A talent recommendation, candidate evidence collection and recruitment report generation method, device and equipment across recruitment platforms and a storage medium

Abstract

This invention discloses a method, apparatus, device, and storage medium for cross-recruitment platform talent recommendation, candidate evidence collection, and recruitment report generation. The method receives job input data and generates a job scorecard and a cross-platform search strategy; it collects initial candidate records through a recruitment platform adapter, opening the details page to obtain evidence such as homepage address, platform identifier, resume number, text, files, and screenshots; it performs name matching, page type identification, number matching, non-blank detection, and loading integrity detection on the evidence; it generates candidate scores, risk points, and reports based on the scorecard, and performs human-machine confirmation control according to the action risk level. This solution improves cross-platform data processing consistency, evidence reliability, and report generation efficiency, while reducing the risk of accidental triggering of external actions.

Description

Technical Field

[0001] This invention relates to the field of computer data processing technology, and in particular to a method, apparatus, equipment, and storage medium for talent recommendation, candidate evidence collection, and recruitment report generation across recruitment platforms. Background Technology

[0002] When recruiters conduct talent searches across regions, job positions, or platforms, they typically need to switch between multiple recruitment platforms, social networks, enterprise ATS systems, resume databases, and spreadsheet files. The field names, candidate identifiers, personal profile structures, resume numbers, attachment formats, export paths, and message statuses of candidates vary across different platforms.

[0003] Existing recruitment tools often focus on keyword searches or single-platform candidate management, making it difficult to convert arbitrary job input into dynamic scorecards and to stably link candidate profile URLs, resume numbers, resume PDFs, screenshot evidence, and recommendation reasons. Especially when generating candidate screening reports, manual screenshotting and copying of information can easily lead to problems such as page errors, candidate mismatches, blank screenshots, missing evidence, and untraceable recommendation conclusions.

[0004] Furthermore, actions on recruitment platforms such as sending messages, saving candidates, posting jobs, obtaining contact information, consuming platform credit, and changing candidate status have external effects. If automated tools do not distinguish between read-only data collection actions and external effect actions, there is a risk of accidentally sending, saving, posting, or consuming credit.

[0005] Therefore, there is a need for a method that can accept input from any job position, collect candidate evidence across recruitment platforms, verify the quality of the evidence, generate interpretable candidate recommendation results and recruitment reports, and perform human-machine confirmation control on external effect actions. Summary of the Invention

[0006] Technical problems to be solved The purpose of this invention is to provide a method for talent recommendation, candidate evidence collection, and recruitment report generation across recruitment platforms, in order to solve the problems of inconsistent candidate information structure across platforms, easy mismatch in evidence collection, lack of evidence chain in report generation, uninterpretable candidate scores, and easy accidental triggering of external actions. Technical solution

[0007] To achieve the above objectives, this invention provides a method for talent recommendation, candidate evidence collection, and recruitment report generation across recruitment platforms, comprising: Receive job-related input data; Perform semantic parsing on the job input data to obtain at least one of the following: job responsibilities, job requirements, hard conditions, priority conditions, exclusion conditions, target industry, target company, location, language, tools, products, customer type, and work style; Generate job evaluation cards based on semantic parsing results; Generate cross-platform search strategies based on job performance cards; Access the recruitment platform and collect initial candidate records through at least one recruitment platform adapter; Open the candidate details page based on the candidate's initial record and collect candidate evidence data; Perform quality checks on the candidate's evidence data; Candidate scoring results are generated based on job scorecards and candidate evidence data that has passed quality verification. A recruitment report is generated based on candidate ratings and candidate evidence data. Human-machine confirmation control is implemented based on the risk level of the action. Beneficial effects

[0008] Compared with the prior art, the present invention has at least the following beneficial effects: 1. Able to convert any job input into a dynamic job scorecard, improving job suitability. 2. Able to unify candidate and evidence fields across different recruitment platforms through a recruitment platform adapter. 3. Able to bind candidate recommendation conclusions with profile URLs, platform candidate IDs, resume numbers, resume files, and screenshot evidence, improving traceability. 4. Able to improve the quality of evidence collection through candidate name matching, page type recognition, and non-blank screenshot detection. 5. Able to automatically generate candidate recommendation reports while maintaining the sequential binding of the recommendation reason page and the complete evidence page. 6. Able to reduce the risks of accidental sending, saving, publishing, and consumption of credit limits through action risk levels and human-machine confirmation controls. Attached Figure Description

[0009] Figure 1 is a schematic diagram of the method flow provided in an embodiment of the present invention.

