Recruitment Methods and Systems Based on Dynamic Skill Loading

By constructing a dynamic skill load set and dynamically adjusting the priority and level of skill requirements, the problem that job skill requirements cannot reflect changes in enterprises in a timely manner in existing recruitment methods is solved, thereby improving the accuracy and adaptability of recruitment matching.

CN123118052APending Publication Date: 2026-11-17GUANGDONG JUNRUN DIGITAL INTELLIGENCE TECH CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202611438066.8
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-09-16
Publication Date
2026-11-17

AI Technical Summary

Technical Problem

Existing recruitment methods match candidates to static job skill requirements, which means that job skill requirements cannot reflect changes in corporate recruitment needs in a timely manner, resulting in low accuracy in matching candidates with jobs.

Method used

By acquiring job requirement data, enterprise recruitment data, candidate resume data, historical recruitment data, and recruitment feedback data, a standard skills set is constructed. Combined with enterprise demand intensity indicators, historical recruitment performance indicators, and recruitment cycle characteristic indicators, a dynamic skills load set is generated. The priority and level of skills requirements are dynamically adjusted to generate a target job skills requirement structure corresponding to the current recruitment environment, and it is continuously updated based on recruitment feedback data.

Benefits of technology

It improves the accuracy of recruitment matching between candidates and positions, enhances the adaptability of recruitment methods to different recruitment stages and environmental changes, and ensures that the structure of job skill requirements can reflect the actual changes in the current recruitment environment in a timely manner.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN123118052A_ABST
    Figure CN123118052A_ABST
Patent Text Reader

Abstract

This application pertains to the field of data processing technology and provides a recruitment method and system based on dynamic skill loads. The method includes: acquiring job requirement data, enterprise recruitment data, candidate resume data, historical recruitment data, and recruitment feedback data corresponding to the target job; generating a skill requirement structure for the target job and a dynamic skill load set corresponding to the target job; updating the dynamic skill load set; and readjusting the skill requirement structure for the target job based on the updated dynamic skill load set. This approach enables recruitment strategies to be continuously iterated and optimized throughout the recruitment process, improving the accuracy of job matching between candidates and target positions. It overcomes the problem that the job skill requirement structure cannot reflect the actual changes in the current recruitment environment in a timely manner, leading to low accuracy in job matching between candidates and positions.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention relates to the field of data processing technology, and in particular to a recruitment method and system based on dynamic skill load. Background Technology

[0002] With the development of digital and intelligent recruitment technologies, recruitment platforms, corporate recruitment management systems, and headhunting agencies are increasingly adopting technologies such as big data analytics, natural language processing, and artificial intelligence to analyze and process job requirements, candidate resumes, and historical recruitment information. This enables intelligent candidate retrieval, job recommendations, and job matching, improving recruitment efficiency and talent selection capabilities. Currently, recruitment systems typically perform matching calculations based on the skill requirements in job descriptions and candidate skill information, and combine this with historical recruitment data to assess candidate suitability, thereby assisting recruiters in candidate screening and recommendation.

[0003] However, existing recruitment methods are usually based on static job skill requirements or fixed weights for skill matching. Because they lack the ability to comprehensively represent the dynamic evolution of skill requirements as companies recruit, historical hiring results and recruitment cycles change, the job skill requirement structure cannot reflect the actual changes in the current recruitment environment in a timely manner. This results in low accuracy of recruitment matching between candidates and positions and makes it difficult to continuously adapt to the talent screening needs of different recruitment stages and environments. Summary of the Invention

[0004] This invention provides a recruitment method and system based on dynamic skill load to solve the problem that the structure of job skill requirements cannot reflect the actual changes in the current recruitment environment in a timely manner, which leads to low accuracy in recruitment matching between candidates and positions.

[0005] In a first aspect, embodiments of this application provide a recruitment method based on dynamic skill loading, comprising: acquiring job requirement data, enterprise recruitment data, candidate resume data, historical recruitment data, and recruitment feedback data corresponding to a target recruitment position; extracting job skill items and candidate skill items based on the job requirement data and candidate resume data, standardizing and mapping the job skill items and candidate skill items, and constructing a standard skill set corresponding to the target recruitment position based on the mapping results; statistically analyzing multiple job posting records corresponding to each skill item in the standard skill set in the enterprise recruitment data to calculate the enterprise demand intensity index corresponding to the skill item; statistically analyzing historical recruitment performance indicators and recruitment cycle characteristic indicators corresponding to historical recruited candidates and historical recruitment processes in the historical recruitment data, based on the enterprise demand intensity index and the recruitment feedback data; and constructing a standard skill set corresponding to the target recruitment position based on the enterprise demand intensity index and the recruitment feedback data based on the recruitment feedback data. The system generates a dynamic skill load set corresponding to the target recruitment position based on the industry demand intensity index, the historical recruitment performance index, and the recruitment cycle characteristic index. Based on the dynamic skill load set, the skill requirements of each skill item in the standard skill set are reconstructed. The priority and level of the skill requirements for each skill item are dynamically adjusted according to changes in the dynamic skill load value, thereby generating a target position skill requirement structure corresponding to the current recruitment environment. The system determines the corresponding skill fit degree based on the correspondence between the target position skill requirement structure and the candidate's skill items, and generates a corresponding candidate recruitment matching result based on the skill fit degree. The system updates the dynamic skill load set using recruitment feedback data, and readjusts the target position skill requirement structure based on the updated dynamic skill load set.

[0006] Secondly, this application provides a recruitment system based on dynamic skill load, including: a data acquisition module for acquiring job requirement data, enterprise recruitment data, candidate resume data, historical recruitment data, and recruitment feedback data corresponding to the target recruitment position; a data extraction module for extracting job skill items and candidate skill items based on the job requirement data and candidate resume data, standardizing and mapping the job skill items and candidate skill items, and constructing a standard skill set corresponding to the target recruitment position based on the mapping results; and an indicator acquisition module for statistically analyzing multiple job posting records corresponding to each skill item in the standard skill set in the enterprise recruitment data to calculate the enterprise demand intensity indicator corresponding to the skill item, statistically analyzing historical recruitment performance indicators and recruitment cycle characteristic indicators corresponding to historical recruited candidates and historical recruitment processes in the historical recruitment data, and based on the enterprise demand intensity indicator, calculating the enterprise demand intensity indicator corresponding to the skill item. The system generates a dynamic skill load set corresponding to the target recruitment position based on industry demand intensity indicators, historical recruitment performance indicators, and recruitment cycle characteristic indicators. A demand reconstruction module reconstructs the skill requirements for each skill item in the standard skill set based on the dynamic skill load set, dynamically adjusting the skill requirement priority and skill level of each skill item according to changes in the dynamic skill load value, to generate a target position skill requirement structure corresponding to the current recruitment environment. A relationship confirmation module determines the corresponding skill fit degree based on the correspondence between the target position skill requirement structure and the candidate's skill items, and generates a corresponding candidate recruitment matching result based on the skill fit degree. A structure adjustment module updates the dynamic skill load set using recruitment feedback data and readjusts the target position skill requirement structure based on the updated dynamic skill load set.

[0007] One approach provided by the aforementioned method and system involves acquiring job requirement data, company recruitment data, candidate resume data, historical recruitment data, and recruitment feedback data for the target job. A standard skillset corresponding to the target job is constructed, and a dynamic skillset is generated by combining company demand intensity indicators, historical hiring performance indicators, and recruitment cycle characteristic indicators. This allows each skill item to dynamically reflect its overall demand status based on changes in the current recruitment environment. Furthermore, the standard skillset is reconstructed using the dynamic skillset, dynamically adjusting the priority and level of each skill item to form a target job skillset structure corresponding to the current recruitment environment. The dynamic skillset is continuously updated and the target job skillset structure is readjusted based on recruitment feedback data, enabling the recruitment strategy to continuously iterate and optimize throughout the recruitment process. This improves the accuracy of matching candidates with target jobs and enhances the adaptability of the recruitment method to different recruitment stages and environmental changes. This overcomes the problem of low accuracy in matching candidates with jobs due to the job skillset structure failing to reflect the actual changes in the current recruitment environment. Attached Figure Description

[0008] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the description of the embodiments of the present invention will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0009] Figure 1 This is a schematic diagram of the operating environment in one embodiment of the present invention; Figure 2 This is a flowchart illustrating a recruitment method based on dynamic skill loading in one embodiment of the present invention; Figure 3 This is a schematic diagram of a sub-process of step S10 in one embodiment of the present invention; Figure 4 This is a schematic diagram of enterprise recruitment data in one embodiment of the present invention; Figure 5 This is a schematic diagram of a sub-process of step S20 in one embodiment of the present invention; Figure 6 This is a schematic diagram of the scoring information of candidate resume data in one embodiment of the present invention; Figure 7 This is a schematic diagram of a sub-process of step S30 in one embodiment of the present invention; Figure 8 This is a schematic diagram of a sub-process of step S40 in one embodiment of the present invention; Figure 9This is a schematic diagram of a sub-process of step S50 in one embodiment of the present invention; Figure 10 This is a schematic diagram of a sub-process of step S60 in one embodiment of the present invention; Figure 11 This is a flowchart illustrating a recruitment system based on dynamic skill loading in one embodiment of the present invention. Detailed Implementation

[0010] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0011] It should be understood that, when used in this specification and the appended claims, the term "comprising" indicates the presence of the described features, integrals, steps, operations, elements, and / or components, but does not exclude the presence or addition of one or more other features, integrals, steps, operations, elements, components, and / or collections thereof. It should also be understood that, as used in this specification and the appended claims, the term "and / or" refers to any combination of one or more of the associated listed items and all possible combinations, and includes such combinations.

