Recruitment information acquisition method, system and device and storage medium

Abstract

The invention relates to the technical field of data processing, and particularly provides a recruitment information acquisition method, system and device and a storage medium, and the method comprises the steps: receiving a question text, and extracting a resume text associated with the question text; identifying a user intention according to the question text and the resume text; acquiring corresponding recruitment data from a pre-constructed knowledge graph according to the user intention; performing statistical processing on recruitment data according to the user intention to obtain recruitment information; the entities of the mapping knowledge domain comprise positions, companies, skills, industries, geographic areas and salary standards. According to the method, highly-related recruitment information can be intelligently associated and retrieved, the matching degree between the result and the user demand is remarkably improved, and invalid browsing of job seekers is reduced.

Description

Technical Field

[0001] This invention belongs to the field of data processing technology, specifically relating to a method, system, device, and storage medium for obtaining recruitment information. Background Technology

[0002] Current job search systems primarily employ keyword-based matching techniques, which have significant limitations. For example, Boolean-based job search engines only support simple AND / OR logical queries, resulting in low relevance between search results and user needs. This forces job seekers to browse numerous irrelevant positions to find suitable opportunities, severely reducing the efficiency of the job market. Furthermore, the system only provides individual job postings and cannot query industry trends or other key indicators. Summary of the Invention

[0003] In view of the above-mentioned shortcomings of the prior art, the present invention provides a method, system, device and storage medium for obtaining recruitment information to solve the above-mentioned technical problems.

[0004] In a first aspect, the present invention provides a method for obtaining recruitment information, comprising: Receive the question text and extract the resume text associated with the question text; Identify user intent based on question text and resume text; Recruit relevant recruitment data from a pre-built knowledge graph based on the user's intent; Based on the user's intent and by statistically processing the recruitment data, recruitment information is obtained; The entities in the knowledge graph include job titles, companies, skills, industries, geographical regions, and salary standards.

[0005] In one optional implementation, identifying user intent based on the question text and resume text includes: A supervised learning algorithm is used to train an intent classifier, and the trained intent classifier is combined with a predefined intent labeling system to identify the intent of the question text. Use named entity recognition technology to extract key entities from question text and resume text; The intent is modified using the key entity.

[0006] In one optional implementation, the method for constructing a knowledge graph includes: Collect publicly available job postings and extract entities from them. Entity types include job title, company, skills, industry, geographic region, and salary standards. Extract the relationships between entities from the recruitment text, including the relationship between position and skills, position and company, company and industry, company and geographic location, skills and skills, industry and skills, and salary standards and entity combinations, where entity combinations include industry, geographic location and position; Add attributes to entities and relations, construct triples based on entities and relations, and build a knowledge graph based on triples.

[0007] In one optional implementation, relevant recruitment data is obtained from a pre-built knowledge graph based on the user intent, including: The user intent is parsed using a pre-built rules engine, and the parameters required for graph queries are generated based on the parsing results. The recruitment data is obtained by performing a graph query based on the parameters using a graph query language.

[0008] In one optional implementation, the rule engine includes: Intent processing rules are used to map user intents to the parameters required for graph queries; Entity expansion rules are used to expand synonyms for entities in user intents; Intelligent expansion rules are used to predict implicit requirements.

[0009] In an optional implementation, statistical processing of the recruitment data based on the user intent includes: Filter the recruitment data according to key conditions in the user's intent; The remaining data is aggregated and statistically analyzed to obtain the number of job postings, salary distribution, and skill matching degree.

[0010] In an optional implementation, the method further includes: The job list is displayed in tabular form, including job title, company, location, salary, skill requirements, and match rate. