to the contents of the initial job posting. The availability of remote work can be changed by inputting a value into the check box CB.

[0117] The options set in each option selection area are inserted in the form of keywords, tags, sentences, etc. into a specified field in the job posting body MB of the new job posting display area DA (in the example of Figure 10, the field "Job description and working conditions").

[0118] The item adding unit 116 may display the contents of the initial job posting referenced when the option to be presented was extracted as the basis for the option. This allows the user to select an option to add to the new job posting and consider whether to add the option while checking the original description in the initial job posting. The basis for the option is displayed, for example, when an option candidate in each option selection area of ​​the new job posting creation screen CD is selected by hovering the cursor over it, clicking, tapping, or the like. The item adding unit 116 displays, for example, a part of the contents of the initial job posting that is the basis for the option (the part from which the option was extracted) on a screen such as a pop-up window. In addition to a word or sentence related to the option, the item adding unit 116 may also display the sentences before and after it (for example, in paragraph units) as the basis for the option. In this case, the item adding unit 116 may highlight the word or sentence related to the option.

[0119] The item adding unit 116, for example, causes the option candidate extraction model, which is a large-scale language model, to output the grounds for the options. Specifically, when causing the option candidate extraction model to output the option candidates, the item adding unit 116 inputs a prompt to the option candidate extraction model that includes an instruction to output the grounds for the options extracted from the description of the previous initial job advertisement in addition to an instruction to output the option candidates corresponding to the description of the initial job advertisement, and causes the option candidate extraction model to output the option candidates and their grounds.

[0120] <Job posting display control unit 117> The job posting display control unit 117 displays the job posting format in which the job posting components are arranged as a new job posting. The job posting display control unit 117 may also display the correspondence between the job posting components in the new job posting and the corresponding job posting components in the initial job posting. This makes it easier for the user to check the correspondence between the initial job posting and the new job posting, and any missing information in the new job posting.

[0121] The job posting display control unit 117 generates correspondence information indicating the correspondence between the job posting components of the new job posting and the job posting components in the initial job posting, for example, by performing a text comparison between the contents of the new job posting and the contents of the initial job posting. The correspondence information is referred to when displaying the correspondence on the employer terminal 20 or the like. The job posting display control unit 117 may also generate the correspondence between the job posting components of the new job posting and the job posting components in the initial job posting by inputting a prompt including an instruction to output the basis of each item of the new job posting (the extracted part of the job posting components in the initial job posting) to the classification model and outputting the basis. The job posting display control unit 117 may also generate the correspondence between the job posting components of the new job posting and the job posting components in the initial job posting by inputting a prompt including an instruction to output the basis of each of the multiple job posting components to be extracted (the extracted part in the initial job posting) to the component extraction model and outputting the basis.

[0122] The job posting display control unit 117 displays the new job posting in a new job posting display area DA of the new job posting creation screen CD shown in FIG. 10, for example. The job posting display control unit 117 may accept edits (deletion, addition, modification, etc. of keywords, tags, sentences, etc.) to the job posting body MB on the new job posting creation screen CD. The job posting display control unit 117 may also accept input of an operation to complete job posting creation on the new job posting creation screen CD. This causes the new job posting to be stored in the job posting database. The registered new job posting becomes a target for job posting searches by job seekers in the job posting / job search system constituted by the recruiting support system 1. In addition, the recruiter can create and send a scouting document to a job seeker based on the new job posting in the job posting / job search system constituted by the recruiting support system 1.

[0123] The job listing display control unit 117 accepts the selection of any job listing component (keyword, sentence, etc.) of the job listing body MB in the new job listing display area DA, and displays the location (correspondence) of the selected job listing component in the initial job listing on a screen such as a pop-up window. The job listing display control unit 117 may display the job listing component in the initial job listing and the sentences before and after it, or the paragraph (item) containing the job listing component, and may highlight the job listing component. The job listing component for which the correspondence with the initial job listing is to be displayed is selected by, for example, the user's operation input such as hovering the cursor over it, clicking, tapping, or dragging it to a specific area.

[0124] The job listing display control unit 117 may display the initial job listing as a whole and the new job listing as a whole so that they can be compared. For example, the job listing display control unit 117 may display the initial job listing and the new job listing side by side, and further display the correspondence by coloring corresponding job listing components in the same color, surrounding them with frames of the same color, connecting them with lines, attaching a common label, or the like.

[0125] <Artificial Intelligence Department 120> The artificial intelligence unit 120 is configured to receive input from each functional unit and return the instructed output. The artificial intelligence used by the server device 10 in each functional unit may be a common one, or may be prepared individually for each functional unit.

[0126] The artificial intelligence unit 120 is an AI (Artificial Intelligence) equipped with a transformer including GPT (Generative Pretrained Transformer, including GPT-1, GPT-2, and GPT-3), BERT (Bidirectional Encoder Representations from Transformers), BART (Bidirectional and Auto-regressive Transformer), etc., and a language model such as a recurrent neural network (RNN), and may include a generative AI.

[0127] The language model is an example of a learning model based on a machine learning algorithm. Specific examples of machine learning algorithms include nearest neighbor methods, naive Bayes methods, decision trees, support vector machines, and deep learning using neural networks. The artificial intelligence unit 120 can apply the above algorithms as appropriate.

[0128] The artificial intelligence unit 120 may have a trained model constructed by a learning method such as supervised learning, unsupervised learning, or self-supervised learning. In supervised learning, machine learning is performed using training data (training data). The training data is composed of a pair of input data for learning and output data (correct answer data). In addition, the language model may not only be one trained for a specific task, but also a general-purpose model that can be used for a wide range of tasks.

[0129] The artificial intelligence unit 120 may be a general-purpose natural language processing learning model such as a large-scale language model (LLM) that has learned a huge amount of data as an artificial intelligence. Such a general-purpose learning model includes a language model that can handle various tasks without fine tuning by one-shot learning, few-shot learning, etc. Also, the general-purpose learning model may be configured to be capable of handling various tasks by zero-shot learning. The artificial intelligence used in each functional unit of the control unit 11 may be a separate learning model, or may be a common general-purpose learning model.

[0130] The learning model included in the artificial intelligence unit 120 (the learning model used in each functional unit, such as the classification model) can perform additional learning as transfer learning or fine tuning. For example, each time a new initial job posting is acquired and a new job posting is created, the artificial intelligence unit 120 may perform additional learning and fine tuning using these as new teacher data. This improves the accuracy of the classification information output from the learning model.

[0131] The learning model included in the artificial intelligence unit 120 may be a learning model (distilled model) obtained by knowledge distillation using an original learning model. In knowledge distillation, a trained model such as a large-scale language model is used as a teacher model, and the parameters of the student model are adjusted so that the output loss (Soft Target Loss) of the student model (distilled model) relative to the output (Soft Target) of the teacher model is reduced, thereby learning the student model, and the student model becomes a distilled model. In addition, the student model may be trained so that the output loss (Hard Target Loss) of the student model relative to the correct label (Hard Target) of the teacher data (combination of input data and output data of the learning model) is reduced. Compared to the original learning model (teacher model), the distilled model has a smaller number of parameters and a smaller processing load while having performance close to the learning model. Therefore, by using the distilled model, the cost of the recruitment support system 1 can be reduced.

[0132] For example, the classification model may be a distilled model trained using a combination of input data and output data in a large-scale language model as training data. In addition, when the recruitment support system 1 is introduced, a large-scale language model may be used as a classification model, and when training data from the large-scale language model is accumulated, a distilled model obtained by knowledge distillation using the training data may be used as the classification model.

[0133] <Display> The display unit 211 of the recruiter terminal 20 and the display unit 311 of the job seeker terminal 30 each display a screen indicated by the screen data transmitted from the server device 10.

[0134] <Operation acquisition section> The operation acquisition unit 212 of the recruiter terminal 20 accepts operations by a user (recruiter) who uses the recruiter terminal 20. The operation acceptance unit 312 of the job seeker terminal 30 accepts operations by a user (job seeker) who uses the job seeker terminal 30.

[0135] 3. Information processing method This section describes an information processing method of the server device 10. This information processing method is executed by a computer, with each unit of the server device 10 acting as each step.

[0136] Specifically, this information processing method includes an acquisition step, a job posting regeneration step, an item addition step, and a job posting display control step. In the acquisition step, an initial job posting is acquired. In the job posting regeneration step, multiple job posting components extracted from the initial job posting are input into a classification model, classification information for each job posting component is output from the classification model, and the multiple job posting components are arranged in a job posting format having a different configuration from the initial job posting based on the classification information. In the item addition step, an option to add to a specified item in the job posting format is presented based on the contents of the initial job posting. In the job posting display control step, the job posting format in which the job posting components are arranged is displayed as a new job posting, and the correspondence between the job posting components in the new job posting and the job posting components in the initial job posting is displayed.

[0137] 11 is an activity diagram showing the flow of information processing (processing for converting an initial job advertisement into a new job advertisement) executed by the hiring support system 1. Below, the information processing will be explained along with each activity in this activity diagram.