[0012] Furthermore, in the description of this invention and the appended claims, the terms "first," "second," "third," etc., are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0013] References to "one embodiment" or "some embodiments" as described in this specification mean that one or more embodiments of the invention include a specific feature, structure, or characteristic described in connection with that embodiment. Therefore, the phrases "in one embodiment," "in some embodiments," "in other embodiments," "in still other embodiments," etc., appearing in different parts of this specification do not necessarily refer to the same embodiment, but rather mean "one or more, but not all, embodiments," unless otherwise specifically emphasized. The terms "comprising," "including," "having," and variations thereof mean "including but not limited to," unless otherwise specifically emphasized.

[0014] It should be understood that the sequence number of each step in the following embodiments does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0015] To illustrate the technical solution of the present invention, specific embodiments are described below.

[0016] Current recruitment platforms generally match jobs based on job descriptions and candidate resumes, and then recommend candidates using keyword searches, rule-based filtering, or fixed-weight scoring methods. However, the demand for different skills changes continuously at different recruitment stages, in different companies, and for different positions. Traditional recruitment methods still use static skill weights for matching calculations, which means that job skill requirements cannot reflect the actual changes in company recruitment in a timely manner. This can easily lead to problems such as unreasonable priority settings for job skills and inconsistencies between candidate skill evaluations and actual recruitment needs.

[0017] To address the aforementioned issues, this application proposes a recruitment method, intelligent agent, and electronic device based on dynamic skill loads. By acquiring job requirement data, enterprise recruitment data, candidate resume data, historical recruitment data, and recruitment feedback data, a standard skill set is constructed. This is then combined with enterprise demand intensity indicators, historical hiring performance indicators, and recruitment cycle characteristic indicators to generate a dynamic skill load set. Based on this dynamic skill load set, the skill requirement structure of the target job is dynamically reconstructed, and candidate skill items are matched hierarchically. The dynamic skill load set is continuously updated based on recruitment feedback data, enabling the skill requirement structure of the target job to dynamically adjust with changes in the recruitment environment, thereby improving the expressive power of job skill requirements and the accuracy of matching candidates with jobs.

[0018] The recruitment method based on dynamic skill loading provided in this embodiment of the invention can be applied to, for example... Figure 1 The operating environment shown includes a server and terminal devices. The terminal devices communicate with the server via a network. The terminal devices are used to submit job posting information, candidate resumes, and recruitment feedback information; the server is used to complete processing steps such as job skill analysis, dynamic skill load calculation, generation of target job skill requirement structure, and calculation of candidate recruitment matching results.

[0019] In practical applications, after a company publishes a target job posting, the server retrieves the corresponding job requirement data and combines it with the company's recruitment data, candidate resume data, historical recruitment data, and recruitment feedback data to establish a dataset corresponding to the target job posting. Then, it extracts job skill items and candidate skill items to construct a standard skill set and calculates the dynamic skill load value for each skill item based on the standard skill set. Based on the dynamic skill load set, it adjusts the priority and level of skill requirements to generate the skill requirement structure for the target job. Based on the skill requirement structure, it calculates the skill fit between candidates and the target job posting, generating candidate recruitment matching results. When new recruitment feedback data is generated, the dynamic skill load set is updated, and the skill requirement structure for the target job is readjusted, enabling continuous iteration of the recruitment strategy.

[0020] The terminal device can be a personal computer, laptop, smartphone, tablet, or other computing device capable of processing recruitment business.

[0021] The server can be implemented using a single server or a computing cluster consisting of multiple servers, with data interaction and task scheduling between the servers completed through a network.

[0022] In one embodiment, such as Figure 2 As shown, a recruitment method based on dynamic skill loading is provided, and its application in the above-mentioned operating environment is illustrated by example. The method includes the following steps: S10: Obtain job requirement data, company recruitment data, candidate resume data, historical recruitment data, and recruitment feedback data for the target job posting.

[0023] S20: Extract job skill items and candidate skill items based on job requirement data and candidate resume data, standardize and map the job skill items and candidate skill items, and construct a standard skill set corresponding to the target recruitment position based on the mapping results.

[0024] S30: For each skill item in the statistical standard skill set, calculate the enterprise demand intensity index corresponding to that skill item by checking multiple job posting records in the enterprise recruitment data. Calculate the historical recruitment performance index and recruitment cycle characteristic index corresponding to historical candidates and historical recruitment processes in the historical recruitment data. Based on the enterprise demand intensity index, historical recruitment performance index, and recruitment cycle characteristic index, generate a dynamic skill load set corresponding to the target recruitment position.

[0025] S40: Based on the dynamic skill load set, the skill requirements of each skill item in the standard skill set are reconstructed. The priority and level of the skill requirements corresponding to each skill item are dynamically adjusted according to the changes in the dynamic skill load value of each skill item, so as to generate the target job skill requirement structure corresponding to the current recruitment environment.

[0026] S50: Determine the corresponding skill fit degree based on the correspondence between the skill requirement structure of the target position and the skill items of the candidates, and generate the corresponding candidate recruitment matching results based on the skill fit degree.

[0027] S60: Update the dynamic skill load set using recruitment feedback data, and readjust the skill requirement structure for the target position based on the updated dynamic skill load set.

[0028] In this embodiment, a standard skill set is constructed, and a dynamic skill load set is generated by combining enterprise recruitment data and historical recruitment data. This allows the demand for each skill item to be dynamically updated as the recruitment environment changes. By dynamically adjusting the priority and level of skill requirements based on the dynamic skill load set to generate the skill requirement structure for the target position, and by continuously revising the dynamic skill load set based on recruitment feedback data, the skill requirement structure for the target position can continuously reflect the actual recruitment needs of the enterprise, thereby improving the accuracy of the skill fit calculation between candidates and target recruitment positions and the reliability of recruitment matching results.

[0029] In one embodiment, such as Figure 3 and Figure 4 As shown, step S10 involves obtaining job requirement data, company recruitment data, candidate resume data, historical recruitment data, and recruitment feedback data for the target job posting. This specifically includes the following steps: S11: Based on the intelligent agent runtime environment, create a recruitment session instance corresponding to the target recruitment position, and maintain the session state of the target recruitment position throughout the recruitment cycle through the recruitment session instance; through the recruitment session instance, call the preset job parsing intelligent agent to connect to the job posting interface of the enterprise recruitment management system to obtain the original job description text corresponding to the target recruitment position, perform structured parsing on the original job description text, and extract job responsibility information, qualification information and salary and benefits information to generate job requirement data.

[0030] When a company creates a target job posting, a unique recruitment session identifier is generated for that posting, and a corresponding recruitment session instance is established. The recruitment session instance is used to associate data objects generated throughout the entire recruitment cycle for the target job posting. The job parsing agent reads the original job description text corresponding to the target job posting and uses existing Natural Language Processing algorithms such as Named Entity Recognition (NER) and rule template parsing to parse the fields of the original job description text. Based on preset field templates, it extracts job responsibilities, qualifications, and compensation and benefits information, forming a unified structure of job requirement data.

[0031] The preset field template consists of fields for job responsibilities, education, major, years of work experience, skill requirements, qualification certificates, work location, salary range, and benefits. Each field corresponds to a unified data type and field name, thus ensuring that job description texts from different sources have a consistent data structure.

[0032] For example, the original job description text contains content such as "Responsible for Java backend development, more than three years of work experience, familiar with Spring Boot, MySQL and Redis, bachelor's degree or above". The job parsing agent recognizes "Java", "Spring Boot", "MySQL" and "Redis" as skill descriptions, "more than three years" as work experience requirement, and "bachelor's degree or above" as education requirement. It then generates the corresponding job requirement data according to the preset field template and saves the job requirement data to the recruitment session instance by associating it with the recruitment session identifier.

[0033] S12: Through the recruitment session instance, call the preset enterprise background analysis agent to connect with the enterprise credit database, enterprise official website information source and enterprise historical recruitment record database to obtain enterprise size information, industry status information, recruitment urgency information and hiring preference information corresponding to the enterprise to which the target recruitment position belongs, so as to generate enterprise recruitment data.