[0010] Figure 2 is a schematic diagram of the system module structure provided in an embodiment of the present invention.

[0011] Figure 3 is a schematic diagram of the candidate evidence chain data structure provided in an embodiment of the present invention.

[0012] Figure 4 is a schematic diagram of the action risk level control process provided in an embodiment of the present invention. Detailed Implementation

[0013] The present invention will be described below with reference to embodiments. It should be understood that the following embodiments are for illustrative purposes only and are not intended to limit the scope of protection of the present invention.

[0014] Example 1: Job Input Analysis and Scoring Card Generation The electronic device receives job input data. Job input data can be natural language job descriptions, job profiles, recruitment requirements, target company lists, target regions, target language requirements, or table fields.

[0015] Electronic devices perform semantic analysis on job input data, extracting job responsibilities, job requirements, hard conditions, priority conditions, exclusion conditions, target industry, target company, location, language, tools, products, customer type, and work style.

[0016] Electronic devices generate job performance evaluation cards based on semantic parsing results. Each job performance evaluation card includes multiple evaluation dimensions, dimension weights, evidence fields, hard thresholds, lenient conditions, and validation questions. Evaluation dimensions may include core job matching, industry or product matching, tool or technical skills, qualifications and results, communication language and stakeholder matching, location and work style matching, etc.

[0017] When the job input data matches a preset job family, the electronic device calls the scoring dimensions and weights of the corresponding job family; when the job input data does not match a preset job family, the electronic device dynamically generates a scorecard based on the semantic parsing results.

[0018] Example 2: Cross-platform search strategy generation Electronic devices generate cross-platform search strategies based on job performance evaluation cards. These strategies include target job title groups, adjacent job title groups, target company groups, target industry groups, keyword groups, Boolean queries, search waves, and platform filtering conditions.

[0019] For example, search waves can include exact match waves, target company waves, adjacent job title waves, competency keyword waves, and applicant waves. Each search wave is recorded as a traceable search source.

[0020] Example 3: Recruitment Platform Adapter Electronic devices access the recruitment platform through a recruitment platform adapter. The recruitment platform adapter is used to identify platform login status, search entry points, candidate card fields, candidate details page addresses, candidate stability indicators, resume numbers, resume export controls, screenshot areas, message controls, save controls, quota prompts, and status change controls.

[0021] The recruitment platform adapter maps candidate homepage addresses, platform candidate identifiers, resume numbers, application numbers, export file paths, and screenshot paths from different recruitment platforms to a unified candidate record.

[0022] Example 4: Candidate Evidence Collection Electronic devices open the candidate details page based on the candidate's initial record. The candidate details page is preferably the candidate's personal homepage, resume details page, or application details page, rather than a search results page, recommendation list page, or applicant list page.

[0023] The electronic device expands the read-only collapsed area in the candidate details page, scrolls the page to trigger lazy loading of content, and collects text content and screenshots from the candidate details page. The collected content may include the candidate's homepage address, platform candidate identifier, resume number, current position, current company, location, work experience, education, skills, languages, certificates, attachment metadata, resume file path, and screenshot path.

[0024] For recruitment platforms that support resume export, electronic devices can generate resume PDFs by reading-only printing or exporting the path, and then bind the resume PDF path to the candidate record.

[0025] Example 5: Quality Verification Electronic devices perform quality checks on candidate evidence data. Quality checks may include candidate name matching, page type identification, resume number matching, screenshot non-blank detection, and content loading integrity detection.

[0026] Page type identification is used to determine whether the current page is a candidate details page, search results page, recommendation list page, applicant list page, or a blank page. When the current page is not a candidate details page, the electronic device stops writing that page as a candidate evidence page into the recruitment report.

[0027] Screenshot non-blank detection can be calculated based on the valid pixel area, text area, or structured content area. When the valid pixel area, text area, or structured content area is lower than a preset threshold, the electronic device marks the screenshot as invalid.

[0028] Example 6: Candidate Deduplication Electronic devices deduplicate candidates from multiple recruitment platforms. Deduplication is based on at least one of the following: candidate homepage address, platform candidate identifier, resume number, combination of name and current company, contact information hash value, resume file fingerprint, and historical communication records.