[0138] The process of converting an initial job posting into a new job posting begins with the selection of an initial job posting by the employer. The employer selects the initial job posting to be converted into a new job posting on the employer terminal 20 (activity A101). The initial job posting is selected by inputting the job posting access information (URL) of the initial job posting, selecting from a list of initial job postings acquired by web scraping, etc.

[0139] The server device 10 acquires information on the initial job posting selected on the employer terminal 20 from an external job posting database or the like (activity A102). Next, the server device 10 extracts intermediate elements obtained by dividing the contents of the initial job posting (activity A103). Furthermore, the server device 10 extracts job posting components obtained by dividing the intermediate elements (activity A104).

[0140] After extracting the job listing components, the server device 10 generates classification information for the job listing components using a classification model (activity A105). Next, the server device 10 arranges the job listing components in the job listing format of a new job listing based on the classification information (activity A106). Furthermore, the server device 10 extracts option candidates to be added to the new job listing (activity A107). Note that the extraction of option candidates does not necessarily have to be performed after the arrangement of the job listing components in activity A106, and can be performed at any timing after the initial job listing is obtained in activity A102.

[0141] After the job posting components have been arranged and option candidates have been extracted, server device 10 outputs the new job posting with the job posting components arranged and the option candidates to recruiter terminal 20 (activity A108). This causes the new job posting and option candidates to be displayed (presented) on recruiter terminal 20. The job seeker selects options and edits the content of the new job posting as necessary on recruiter terminal 20 (activity A110). Server device 10 accepts input of options and edits from recruiter terminal 20, and registers the new job posting reflecting these in its internal job posting database (activity A111).

[0142] 4. Effect The operation of this embodiment can be summarized as follows. That is, the contents of the initial job posting can be arranged to fit the new job posting format. Therefore, the employer can easily modify the job posting to fit the format of the job search service.

[0143] Although the embodiment of the present invention has been described above, the present invention is not limited to this, and can be modified as appropriate without departing from the technical concept of the invention.

[0144] 5.Other In the above embodiment, the server device 10 performs various storage and control, but multiple external devices may be used instead of the server device 10. That is, various information and programs may be distributed and stored in multiple external devices using block chain technology or the like. In particular, the artificial intelligence unit 120 may be an external configuration of the server device 10. In that case, the artificial intelligence unit 120, which is an external configuration, is configured to receive input from each functional unit of the server device 10 and return the instructed output to the server device 10.

[0145] The control unit 11 does not necessarily have to include the job posting evaluation unit 113. In other words, the server device 10 does not necessarily have to calculate the score of the initial job posting. In addition, the control unit 11 does not necessarily have to include the item addition unit 116. In other words, the server device 10 does not necessarily have to present an option to add to a new job posting.

[0146] The aspect of the present embodiment is not limited to the hiring support system 1, and may be an information processing method or a program. The hiring support method includes each step executed by the hiring support system 1. The program causes a computer to execute each step of the hiring support system 1.

[0147] It may be provided in any of the following ways:

[0148] (1) A recruiting support system comprising a processor configured to execute the following steps: an acquisition step acquires an initial job posting; a job posting regeneration step inputs a plurality of job posting components extracted from the initial job posting into a classification model, causes the classification model to output classification information for each of the job posting components, and arranges the plurality of job posting components into a job posting format having a different configuration from the initial job posting based on the classification information, wherein the job posting components include at least one word or sentence, the classification information is information indicating an item in which the job posting components are arranged in the job posting format, and the classification model is a learning model trained to be able to input the job posting components and output the classification information.

[0149] (2) In the recruitment support system described in (1) above, in the job posting regeneration step, instructions for generating the classification information for each job posting component are input into the classification model.

[0150] (3) In the recruiting support system described in (1) or (2) above, in the job posting regeneration step, the classification model is caused to output an identifier individually assigned to the job posting component and the classification information associated with the identifier, and the job posting component corresponding to the identifier associated with the classification information is placed in an item indicated by the classification information in the job posting format.

[0151] (4) In the recruitment support system described in any one of (1) to (3) above, in the job posting regeneration step, a first division process is executed to divide the contents of the initial job posting into at least one intermediate element, and a second division process is executed to divide the intermediate element into multiple job posting components.

[0152] (5) In the recruitment support system described in (4) above, the first division process divides the content into at least one of the intermediate elements based on a delimiter included in the initial job posting.