[0034] The enterprise background analysis agent obtains enterprise background information from multiple data sources based on the enterprise's unified social credit code or enterprise identifier, and integrates it according to unified fields. Enterprise size information includes the number of employees, the number of branches, and registered capital; industry position information includes the industry category, market ranking, or industry level; recruitment urgency information is calculated based on the job opening duration, planned recruitment number, and remaining recruitment number; and hiring preference information is obtained through statistical analysis of the educational level, years of work experience, and skill combinations of recruited personnel in historical recruitment records.

[0035] The recruitment urgency information can be categorized according to preset rules. When the job completion rate is lower than the preset percentage and the job is open for more than the preset number of days, the recruitment urgency will be marked as high; otherwise, it will be marked as medium or low.

[0036] For example, if a company's historical data shows that it plans to recruit 10 people for a target position, has currently hired 2 people, and the position has been posted for 25 days, then the company's background analysis agent calculates the recruitment completion rate as 20%. It generates a high recruitment urgency indicator based on the job posting time and statistics show that more than 80% of the employees hired in the past three years have distributed development experience, thereby generating corresponding company recruitment data.

[0037] S13: Through the recruitment session instance, the preset search strategy agent and resume screening agent are invoked to search and recall candidates in the candidate resume database and professional social platform based on the job requirements data. The candidate's educational background information, work experience information, project experience information and skill certificate information are obtained. The recalled candidate information is deduplicated and the field format is standardized to generate candidate resume data.

[0038] The search strategy agent generates candidate search criteria based on job responsibilities, qualifications, and skills requirements in the job demand data, and performs candidate searches in the candidate resume database and professional social networking platform; the resume screening agent reads the educational background, work experience, project experience, and skills certificate information of the recalled candidates, and converts them into a format according to a unified field template.

[0039] The deduplication process uses a combination of ID card number, mobile phone number, email address, and name and work experience for multi-field joint matching. For candidates that are successfully matched, only the data record with the highest information completeness is retained.

[0040] For example, if the same candidate has resume information in both the company's talent pool and professional social networking platforms, the system will verify the candidate's identity based on their mobile phone number and email address, and retain a candidate resume data containing complete project experience and skills certificate information.

[0041] S14: By calling the preset context assembly module through the recruitment session instance, retrieve historical recruitment session data corresponding to the same or similar positions as the target recruitment position from the historical session records stored in state persistence, and extract historical candidate recommendation records, historical interview evaluation records and historical recruitment result records from the historical recruitment session data to generate historical recruitment data.

[0042] The context assembly module calculates job similarity based on job title, job category, and job skills, and retrieves historical recruitment conversation data that meets a preset similarity threshold. Job similarity is calculated using existing text vector similarity algorithms, such as the Sentence-BERT text vector algorithm combined with cosine similarity.

[0043] When the job similarity reaches the preset similarity threshold, historical candidate recommendation records, historical interview evaluation records, and historical hiring result records are extracted from the corresponding historical recruitment session data and organized into historical recruitment data.

[0044] For example, if the current job posting is "Senior Java Development Engineer", the context assembly module retrieves recruitment session data for historical job postings such as "Java R&D Engineer" and "Backend Development Engineer". The job similarity scores are 0.91 and 0.88, respectively, which are higher than the preset similarity threshold of 0.85. Therefore, the corresponding historical recruitment records are extracted to form historical recruitment data.

[0045] S15: Through the recruitment session instance, the preset telephone screening agent, interview coaching agent, and onboarding follow-up agent are sequentially invoked to obtain feedback information from the candidate's telephone communication, the interviewer's evaluation feedback information on the candidate, and the candidate's performance and retention status information after onboarding. The obtained feedback information is then aligned and structured according to the preset feedback template to generate recruitment feedback data. Among them, job requirement data, company recruitment data, candidate resume data, historical recruitment data, and recruitment feedback data are uniformly associated with the session identifier corresponding to the target recruitment position through the recruitment session instance, and are transferred and shared among the agents through the tool call loop mechanism.

[0046] The telephone screening agent records the communication results, salary intentions, and start date of candidates; the interview coaching agent records the evaluation of each round of interviews, skills evaluation, and overall evaluation; and the onboarding follow-up agent records the candidate's performance information, probation period results, and onboarding retention status.

[0047] The preset feedback template includes fields for feedback stage, feedback time, feedback object, evaluation result, evaluation content, and processing status. Feedback data generated by each agent is organized uniformly according to the above fields and associated with the session identifier corresponding to the target recruitment position through the recruitment session instance, so that the data generated within the same recruitment cycle can be managed uniformly.

[0048] The tool call loop mechanism sequentially completes the data transfer between the job analysis agent, the enterprise background analysis agent, the search strategy agent, the resume screening agent, the telephone screening agent, the interview coaching agent, and the onboarding follow-up agent according to the preset call order. Each agent reads existing data and writes new processing results through the recruitment session instance.

[0049] For example, if a candidate indicates during the phone interview that they can start work within a week, receives an A rating in the technical interview, maintains an A performance rating for three months after joining the company, and has a normal retention status, the initial phone screening agent, the interview coaching agent, and the onboarding follow-up agent will each generate corresponding feedback records. These records will be uniformly organized into recruitment feedback data according to a preset feedback template and linked to the corresponding recruitment conversation identifier.

[0050] In this embodiment, job requirement data, enterprise recruitment data, candidate resume data, historical recruitment data, and recruitment feedback data are uniformly organized through recruitment session instances. Multiple intelligent agents are used to collect, parse, and structure recruitment data from different sources, ensuring that data generated at different stages maintains a unified correlation, reducing information silos between recruitment data, and providing a consistent data foundation for the subsequent construction of standard skill sets and calculation of dynamic skill load sets. At the same time, it reduces the impact of differences in the formats of recruitment data from multiple sources on subsequent processing.

[0051] In one embodiment, such as Figure 5 and Figure 6 As shown, step S20 involves extracting job skill items and candidate skill items based on job requirement data and candidate resume data, standardizing and mapping these skill items, and constructing a standard skill set corresponding to the target job position based on the mapping results. This specifically includes the following steps: S21: Perform named entity recognition and text segmentation on the job qualification information and job responsibility information in the job requirement data to extract a set of job skill phrases corresponding to the target recruitment position.

[0052] The job description parsing module reads the job qualification and responsibility information from the job requirements data and performs Chinese word segmentation and named entity recognition on the text. For Chinese word segmentation, it uses either the Jieba or LTP word segmentation algorithms from existing natural language processing. For named entity recognition, it uses the BERT-BiLSTM-CRF named entity recognition algorithm from existing natural language processing to identify skill entities such as skill names, development frameworks, programming languages, databases, middleware, software tools, and professional certificates.

[0053] Each extracted skill phrase retains its source field, its location in the text, and its frequency of occurrence, and these are combined to form a set of job skill phrases.

[0054] For example, job requirement data may contain descriptions such as "familiar with Java development", "proficient in Spring Boot framework", "familiar with MySQL database", and "has Redis development experience". After word segmentation and named entity recognition, job skill phrases such as "Java", "Spring Boot", "MySQL" and "Redis" are extracted respectively, and the source field and the position of each job skill phrase are recorded.

[0055] S22: Perform named entity recognition and text segmentation on the work experience, project experience, and skill certificate information in the candidate's resume data to extract the candidate's skill phrase set.

[0056] Using the same Chinese word segmentation algorithm and named entity recognition algorithm as in step S21, skill entity extraction is performed on the work experience information, project experience information, and skill certificate information in the candidate's resume data.

[0057] Each candidate's skill phrase is also recorded with the corresponding years of work experience, most recent usage time, project affiliation, and skill source fields, which reflect the candidate's actual mastery of the skill.

[0058] For example, if a candidate's resume data contains information such as "responsible for three years of Spring Cloud microservice development", "has Docker deployment experience", and "obtained Oracle database certification", named entity recognition will be used to obtain candidate skill phrases such as "Spring Cloud", "Docker", and "Oracle database", and the corresponding work experience and project sources will be recorded.

[0059] S23: Call the preset skill knowledge graph, and perform synonym normalization on each skill phrase in the job skill phrase set and the candidate skill phrase set. Map skill phrases with different expressions but the same meaning to a unified standard skill name to obtain standardized job skill items and standardized candidate skill items.

[0060] Establish a pre-defined skill knowledge graph. The skill knowledge graph is organized using a node-relationship structure, where nodes represent standard skill names, and relationships represent synonymy, inclusion, hierarchical, and associative relationships between skills.

[0061] The skill knowledge graph can be constructed based on publicly available occupational skill standards, historical recruitment data from enterprises, and skill dictionaries. For newly added skill names, if the semantic similarity between the skill name and an existing standard skill name exceeds a preset normalization threshold, the skill name is mapped to the corresponding standard skill name; otherwise, a new standard skill node is created.

[0062] The semantic similarity between skill phrases was obtained by generating text vectors using the existing text vector algorithm Sentence-BERT and calculating cosine similarity.

[0063] The preset normalization threshold can be set to 0.90.