[0029] After deduplication, the electronic device merges evidence from different sources of duplicate candidates into the same candidate record.

[0030] Example 7: Candidate Scoring and Recommendation The electronic device generates candidate ratings based on the job scorecard and candidate evidence data that has passed quality verification. The candidate ratings include a matching score, matching tags, reasons for recommendation, risk points, missing verification items, and suggested actions.

[0031] In one embodiment, the candidate evaluation results also include a recent job relevance score. The recent job relevance score is generated based on the degree of match between the candidate's current position or most recent substantive work experience and the core responsibilities, technical scope, product scope, customer type, or capability requirements in the job scorecard.

[0032] Example 8: Recruitment Report Generation The electronic device generates a recruitment report based on candidate ratings and candidate evidence data. The recruitment report can be in PPT, PDF, DOCX, HTML, or spreadsheet format.

[0033] The recruitment report should include at least the following pages: a job evaluation card page, a talent map coverage page, a candidate summary page, a candidate recommendation reason page, and a candidate evidence page. Following each candidate's recommendation reason page should be a candidate evidence page, which may include a screenshot of the candidate's homepage, a screenshot of their resume, a PDF version of their resume, or a combination thereof.

[0034] Before generating the report, the electronic device performs screenshot quality verification, deletes blank or nearly blank pages, and ensures that the candidate's recommendation reason page and the candidate's evidence page are bound in sequence.

[0035] Example 9: Motion Risk Level and Human-Machine Verification Control Electronic devices perform human-machine verification controls based on the risk level of the action. The risk level of the action includes read-only actions, evidence delivery actions, and external effect actions.

[0036] Read-only actions include opening a page, scrolling the page, expanding read-only content, and reading text visible to candidates. Evidence delivery actions include taking screenshots of candidates, exporting resumes, and generating reports. External effect actions include sending messages, saving candidates, posting job openings, changing candidate status, obtaining contact information, consuming platform credit, and submitting forms.

[0037] When performing human-machine confirmation control on external effect actions, the electronic device outputs the action object, action content, target candidate or position, platform name, and possible external effects before execution; receives user confirmation instructions; and executes external effect actions only after receiving user confirmation instructions. Industrial applicability

[0038] This invention can be applied to recruitment platforms, enterprise recruitment systems, recruitment agents, candidate management systems, and recruitment report generation systems. It can improve the efficiency and accuracy of cross-platform candidate evidence collection, recommendation scoring, and report generation, and has industrial applicability.

Claims

1. A method for talent recommendation, candidate evidence collection, and recruitment report generation across recruitment platforms, characterized in that, The method, executed by an electronic device, includes: receiving job input data, the job input data including job description, job profile, recruitment requirements, target company, target region, language requirements, hard conditions, or a combination thereof; performing semantic parsing on the job input data to obtain at least one of job responsibilities, job requirements, hard conditions, priority conditions, exclusion conditions, target industry, target company, location, language, tools, products, customer type, and work style; generating a job rating card based on the semantic parsing results, the job rating card including multiple rating dimensions, dimension weights, evidence fields, hard thresholds, relaxed conditions, and verification questions; generating a cross-platform search strategy based on the job rating card, the cross-platform search strategy including at least one of target title groups, adjacent title groups, target company groups, keyword groups, Boolean search expression, search wave, and platform filtering conditions; accessing a recruitment platform and collecting candidate initial records through at least one recruitment platform adapter, the recruitment platform adapter being used to identify at least one of candidate details page address, candidate stability identifier, resume number, resume export control, screenshot area, message control, save control, quota prompt, and status change control; and based on the candidate initial records... Open the candidate details page and collect candidate evidence data, which includes at least one of the following: candidate homepage address, platform candidate identifier, resume number, current position, current company, experience text, skills text, language information, resume file path, and screenshot path. Perform quality verification on the candidate evidence data, which includes at least one of the following: candidate name matching, page type identification, resume number matching, screenshot non-blank detection, and content loading integrity detection. Generate a candidate rating result based on the job scoring card and the candidate evidence data that has passed the quality verification, which includes at least one of the following: matching score, matching tag, recommendation reason, risk point, missing verification item, and suggested action. Generate a recruitment report based on the candidate rating result and candidate evidence data, wherein the recommendation reason page for each candidate is bound to the candidate evidence page in a preset order. Perform human-machine confirmation control according to the action risk level, wherein read-only collection actions and actions that will produce external effects are processed separately. Actions that will produce external effects include at least sending messages, saving candidates, posting jobs, changing candidate status, obtaining contact information, consuming platform quota, or submitting forms.