[0153] (6) In the recruitment support system described in (4) above, in the first division process, the description content is input into a first division model and the first division model is caused to output at least one of the intermediate elements, and here, the first division model is a learning model that has been trained to be able to input the description content and output a plurality of the intermediate elements based on predetermined intermediate element classification information.

[0154] (7) In the recruiting support system described in (5) above, in the second splitting process, the intermediate element is split into multiple job listing components based on a delimiter included in the initial job listing that is different from the delimiter used in the first splitting process.

[0155] (8) A recruiting support system according to any one of (4) to (6) above, wherein in the second splitting process, the intermediate element is input into a second splitting model, and the second splitting model is caused to output a plurality of the job posting components, and wherein the second splitting model is a learning model trained to be capable of inputting the intermediate element and outputting a plurality of the job posting components in which the sentences contained in the intermediate element are grouped together based on a common theme.

[0156] (9) In the recruiting support system described in any one of (1) to (8) above, the processor is configured to further execute the following steps, in which in a job posting display control step, the job posting format in which the job posting components are arranged is displayed as a new job posting, and a correspondence between the job posting components in the new job posting and the job posting components in the initial job posting is displayed.

[0157] (10) In the recruitment support system described in any one of (1) to (9) above, the processor is configured to further perform the following step, in which in the item addition step, based on the contents of the initial job posting, options to be added to specific items in the job posting format are presented, and the options are extracted from among pre-set options.

[0158] (11) In the recruitment support system described in (10) above, in the item addition step, the contents of the initial job description referenced when extracting the option to be presented are displayed as the basis for that option.

[0159] (12) In the recruitment support system described in any one of (1) to (11) above, in the acquisition step, website access information for accessing a job posting reference website including job posting access information for accessing the initial job posting is acquired, and the job posting access information that can be acquired from the website access information is presented.

[0160] (13) In the recruiting support system described in (12) above, the job posting access information includes a URL, and in the acquisition step, an access information group is created by grouping multiple pieces of job posting access information by analyzing the URL, and selection is accepted on an access information group basis, and the initial job posting is acquired from the job posting access information included in the selected access information group.

[0161] (14) In the recruitment support system described in (12) or (13) above, in the acquisition step, the initial job listing is acquired from each of the multiple job listing access information, and the processor is configured to further execute the following step, and in the list display control step, information on the multiple initial job listings is displayed in a list.

[0162] (15) In the recruitment support system described in (14) above, the processor is configured to further execute the following steps: in the job posting evaluation step, the scores of the multiple initial job postings are calculated based on reference information including a correlation between the content of the job postings and the scores of the job postings; and in the list display control step, information on the multiple initial job postings is displayed in a list together with the scores.

[0163] (16) A recruitment support method comprising steps executed by a recruitment support system described in any one of (1) to (15) above.

[0164] (17) A program that causes a computer to execute each step of the recruitment support system described in any one of (1) to (15) above. Of course, this is not the case.

[0165] Finally, although various embodiments according to the present disclosure have been described, these are presented as examples and are not intended to limit the scope of the invention. The novel embodiment can be implemented in various other forms, and various omissions, substitutions, and modifications can be made without departing from the spirit of the invention. The embodiments and their modifications are included in the scope and spirit of the invention, and are included in the scope of the invention and its equivalents described in the claims. [Explanation of symbols]

[0166] 1: Recruitment support system 2: Communication lines 10: Server device 11: Control section 12: Storage section 13: Communications Department 14: Communication bus 20: Recruiter terminal 21: Control section 22: Storage section 23: Communications Department 24: Input section 25: Output section 26: Communication bus 30: Job Seeker Terminal 31: Control section 32: Storage section 33: Communications Department 34: Input section 35: Output section 36: Communication bus 111: Basic display control unit 112: Acquisition Department 113: Job posting evaluation department 114: List display control unit 115: Job posting regeneration department 116: Item addition part 117: Job posting display control unit 120: Artificial Intelligence Department 211:Display section 212: Operation acquisition section 311: Display section 312: Operation reception section

Claims

1. A recruitment support system, A processor is provided. The processor is configured to perform the steps of: In the acquisition step, the initial job posting is acquired, In the job posting regeneration step, a plurality of job posting components extracted from the initial job posting are input to a classification model, classification information for each of the job posting components is output from the classification model, and the plurality of job posting components are arranged in a job posting format having a different configuration from the initial job posting based on the classification information; Here, the job posting component includes at least one word or sentence, The classification information is information indicating an item in which the job posting component is arranged in the job posting format, The classification model is a learning model that has been trained to be able to input the job posting components and output the classification information, In the item addition step, an option to be added to a predetermined item of the job posting format is presented based on the description of the initial job posting, and the option is a keyword related to the description of the initial job posting extracted from a preset keyword list, or the description of the initial job posting is input to an option candidate extraction model and output to the option candidate extraction model; A recruitment support system, wherein the option candidate extraction model is a learning model that is trained using data on the contents of the initial job posting and data on the corresponding options as training data.