[0064] For example, the job skill phrase set contains "JavaEE" and "Java development", the candidate skill phrase set contains "J2EE", and all three correspond to the standard skill node "Java" in the skill knowledge graph. The semantic similarity between the three and the standard skill node is greater than 0.90, so they are uniformly mapped to the standard skill name "Java". The job skill phrase "SpringBoot" and the candidate skill phrase "Spring Boot" are uniformly mapped to the standard skill name "Spring Boot".

[0065] S24: Associate and map the standardized job skills items with the standardized candidate skills items, and construct a set of standard skills corresponding to the target recruitment position based on the mapping results.

[0066] The standardized job skills items and standardized candidate skills items are matched item by item. A unified skill mapping relationship is established based on the correspondence between the skills items, and a standard skill set corresponding to the target recruitment position is generated.

[0067] The standard skill set is organized using a unified data structure. Each skill item includes at least the standard skill name, skill category label, skill level label, initial skill weight, and associated skill identifier.

[0068] Among them, the skill category tag is used to identify categories such as programming language, database, development framework, middleware, operation and maintenance tools or professional certificates; the skill level tag is used to identify basic skills, core skills or extended skills; the initial skill weight can be initialized according to the frequency of occurrence in the job requirement data.

[0069] For example, if the job skill items include "Java", "Spring Boot", and "Redis", and the candidate skill items include "Java", "MySQL", and "Redis", then the standard skill set includes "Java", "Spring Boot", "Redis", and "MySQL". "Java" is associated with both the job skill items and the candidate skill items, "Spring Boot" is associated with only the job skill items, and "MySQL" is associated with only the candidate skill items.

[0070] Specifically, step S24 is implemented as follows: Calculate the semantic similarity between standardized job skill items and standardized candidate skill items. Establish an association mapping relationship between job skill items and candidate skill items with semantic similarity greater than a preset similarity threshold to obtain an association mapping table between job skill items and candidate skill items.

[0071] Standardized job skill items and standardized candidate skill items are generated separately, and the semantic similarity between each skill item is calculated using cosine similarity.

[0072] When the semantic similarity is greater than the preset similarity threshold, a corresponding relationship is established in the association mapping table; otherwise, they are retained as independent skill items.

[0073] The preset similarity threshold can be determined based on historical recruitment data statistics, for example, set to 0.85.

[0074] The association mapping table includes job skill number, candidate skill number, standard skill number, and semantic similarity value.

[0075] For example, the semantic similarity between the job skill item "Redis Cluster" and the candidate skill item "Redis Cluster Deployment" is 0.93, so a relationship is established; however, the semantic similarity between the job skill item "Kubernetes" and the candidate skill item "Docker" is 0.48, so no relationship is established.

[0076] Based on the association mapping table, the standardized job skill items and standardized candidate skill items are merged and deduplicated, and skill category labels, skill level labels and initial skill weights are configured for each merged skill item to generate a standard skill set corresponding to the target recruitment position.

[0077] Based on the association mapping table, skill items mapped to the same standard skill number are merged, and only one standard skill record is retained.

[0078] Subsequently, corresponding skill category labels are determined based on the skill knowledge graph, initial skill weights are determined based on the frequency of occurrence of skill items in the job requirement data and the fields they are located in, and skill level labels are determined based on job responsibility information and qualification information.

[0079] Among them, the initial skill weight of skills explicitly required in the job qualification information is higher than that of skills in the job responsibility information; and the initial skill weight of skills that appear in both the job responsibility information and the job qualification information is further increased.

[0080] For example, the standard skill name "Java" comes from the job skill item "Java Development" and the candidate skill item "J2EE". After merging the two, only the standard skill name "Java" is retained. The skill category label is configured as "Development Language", the skill level label is configured as "Core Skill", and the initial skill weight is calculated to be 0.92 based on the frequency and field position in the job requirement data.

[0081] In this embodiment, a unified skill knowledge graph is established to standardize the set of job skill phrases and the set of candidate skill phrases. An association mapping table is used to complete the unified association between job skill items and candidate skill items, so that skills from different sources and with different expressions can be mapped to a unified standard skill name. This reduces the impact of skill expression differences on subsequent dynamic skill load calculation, improves the consistency and completeness of the standard skill set, and provides a stable data foundation for the subsequent generation of dynamic skill load sets and the reconstruction of skill requirements.

[0082] In one embodiment, such as Figure 7 As shown, step S30 involves statistically analyzing multiple job posting records corresponding to each skill item in the standard skill set in the enterprise recruitment data to calculate the enterprise demand intensity index corresponding to that skill item. It also involves statistically analyzing historical recruitment performance indicators and recruitment cycle characteristic indicators corresponding to historical recruited candidates and historical recruitment processes in historical recruitment data. Based on the enterprise demand intensity index, historical recruitment performance indicators, and recruitment cycle characteristic indicators, a dynamic skill load set corresponding to the target recruitment position is generated. Specifically, this includes the following steps: S31: For each skill item in the statistical standard skill set, calculate the enterprise demand intensity index corresponding to the skill item by weighting it in multiple job posting records corresponding to the enterprise recruitment data, including the frequency of occurrence, number of recruits, duration of job openness, recruitment urgency, position weight in the job description text, and the labeling type of being marked as a required skill or priority skill.

[0083] Analyze the demand for each skill item in the standard skill set within enterprise recruitment data, and establish a statistical record of enterprise demand. This record should include at least the skill item identifier, frequency of occurrence, number of openings, duration the position remains open, recruitment urgency, position weight, and labeling type.

[0084] The position weight is determined based on the position of the skill item in the job description text. The position weight of the skill item in the job qualification information is greater than that of the skill item in the job responsibility information. The labeling types include required skills, priority skills and common skills, and different labeling types correspond to different weight coefficients.

[0085] The enterprise demand intensity index is obtained by weighting the above statistical items according to the preset weight coefficients. The preset weight coefficients can be configured based on the statistical results of historical recruitment data and can be adjusted according to the enterprise's recruitment habits.

[0086] For example, the skill "Java" appeared in 35 job postings in the past three months, with a corresponding number of 82 openings. The job postings were open for an average of 18 days and were marked as an essential skill. The demand intensity index of "Redis Cluster" was significantly higher than that of "Redis Cluster," which only appeared in 5 job postings and was marked as a common skill.

[0087] S32: Extract the percentage of occurrence of each skill item in the standard skill set from historical recruitment data, the historical interview pass rate, and the historical post-employment performance rating. Calculate the historical recruitment performance index corresponding to each skill item by weighting the percentage of occurrence, the historical interview pass rate, and the historical post-employment performance rating.

[0088] For each skill item in the standard skill set, we statistically analyze its appearance in historical recruited candidates and evaluate the person-job fit effect corresponding to the skill item by combining historical interview results and onboarding performance.

[0089] Historical recruitment performance indicators are calculated by comprehensively considering the percentage of candidates recruited in the past, the historical interview pass rate, and the historical post-employment performance ratings. The historical post-employment performance ratings can be quantified using the company's existing performance rating system; for example, A, B, and C levels correspond to different scoring values.

[0090] For example, although the skill item "Spring Boot" has a high demand from enterprises, the historical performance ratings of those hired have consistently remained at the A level and the interview pass rate has reached 82%. Therefore, its historical hiring performance index is higher than that of skill items that only have high recruitment demand but have average historical performance.

[0091] S33: Extract the recruitment cycle duration corresponding to the historical recruitment process for each skill item in the standard skill set from historical recruitment data, calculate the correlation coefficient between the satisfaction level of each skill item and the recruitment cycle duration, and determine the skill scarcity of the corresponding skill item based on the correlation coefficient, and use the skill scarcity as the recruitment cycle characteristic indicator of the corresponding skill item.

[0092] We statistically analyzed the historical recruitment cycle length for each skill category and corresponding job positions, and calculated the correlation between the candidate's skill satisfaction and the recruitment cycle length.

[0093] The correlation coefficient was calculated using the Pearson correlation coefficient from existing statistical analysis to reflect the impact of changes in skill satisfaction on the recruitment cycle length.

[0094] When the skill satisfaction level is low and the recruitment cycle is long, the corresponding skill item is judged to have a high degree of skill scarcity; conversely, the skill scarcity level is judged to be low, and a recruitment cycle characteristic index is generated accordingly.

[0095] For example, historical data shows that there are relatively few candidates with "Kubernetes" skills, and the average recruitment cycle for the corresponding positions is 52 days. There is a strong negative correlation between the degree of skill satisfaction and the recruitment cycle, so the corresponding skill item has a high recruitment cycle characteristic index.

[0096] S34: Based on the enterprise demand intensity index, historical recruitment performance index and recruitment cycle characteristic index corresponding to each skill item, the dynamic skill load value of the corresponding skill item is obtained by fusion calculation according to the preset weight coefficient, and the dynamic skill load set corresponding to the target recruitment position is generated based on the dynamic skill load value.

[0097] For each skill item in the standard skill set, we obtain the enterprise demand intensity index, historical recruitment performance index, and recruitment cycle characteristic index, and use a weighted fusion method to calculate the dynamic skill load value.