2. The method according to claim 1, characterized in that, The generation of job scorecards based on semantic parsing results includes: when the job input data matches a preset job family, calling the scoring dimensions and weights corresponding to the preset job family; when the job input data does not match a preset job family, dynamically generating scoring dimensions and dimension weights based on job responsibilities, job requirements, industry, product, customer type, tools, language, and location.

3. The method according to claim 1, characterized in that, The recruitment platform adapter includes a field mapping module, which maps candidate homepage addresses, platform candidate identifiers, resume numbers, application numbers, export file paths, and screenshot paths from different recruitment platforms to a unified candidate record.

4. The method according to claim 1, characterized in that, The process of collecting candidate evidence data includes: opening the candidate details page; expanding the read-only collapsed area in the candidate details page; scrolling the candidate details page to trigger lazy loading of content; collecting text content and page screenshots from the candidate details page; and binding the text content and page screenshots to the candidate stability identifier.

5. The method according to claim 1, characterized in that, The page type identification includes determining whether the current page is a candidate details page, search results page, recommendation list page, applicant list page, or blank page; when the current page is not a candidate details page, the page is stopped from being written into the recruitment report as a candidate evidence page.

6. The method according to claim 1, characterized in that, The screenshot non-blank detection includes calculating the effective pixel area, text area, or structured content area in the screenshot. When the effective pixel area, text area, or structured content area is lower than a preset threshold, the screenshot is marked as an invalid screenshot.

7. The method according to claim 1, characterized in that, The method also includes deduplicating candidates from multiple recruitment platforms, based on at least one of the following: candidate homepage address, platform candidate identifier, resume number, combination of name and current company, contact information hash value, resume file fingerprint, and historical communication records.

8. The method according to claim 1, characterized in that, Generating a recruitment report includes: generating a job rating card page; generating a talent map overlay page; generating a candidate summary page; generating a recommendation reason page for each candidate; and inserting a candidate evidence page after each candidate's recommendation reason page, wherein the candidate evidence page includes a screenshot of the candidate's homepage, a screenshot of the resume, a resume PDF page, or a combination thereof.

9. The method according to claim 1, characterized in that, Action risk levels include read-only actions, evidence delivery actions, and external effect actions; read-only actions include opening a page, scrolling a page, expanding read-only content, and reading candidate visible text; evidence delivery actions include taking screenshots of candidates, exporting resumes, and generating reports; external effect actions include sending messages, saving candidates, posting job openings, changing candidate status, obtaining contact information, consuming platform credit, and submitting forms.

10. The method according to claim 9, characterized in that, Human-machine confirmation control for external effect actions includes: outputting the action object, action content, target candidate or position, platform name, and possible external effects before executing the external effect action; receiving user confirmation instructions; and executing the external effect action only after receiving the user confirmation instructions.

11. The method according to claim 1, characterized in that, The candidate scoring results also include a recent job relevance score, which is generated based on the degree of matching between the candidate's current position or most recent substantive work experience and the core responsibilities, technical scope, product scope, customer type, or ability requirements in the job scorecard.

12. A device for talent recommendation, candidate evidence collection, and recruitment report generation across recruitment platforms, characterized in that, include: The job analysis module is used to receive and parse job input data. The scorecard generation module is used to generate job scorecards; the search strategy generation module is used to generate cross-platform search strategies. The platform adaptation module is used to access the recruitment platform and collect initial candidate records through the recruitment platform adapter; The evidence collection module is used to open the candidate details page and collect candidate evidence data; The quality verification module is used to perform quality verification on candidate evidence data; The rating and recommendation module is used to generate candidate rating results; The report generation module is used to generate recruitment reports; The motion control module is used to perform human-machine confirmation control based on the motion risk level.

13. An electronic device, characterized in that, It includes a processor and a memory, wherein the memory stores a computer program that, when executed by the processor, implements the method of any one of claims 1 to 11.

14. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, implements the method of any one of claims 1 to 11.

Images