2. The recruitment support system according to claim 1, In the job posting regeneration step, an instruction to generate the classification information for each job posting component is input to the classification model, in the recruitment support system.

3. The recruitment support system according to claim 1, In the job posting regeneration step, the classification model outputs an identifier individually assigned to the job posting component and the classification information associated with the identifier, and the job posting component corresponding to the identifier associated with the classification information is placed in the item indicated by the classification information in the job posting format, in a recruitment support system.

4. The recruitment support system according to claim 1, In the job posting regeneration step, a first division process is executed to divide the content of the initial job posting into at least one intermediate element, and a second division process is executed to divide the intermediate element into a plurality of job posting components; In the first division process, the description content is divided into at least one of the intermediate elements based on a delimiter included in the initial job posting, In the second division process, the intermediate element is divided into multiple job posting components based on a delimiter symbol contained in the initial job posting that is different from the delimiter symbol used in the first division process, in a recruitment support system.

5. The recruitment support system according to claim 1, In the job posting regeneration step, a first division process is executed to divide the content of the initial job posting into at least one intermediate element, and a second division process is executed to divide the intermediate element into a plurality of job posting components; In the first division process, the description is input to a first division model, and at least one of the intermediate elements is output from the first division model, where the first division model is a learning model that is trained to be capable of inputting the description and outputting a plurality of the intermediate elements obtained by dividing the description into categories defined by predetermined intermediate element classification information, The category defined by the intermediate element classification information is the same category as the item in the job posting format, a category that is a higher-level conceptualization of the item, or a category that is a subdivision of the item, In the second division process, the intermediate element is divided into a plurality of the job posting components based on the subject of the text, in a recruitment support system.

6. 6. The recruitment support system according to claim 5, In the second division process, the intermediate element is input to a second division model, and the second division model is made to output a plurality of the job advertisement components, and the second division model is a learning model or a large-scale language model that is trained so as to be capable of inputting the intermediate element and outputting a plurality of the job advertisement components in which the sentences included in the intermediate element are grouped together by the commonality of the theme, using the data of the intermediate element and the corresponding data of the job advertisement components in which sentences are grouped together by the commonality of the theme as training data, and When the second segmentation model is a large-scale language model, the second segmentation process inputs the intermediate elements and inputs instructions to the second segmentation model to output multiple job posting components that are organized by thematic commonality of the sentences contained in the intermediate elements.

7. The recruitment support system according to claim 1, The processor is further configured to perform the steps of: In the job posting display control step, the recruitment support system displays the job posting format in which the job posting components are arranged as a new job posting, and displays the correspondence between the job posting components in the new job posting and the job posting components in the initial job posting.

8. The recruitment support system according to claim 1, In the item addition step, the recruitment support system displays the contents of the initial job posting referenced when the option to be presented was extracted as the basis for that option.

9. The recruitment support system according to claim 1, In the acquisition step, a recruitment support system acquires website access information for accessing a job posting reference website, including job posting access information for accessing the initial job posting, and presents the job posting access information that can be acquired from the website access information.

10. The recruitment support system according to claim 9, The job posting access information includes a URL, In the acquisition step, a recruitment support system creates an access information group by grouping multiple job posting access information by analyzing the URL, accepts selection on an access information group basis, and acquires the initial job posting from the job posting access information included in the selected access information group.

11. The recruitment support system according to claim 9, In the obtaining step, the initial job posting is obtained from each of the plurality of job posting access information, The processor is further configured to perform the steps of: In the list display control step, information on a plurality of the initial job postings is displayed in a list, in the recruitment support system.

12. The recruitment support system according to claim 11, The processor is further configured to perform the steps of: In the job posting evaluation step, the scores of the initial job postings are calculated based on reference information including a correlation between the contents of the job postings and the scores of the job postings; In the list display control step, information on a plurality of the initial job postings is displayed in a list together with the scores, in a recruitment support system.

13. A recruitment support method, comprising: A recruitment support method comprising the steps executed by the recruitment support system according to any one of claims 1 to 12.

14. A program, A program causing a computer to execute each step of the recruitment support system according to any one of claims 1 to 12.

Images