[0098] The dynamic skill load value is used to describe the overall demand status of the corresponding skill in the current recruitment environment. The higher the value, the higher the overall demand for the corresponding skill in the current recruitment environment.

[0099] The dynamic skill load set is organized using a key-value structure. Each record includes a skill item identifier, dynamic skill load value, update time, and associated skill identifier.

[0100] The preset weighting coefficients can be determined based on historical recruitment data statistics. For example, the weight of the enterprise demand intensity indicator is 0.45, the weight of the historical recruitment performance indicator is 0.35, and the weight of the recruitment cycle characteristic indicator is 0.20.

[0101] For example, if the enterprise demand intensity index for the skill item "Java" is 0.92, the historical recruitment performance index is 0.89, and the recruitment cycle characteristic index is 0.76, then the dynamic skill load value of 0.878 is calculated according to the preset weights and written into the dynamic skill load set.

[0102] Furthermore, the specific steps for generating the dynamic skill load set corresponding to the target recruitment position based on the dynamic skill load value are as follows: Establish skill association relationships between each skill item in the standard skill set, analyze the association strength between skill items based on the skill association relationships, and construct a skill association network.

[0103] A skill association network is established based on a standard skill set. The skill association network is organized using a graph structure, where skill items are nodes and the relationships between different skill items are edges.

[0104] Skill associations are determined based on the frequency of co-occurrence of skills in historical recruitment data, the co-occurrence of candidate skills, and the associations in the skill knowledge graph; the association strength is calculated using a weighted statistical method and stored as edge weights.

[0105] For example, "Spring Boot" and "Java" appear together in job requirement data for a long time, so they have a high correlation strength; "Redis" and "MySQL" appear together less frequently, so the correlation strength is lower.

[0106] Based on the skill association network, the change in dynamic skill load corresponding to the target skill item whose dynamic skill load value changes is propagated to the associated skill items that are related to the target skill item, so as to obtain the propagated and corrected dynamic skill load value corresponding to each skill item.

[0107] When the dynamic load value of a target skill item changes, the change is propagated proportionally to the associated skill item according to the corresponding edge weight in the skill association network.

[0108] The propagation ratio is determined based on the correlation strength; the higher the correlation strength, the higher the corresponding propagation ratio. When the correlation strength is lower than the preset propagation threshold, no propagation correction is performed.

[0109] For example, the dynamic load value of the "Java" skill increases by 0.10, while the value of "Spring Boot" (with a correlation strength of 0.80) increases by 0.08, and the value of "MySQL" (with a correlation strength of 0.30) increases by 0.03, thus reflecting the synchronous change in the demand for related skills.

[0110] The dynamic skill load values ​​of each skill item, after propagation correction, are associated and organized to generate a set of dynamic skill loads corresponding to the target recruitment position.

[0111] Update the propagation-corrected dynamic skill load values ​​for each skill item separately, and organize them uniformly according to the skill item identifier to form a new set of dynamic skill loads.

[0112] Each record in the dynamic skill load set includes a skill item identifier, a dynamic skill load value, an update time, an associated skill identifier, and a propagation correction identifier, which are used to identify whether the corresponding dynamic skill load value has undergone associated propagation correction.

[0113] For example, after the skill association propagation is completed, skills such as "Java", "Spring Boot", "Redis", and "Docker" will update their corresponding dynamic skill load values ​​and be reorganized into a new set of dynamic skill loads.

[0114] In this embodiment, the dynamic skill load value corresponding to each skill item is calculated by integrating enterprise recruitment data, historical recruitment data, and recruitment cycle information. The dynamic skill load value is then corrected by association propagation using a skill association network. This allows the dynamic skill load set to not only reflect the demand changes of individual skill items but also the correlation and influence between skills. This more accurately describes the comprehensive demand status of each skill item in the current recruitment environment and provides a more quantitative basis for subsequent skill demand reconstruction that is more in line with actual recruitment needs.

[0115] In one embodiment, such as Figure 8 As shown, in step S40, the skill requirements of each skill item in the standard skill set are reconstructed based on the dynamic skill load set. The priority and level of the skill requirements corresponding to each skill item are dynamically adjusted according to the changes in the dynamic skill load value of each skill item, so as to generate the target job skill requirement structure corresponding to the current recruitment environment. Specifically, this includes the following steps: S41: Sort the dynamic skill load values ​​corresponding to each skill item in the dynamic skill load set in descending order to obtain the skill requirement priority corresponding to each skill item.

[0116] Read the dynamic skill load values ​​corresponding to each skill item in the dynamic skill load set, and sort them from highest to lowest. A larger dynamic skill load value indicates a higher overall demand for the corresponding skill item in the current recruitment environment, and thus a higher skill requirement priority; a smaller dynamic skill load value indicates a relatively lower current recruitment demand for the corresponding skill item, and thus a lower skill requirement priority. After sorting, reallocate skill requirement priorities for each skill item in the standard skill set to form a skill requirement priority sequence corresponding to the current recruitment cycle.

[0117] For example, for the four skills "Python", "Machine Learning", "Docker" and "Linux" in the standard skill set, their dynamic skill load values ​​are 0.93, 0.88, 0.71 and 0.66 respectively. After sorting them from high to low according to their dynamic skill load values, the corresponding skill demand priorities are "Python", "Machine Learning", "Docker" and "Linux", which is used to reflect the importance of each skill to the current job posting.

[0118] S42: Skills with dynamic skill load values ​​greater than the preset first dynamic load threshold are assigned to the core essential skills layer; skills with dynamic skill load values ​​between the first dynamic load threshold and the preset second dynamic load threshold are assigned to the important preferred skills layer; and skills with dynamic skill load values ​​less than or equal to the second dynamic load threshold are assigned to the auxiliary bonus skills layer, thus obtaining the skill requirement structure.

[0119] Based on the relationship between the dynamic skill load value corresponding to each skill item and the preset first and second dynamic load thresholds, the skill items in the standard skill set are hierarchically divided. The preset first and second dynamic load thresholds are configured according to the type of job posting, historical recruitment data statistics, and corporate recruitment strategies, with the first dynamic load threshold being greater than the second dynamic load threshold. Skill items with dynamic skill load values ​​greater than the first dynamic load threshold are classified into the core essential skills layer; skill items with dynamic skill load values ​​between the first and second dynamic load thresholds are classified into the important preferred skills layer; and skill items with dynamic skill load values ​​less than or equal to the second dynamic load threshold are classified into the auxiliary bonus skills layer, thus forming a skill demand structure corresponding to the current recruitment environment.

[0120] For example, the first dynamic load threshold is set to 0.85, and the second dynamic load threshold is set to 0.60. When the dynamic load value of "Python" is 0.93, it is classified into the core essential skills layer; when the dynamic load value of "Docker" is 0.72, it is classified into the important preferred skills layer; and when the dynamic load value of "Hadoop" is 0.48, it is classified into the auxiliary bonus skills layer, thereby obtaining the skill requirement structure corresponding to the current recruitment cycle of the target recruitment position.

[0121] S43: Calculate the change in the dynamic skill load value of each skill item relative to the previous recruitment cycle to determine whether each skill item meets the preset skill level migration conditions. If it does, adjust the corresponding skill item to the new skill level to obtain a new skill requirement structure, and simultaneously adjust the skill requirement priority of related skill items that are related to the skill item.

[0122] Extract the dynamic skill load values ​​for the same skill items in the current recruitment cycle and the previous recruitment cycle, calculate the change range of the dynamic skill load values, and determine whether the corresponding skill item has undergone skill level migration based on preset skill level migration conditions. The preset skill level migration conditions are set based on the change range of the dynamic skill load values, the number of recruitment cycles that continuously meet the migration conditions, and the current skill level of the corresponding skill item. When the skill level migration conditions are met, the corresponding skill item is adjusted to the new skill level; for related skill items that have a skill association relationship with this skill item, the corresponding skill demand priority is adjusted synchronously according to the association strength in the skill association network, so that the related skill items can maintain a consistent demand change trend with the target skill item, resulting in a new skill demand structure.

[0123] For example, if the dynamic skill load value of the "Large Language Model Application" skill item continues to increase for three consecutive recruitment cycles and exceeds the first dynamic load threshold, it will be adjusted from the important preferred skill layer to the core essential skill layer. Since there is a high correlation between it and the "Prompt Engineering" skill item, the priority of the skill requirement corresponding to the "Prompt Engineering" skill item will be increased simultaneously, so that the skill requirement structure can reflect the synergistic change relationship between skills.

[0124] S44: Based on the new skill requirement structure and the adjusted skill requirement priority of each skill item, the boundaries of each level are redefined, and the adjusted target job skill requirement structure is generated.

[0125] Based on the new skill requirement structure and the adjusted skill requirement priorities for each skill item, the core essential skill layer, important preferred skill layer, and auxiliary bonus skill layer are reorganized. The skill items included in each skill level and their order are updated, and an adjusted target job skill requirement structure is generated. The generated target job skill requirement structure records the skill level, skill requirement priority, and dynamic skill load value for each skill item, and serves as the job skill requirement structure used when matching candidates for recruitment in the current recruitment cycle.

[0126] For example, when the "Agent Development" skill item is moved from the important preferred skill layer to the core essential skill layer, the skill order in the core essential skill layer is reorganized, with skills with higher dynamic skill load values ​​placed first; at the same time, the adjusted three-layer skill structure is saved as the target job skill requirement structure for use in the candidate recruitment matching stage.

[0127] This embodiment prioritizes, categorizes, and migrates skill requirements within a standard skill set based on the dynamic skill load values ​​corresponding to each skill item in the dynamic skill load set. It also synchronously adjusts the skill requirement priorities of associated skill items based on skill relationships, generating a target job skill requirement structure that dynamically reflects changes in the current recruitment environment. Compared to using fixed skill weights for job matching, this embodiment continuously adjusts the job skill structure in response to changes in corporate recruitment needs, historical recruitment results, and recruitment cycles. This improves the responsiveness of the target job skill requirement structure to changes in recruitment demands and enhances the consistency between candidate matching results and the company's actual staffing needs.

[0128] In one embodiment, such as Figure 9 As shown, step S50, which involves determining the corresponding skill fit based on the correspondence between the target job skill requirement structure and the candidate's skill items, and generating the corresponding candidate recruitment matching result based on the skill fit, specifically includes the following steps: S51: Compare the candidate's skills with the skills included in the core essential skills layer, important preferred skills layer, and auxiliary bonus skills layer in the skill requirement structure of the target position to determine the corresponding hit skills for each skill level.

[0129] The system reads the candidate's skill items and the skill items included in each skill level of the target position's skill requirement structure, and matches them item by item against the standard skill names in the standard skill set. When a candidate's skill item has the same standard skill name as a skill item in the target position's skill requirement structure, the corresponding skill item is considered a match; similarly, when a candidate's skill item can be mapped to the same standard skill name based on the association mapping relationship in the standard skill set, the corresponding skill item is also considered a match. Matched skill items are counted separately for the core essential skill layer, the important preferred skill layer, and the auxiliary bonus skill layer. The skill level to which each skill item belongs and its corresponding dynamic skill load value are recorded, forming a set of matched skill items for each skill level.

[0130] For example, the core essential skills layer of the target position's skill requirements structure includes "Python," "Machine Learning," and "Deep Learning," the important preferred skills layer includes "Docker" and "Linux," and the auxiliary bonus skills layer includes "Kubernetes." The candidate's skills include "Python Development," "TensorFlow," "Docker," and "Linux." Based on the mapping relationships in the standard skill set, "Python Development" is mapped to "Python," and "TensorFlow" is mapped to "Deep Learning." Therefore, the candidate matches "Python" and "Deep Learning" in the core essential skills layer, "Docker" and "Linux" in the important preferred skills layer, but no skills are matched in the auxiliary bonus skills layer.

[0131] S52: Determine the skill contribution value corresponding to each hit skill item based on the dynamic skill load value corresponding to the hit skill item, calculate the level matching degree corresponding to each skill level based on the skill contribution value, and merge the level matching degrees corresponding to each skill level to obtain the skill suitability degree corresponding to the candidate.

[0132] The system reads the dynamic skill load value corresponding to each matched skill item and calculates the corresponding skill contribution value based on the skill level to which the skill item belongs. Each skill level is assigned a corresponding level weight, which is pre-set based on the company's recruitment strategy, job level, and job type. The level weight of the core essential skills level is greater than that of the important preferred skills level, and the level weight of the important preferred skills level is greater than that of the auxiliary bonus skills level. The system calculates the level matching degree of the corresponding level based on the skill contribution values ​​of all matched skill items within the same skill level. Then, it weights and merges the level matching degrees according to the level weights of each skill level to obtain the skill suitability of the candidate.

[0133] The preset level weights can be set based on the company's historical recruitment data statistics. For example, the core essential skills level can be set to 0.60, the important preferred skills level to 0.30, and the auxiliary bonus skills level to 0.10. The sum of the weights of each level is 1.

[0134] For example, a candidate hits two skills in the core essential skills layer with dynamic skill load values ​​of 0.96 and 0.91 respectively; hits two skills in the important preferred skills layer with dynamic skill load values ​​of 0.74 and 0.69 respectively; and hits one skill in the auxiliary bonus skills layer with a dynamic skill load value of 0.55. The skill contribution value is calculated based on the dynamic skill load value corresponding to each hit skill, and then the level matching degree corresponding to the three skill levels is calculated separately. Finally, the candidate's skill fit degree of 0.89 is obtained by merging according to the level weight.

[0135] S53: Compare the skill fit with the preset matching threshold, filter out candidates whose skill fit is greater than or equal to the preset matching threshold, and sort them from high to low according to their skill fit to generate the corresponding candidate recruitment matching results.

[0136] The skill fit of each candidate is compared with a preset matching threshold. When the skill fit is greater than or equal to the preset matching threshold, the candidate is deemed to meet the skill requirements of the target recruitment position; when the skill fit is lower than the preset matching threshold, the candidate is not included in the candidate recruitment matching results. For candidates who meet the matching conditions, they are sorted from highest to lowest skill fit. When multiple candidates have the same skill fit, the matching degree of the core essential skill level, the cumulative value of the dynamic skill load value corresponding to the matched skill items, and the comprehensive evaluation value corresponding to the historical recruitment performance indicators are compared in sequence to obtain the final candidate recruitment matching result.

[0137] The preset matching threshold is configured based on the job level and job category. For example, the threshold is set to 0.85 for senior positions, 0.75 for intermediate positions, and 0.65 for junior positions.

[0138] For example, if the skill fit of five candidates for the target job is 0.93, 0.89, 0.82, 0.76, and 0.71 respectively, and the preset matching threshold is set to 0.80, the top three candidates are selected and ranked from highest to lowest according to their skill fit to generate the candidate recruitment matching results. If the skill fit of the first and second candidates is 0.89, the matching degree of their core essential skills is compared first, and the candidate with the higher matching degree is ranked higher.

[0139] This embodiment performs tiered matching of candidate skills based on the skill requirements structure of the target position. It calculates the skill contribution value and the tiered matching degree for each skill level by combining the dynamic skill load value corresponding to each skill item. This ensures that the skill fit not only reflects the skill coverage between the candidate and the target recruitment position but also comprehensively reflects the actual demand for each skill item in the current recruitment environment. Compared to recruitment matching using fixed skill weights, this embodiment can adjust the skill matching results in real time based on the dynamic skill load set, improving the consistency between candidate recruitment matching results and the actual recruitment needs of the enterprise, and enhancing the accuracy of recruitment matching.

[0140] In one embodiment, such as Figure 10 As shown, step S60 involves updating the dynamic skill load set using recruitment feedback data, and then readjusting the skill requirement structure for the target position based on the updated dynamic skill load set. This specifically includes the following steps: S61: Extract feedback information from the candidate recruitment feedback process from the recruitment feedback data, and calculate the adaptation deviation value for each skill item based on the degree of deviation between the skill fit corresponding to the candidate recruitment matching result and the actual recruitment result corresponding to the feedback information.

[0141] The process involves retrieving feedback information from the recruitment feedback data corresponding to the candidate's recruitment process. This feedback information includes results of telephone communication, interview evaluations, hiring decisions, probationary period assessments, and employee retention status. A consistency analysis is performed between the skill fit of the candidate's recruitment matching results and the actual recruitment results corresponding to the feedback information. Based on the degree of deviation between the two, a fit deviation value is calculated for each skill item. A larger fit deviation value indicates a significant discrepancy between the current dynamic skill load value of the corresponding skill item and the company's actual recruitment needs; a smaller fit deviation value indicates that the current dynamic skill load value of the corresponding skill item accurately reflects the company's actual recruitment needs.

[0142] The fit deviation value can be calculated based on the difference between the skill fit degree and the evaluation value corresponding to the actual recruitment result. The evaluation value corresponding to the actual recruitment result is configured according to the completion status of each stage of the recruitment process. For example, the evaluation value of hiring and completing the probation period is higher than the evaluation value of only completing telephone communication.

[0143] For example, if a candidate's skill fit score is 0.92, but they fail the interview, it indicates that the dynamic skill load values ​​for some of the matched skills are too high, and a larger fit deviation value is calculated for the corresponding skill items. Another candidate's skill fit score is 0.74, but their performance after being hired is excellent, indicating that the dynamic skill load values ​​for some of the matched skills are too low, and the fit deviation value for the corresponding skill items is also calculated to reflect the difference between the current skill demand assessment results and the actual recruitment results.

[0144] S62: Based on the adaptation deviation value, correct the enterprise demand intensity index, historical recruitment performance index, and recruitment cycle characteristic index of the corresponding skill items in the dynamic skill load set, and recalculate the dynamic skill load value of the corresponding skill items; calculate the time decay factor based on the update time of the dynamic skill load value of each skill item, and adjust the decay of the dynamic skill load value of the skill items that have exceeded the preset update cycle and have not changed in recruitment demand; re-execute skill association propagation based on the recalculated dynamic skill load value to obtain the updated dynamic skill load set.

[0145] Based on the adaptation deviation values ​​corresponding to each skill item, the enterprise demand intensity index, historical recruitment performance index, and recruitment cycle characteristic index of the corresponding skill items in the dynamic skill load set are respectively corrected, and the dynamic skill load value of the corresponding skill item is recalculated based on the corrected indexes. For skill items that have exceeded the preset update cycle and have not experienced changes in recruitment demand, a time decay factor is calculated based on the update time corresponding to the dynamic skill load value, and the corresponding dynamic skill load value is adjusted to reduce the impact of long-term unupdated skill items on the current recruitment strategy. After completing the update of the dynamic skill load value, the skill association propagation is re-executed based on the skill association network to propagate the updated dynamic skill load changes to the associated skill items, and the dynamic skill load values ​​corresponding to the associated skill items are corrected synchronously to generate an updated dynamic skill load set.

[0146] The preset update cycle is configured based on job category, industry change frequency, and company recruitment cycle. For example, the update cycle is set to 30 days for internet R&D positions, 60 days for manufacturing positions, and 45 days for finance positions. The time decay factor is calculated based on the length of time since the skill item was last updated. The longer the update time, the smaller the corresponding time decay factor.

[0147] For example, if the "Large Language Model Application" skill item shows a large mismatch value for several consecutive recruitment cycles, the corresponding historical recruitment performance indicators will be revised and the corresponding dynamic skill load value will be increased. If the "Hadoop Development" skill item has not had any new recruitment demand for several consecutive update cycles and exceeds the preset update cycle, the corresponding dynamic skill load value will be reduced according to the time decay factor. At the same time, the dynamic skill load value corresponding to the "MapReduce Development" skill item associated with the "Hadoop Development" skill item will be revised synchronously through the skill association network.

[0148] S63: Re-execute the skill requirement reconstruction based on the updated dynamic skill load set to adjust the skill requirement priority of each skill level and skill item in the target job skill requirement structure, and use the adjusted target job skill requirement structure as the initial job skill requirement structure for the next recruitment cycle.

[0149] Based on the updated dynamic skill load set, a new skill requirement reconstruction is performed. This involves re-prioritizing each skill item in the standard skill set, re-classifying skill levels, and determining skill level migration, thereby regenerating the target job's skill requirement structure. The adjusted target job skill requirement structure is saved to the historical recruitment data corresponding to the target job and serves as the initial skill requirement structure for the next recruitment cycle, enabling continuous dynamic iteration and updates based on recruitment feedback data.

[0150] For example, if the dynamic skill load value of the "Agent Development" skill item continues to increase after several consecutive recruitment cycles, it will be adjusted from the important preferred skill layer to the core essential skill layer; if the dynamic skill load value of the "Traditional Hadoop Development" skill item continues to decrease, it will be adjusted from the important preferred skill layer to the auxiliary bonus skill layer, and the updated target job skill requirement structure will be used as the job skill requirement structure for the next recruitment cycle to continue candidate recruitment matching.

[0151] This embodiment continuously adjusts the enterprise demand intensity index, historical recruitment performance index, and recruitment cycle characteristic index corresponding to each skill item in the dynamic skill load set by utilizing recruitment feedback data. It also performs time decay processing and skill association propagation updates based on the update time corresponding to the dynamic skill load value, enabling the dynamic skill load set to continuously reflect changes in the actual recruitment needs of enterprises. Furthermore, it re-executes skill requirement reconstruction based on the updated dynamic skill load set, allowing the skill requirement structure of the target position to be dynamically adjusted with recruitment feedback data. This forms a closed-loop iterative process covering job requirement analysis, candidate recruitment matching, recruitment feedback correction, and job skill requirement updates, improving the adaptability of the target job skill requirement structure to changes in the recruitment environment and enhancing the consistency between candidate recruitment matching results and actual enterprise recruitment results.

[0152] It should be understood that the sequence number of each step in the above embodiments does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0153] In one embodiment, a recruitment system based on dynamic skill load is provided, which corresponds one-to-one with a recruitment method based on dynamic skill load in the above embodiments. For example... Figure 11 As shown, this recruitment system based on dynamic skill loading includes a data acquisition module, a data extraction module, an indicator acquisition module, a demand restructuring module, a relationship confirmation module, and a structure adjustment module. Detailed descriptions of each functional module are as follows: The data acquisition module is used to acquire job requirement data, company recruitment data, candidate resume data, historical recruitment data, and recruitment feedback data for the target job posting.

[0154] The data extraction module is used to extract job skill items and candidate skill items based on job requirement data and candidate resume data, standardize and map the job skill items and candidate skill items, and construct a standard skill set corresponding to the target recruitment position based on the mapping results.

[0155] The indicator acquisition module is used to obtain enterprise demand intensity indicators corresponding to enterprise recruitment data, historical recruitment performance indicators, and recruitment cycle characteristic indicators corresponding to historical recruitment data, using each skill item in the standard skill set, in order to generate a dynamic skill load set corresponding to the target recruitment position.

[0156] The requirement refactoring module is used to refactor the skill requirements of each skill item in the standard skill set based on the dynamic skill load set. It dynamically adjusts the priority and skill level of the skill requirements corresponding to each skill item according to the changes in the dynamic skill load value of each skill item, so as to generate the target job skill requirement structure corresponding to the current recruitment environment.

[0157] The relationship confirmation module is used to determine the corresponding skill fit degree based on the correspondence between the skill requirements structure of the target position and the skill items of the candidates, and to generate the corresponding candidate recruitment matching results based on the skill fit degree.

[0158] The structural adjustment module is used to update the dynamic skill load set based on recruitment feedback data, and to readjust the skill requirement structure of the target position based on the updated dynamic skill load set.

[0159] It should be noted that the information interaction and execution process between the above systems / modules are based on the same concept as the method embodiments of this application. For details on their specific functions and technical effects, please refer to the method embodiments section, and they will not be repeated here.

[0160] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the above-described division of functional units and modules is merely an example. In practical applications, the above functions can be assigned to different functional units and modules as needed, that is, the internal structure of the system can be divided into different functional units or modules to complete all or part of the functions described above. The functional units and modules in the embodiments can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit. Furthermore, the specific names of the functional units and modules are only for easy differentiation and are not intended to limit the scope of protection of this application. The specific working process of the units and modules in the above system can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0161] In the above embodiments, the descriptions of each embodiment have different focuses. For parts that are not described in detail or recorded in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0162] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0163] In the embodiments provided in this application, it should be understood that the disclosed systems / devices and methods can be implemented in other ways. For example, the system / device embodiments described above are merely illustrative. For instance, the division of modules or units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between systems or units may be electrical, mechanical, or other forms.

[0164] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0165] The above-described embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application, and should all be included within the protection scope of this application.

Claims

1. A recruitment method based on dynamic skill loading, characterized in that, include: Obtain job requirement data, company recruitment data, candidate resume data, historical recruitment data, and recruitment feedback data for the target job openings; Based on the job requirements data and the candidate resume data, job skills items and candidate skills items are extracted, the job skills items and candidate skills items are standardized and associated with each other, and a standard skills set corresponding to the target recruitment position is constructed based on the mapping results. The system statistically analyzes multiple job posting records corresponding to each skill item in the standard skill set in the enterprise recruitment data to calculate the enterprise demand intensity index corresponding to that skill item. It also statistically analyzes the historical recruitment performance index and recruitment cycle characteristic index corresponding to historical recruited candidates and historical recruitment processes in the historical recruitment data. Based on the enterprise demand intensity index, the historical recruitment performance index, and the recruitment cycle characteristic index, a dynamic skill load set corresponding to the target recruitment position is generated. Based on the dynamic skill load set, the skill requirements of each skill item in the standard skill set are reconstructed. The priority and level of the skill requirements corresponding to each skill item are dynamically adjusted according to the changes in the dynamic skill load value of each skill item, so as to generate the target job skill requirement structure corresponding to the current recruitment environment. The corresponding skill fit is determined based on the correspondence between the target job skill requirement structure and the candidate skill items, and the corresponding candidate recruitment matching result is generated based on the skill fit. The dynamic skill load set is updated using the recruitment feedback data, and the skill requirement structure for the target position is readjusted based on the updated dynamic skill load set.

2. The method as described in claim 1, characterized in that, The steps for obtaining job requirement data, company recruitment data, candidate resume data, historical recruitment data, and recruitment feedback data corresponding to the target job posting include: Based on the intelligent agent runtime environment, a recruitment session instance corresponding to the target recruitment position is created. The recruitment session instance calls the preset intelligent agent interface for the enterprise recruitment management system to obtain the original job description text corresponding to the target recruitment position. The original job description text is then parsed in a structured manner, and job requirement data is extracted. The recruitment session instance calls a preset intelligent agent to connect with the enterprise credit database, the enterprise website information source, and the enterprise historical recruitment record database to obtain enterprise recruitment data corresponding to the enterprise to which the target recruitment position belongs; The recruitment session instance invokes a preset agent to search and recall candidates in the candidate resume database and professional social platform based on the job requirements data, and obtains the candidate resume data. The recruitment session instance calls the preset historical recruitment session data and extracts the historical recruitment data from the historical recruitment session data; The recruitment session instance invokes a preset agent to obtain feedback information from the candidate's telephone communication and the interviewer's evaluation feedback on the candidate, in order to generate recruitment feedback data.

3. The method as described in claim 1, characterized in that, The steps of extracting job skill items and candidate skill items based on the job requirement data and candidate resume data, standardizing and mapping the job skill items and candidate skill items, and constructing a standard skill set corresponding to the target recruitment position based on the mapping results include: Named entity recognition and text segmentation are performed on the job qualification information and job responsibility information in the job requirement data to extract a set of job skill phrases corresponding to the target recruitment position. Named entity recognition and text segmentation are performed on the work experience information, project experience information and skill certificate information in the candidate resume data to extract the candidate skill phrase set corresponding to the candidate. Each skill phrase in the set of job skill phrases and the set of candidate skill phrases is subjected to synonym normalization processing. Skill phrases with different expressions but the same meaning are mapped to a unified standard skill name, resulting in standardized job skill items and standardized candidate skill items. The standardized job skills items are associated and mapped with the standardized candidate skills items, and a standard skills set corresponding to the target recruitment position is constructed based on the mapping results.

4. The method as described in claim 3, characterized in that, The step of associating and mapping the standardized job skill items with the standardized candidate skill items, and constructing a standard skill set corresponding to the target recruitment position based on the mapping result, includes: Calculate the semantic similarity between the standardized job skill items and the standardized candidate skill items, and establish an association mapping relationship between the job skill items with semantic similarity greater than a preset similarity threshold and the candidate skill items to obtain an association mapping table; Based on the association mapping table, the standardized job skill items and the standardized candidate skill items are merged and deduplicated to generate the standard skill set corresponding to the target recruitment position.

5. The method as described in claim 1, characterized in that, The steps of statistically analyzing each skill item in the standard skill set across multiple job posting records corresponding to the enterprise recruitment data to calculate the enterprise demand intensity index corresponding to that skill item, statistically analyzing historical recruitment performance indicators and recruitment cycle characteristic indicators corresponding to historical recruited candidates and historical recruitment processes in the historical recruitment data, and generating a dynamic skill load set corresponding to the target recruitment position based on the enterprise demand intensity index, the historical recruitment performance indicators, and the recruitment cycle characteristic indicators include: The system statistically analyzes multiple job posting records corresponding to each skill item in the standard skill set in the enterprise recruitment data to calculate the enterprise demand intensity index corresponding to that skill item. Extract the percentage of each skill item in the standard skill set in the historical recruitment data, the historical interview pass rate, and the historical post-employment performance rating of each skill item, so as to calculate the historical recruitment performance index corresponding to that skill item. Extract the recruitment cycle duration corresponding to the historical recruitment process of each skill item in the standard skill set from the historical recruitment data, calculate the correlation coefficient between the satisfaction level of each skill item and the recruitment cycle duration, and determine the skill scarcity of the corresponding skill item based on the correlation coefficient, and use the skill scarcity as the recruitment cycle characteristic indicator of the corresponding skill item. The dynamic skill load value of each skill item is obtained by integrating the enterprise demand intensity index, the historical recruitment performance index, and the recruitment cycle characteristic index according to the preset weight coefficient, and a dynamic skill load set corresponding to the target recruitment position is generated based on the dynamic skill load value.

6. The method as described in claim 5, characterized in that, The step of generating the dynamic skill load set corresponding to the target recruitment position based on the dynamic skill load value includes: Establish skill association relationships between each skill item in the standard skill set, analyze the association strength between skill items based on the skill association relationships, and construct a skill association network; Based on the skill association network, the dynamic skill load change corresponding to the target skill item whose dynamic skill load value has changed is propagated to the associated skill items that are related to the target skill item, so as to obtain the propagated and corrected dynamic skill load value corresponding to each skill item. The propagation-corrected dynamic skill load values ​​corresponding to each skill item are associated and organized to generate a dynamic skill load set corresponding to the target recruitment position.

7. The method as described in claim 1, characterized in that, The step of reconstructing skill requirements for each skill item in the standard skill set based on the dynamic skill load set, and dynamically adjusting the skill requirement priority and skill level corresponding to each skill item according to the changes in the dynamic skill load value of each skill item, in order to generate a target job skill requirement structure corresponding to the current recruitment environment, includes: The skill requirement priority of each skill item is obtained by sorting the dynamic skill load values ​​corresponding to each skill item in descending order according to the dynamic skill load set. Skills with dynamic skill load values ​​greater than a preset first dynamic load threshold are assigned to the core essential skills layer; skills with dynamic skill load values ​​between the first dynamic load threshold and a preset second dynamic load threshold are assigned to the important preferred skills layer; and skills with dynamic skill load values ​​less than or equal to the second dynamic load threshold are assigned to the auxiliary bonus skills layer, thus obtaining the skill requirement structure. Calculate the change in the dynamic skill load value of each skill item relative to the previous recruitment cycle to determine whether each skill item meets the preset skill level migration conditions. If it does, adjust the corresponding skill item to the new skill level to obtain a new skill requirement structure, and simultaneously adjust the skill requirement priority of related skill items that are related to the skill item. Based on the new skill requirement structure and the adjusted skill requirement priority of each skill item, the boundaries of each level are redefined, and the adjusted target job skill requirement structure is generated.

8. The method as described in claim 7, characterized in that, The step of determining the corresponding skill fit degree based on the correspondence between the skill requirement structure of the target position and the skill items of the candidate, and generating the corresponding candidate recruitment matching result based on the skill fit degree, includes: The candidate's skills are compared with the skills in the core essential skills layer, important preferred skills layer, and auxiliary bonus skills layer in the skill requirement structure of the target position to determine the corresponding matching skills. The skill contribution value corresponding to each hit skill item is determined based on the dynamic skill load value corresponding to the hit skill item. The level matching degree corresponding to each skill level is calculated based on the skill contribution value. The level matching degrees corresponding to each skill level are merged and calculated to obtain the skill suitability degree corresponding to the candidate. The skill fit is compared with a preset matching threshold, and candidates with a skill fit greater than or equal to the preset matching threshold are selected and sorted from high to low according to their skill fit to generate corresponding candidate recruitment matching results.

9. The method as described in claim 1, characterized in that, The step of updating the dynamic skill load set using the recruitment feedback data, and readjusting the skill requirement structure of the target position based on the updated dynamic skill load set, includes: The feedback information from the candidate recruitment feedback process is extracted from the recruitment feedback data. Based on the degree of deviation between the skill fit degree corresponding to the candidate recruitment matching result and the actual recruitment result corresponding to the feedback information, the fit deviation value corresponding to each skill item is calculated. Based on the adaptation deviation value, the enterprise demand intensity index, historical recruitment performance index, and recruitment cycle characteristic index of the corresponding skill item in the dynamic skill load set are corrected respectively, and the dynamic skill load value of the corresponding skill item is recalculated to obtain the updated dynamic skill load set. The skill requirement reconstruction is re-executed based on the updated dynamic skill load set to adjust the skill requirement priority of each skill level and skill item in the target job skill requirement structure.

10. A recruitment system based on dynamic skill loading, characterized in that, include: The data acquisition module is used to acquire job requirement data, company recruitment data, candidate resume data, historical recruitment data, and recruitment feedback data corresponding to the target recruitment position. The data extraction module is used to extract job skill items and candidate skill items based on the job requirement data and the candidate resume data, perform standardization processing and association mapping on the job skill items and the candidate skill items, and construct a standard skill set corresponding to the target recruitment position based on the mapping results. The indicator acquisition module is used to count the multiple job posting records corresponding to each skill item in the standard skill set in the enterprise recruitment data, so as to calculate the enterprise demand intensity index corresponding to the skill item, count the historical recruitment performance index and recruitment cycle characteristic index corresponding to the historical candidates and historical recruitment processes in the historical recruitment data, and generate a dynamic skill load set corresponding to the target recruitment position based on the enterprise demand intensity index, the historical recruitment performance index and the recruitment cycle characteristic index. The requirement refactoring module is used to refactor the skill requirements of each skill item in the standard skill set according to the dynamic skill load set, and dynamically adjust the skill requirement priority and skill level of each skill item according to the change of the dynamic skill load value of each skill item, so as to generate the target job skill requirement structure corresponding to the current recruitment environment. The relationship confirmation module is used to determine the corresponding skill fit degree based on the correspondence between the skill requirement structure of the target position and the skill items of the candidate, and to generate the corresponding candidate recruitment matching result based on the skill fit degree. The structure adjustment module is used to update the dynamic skill load set through the recruitment feedback data, and to readjust the skill requirement structure of the target position based on the updated dynamic skill load set.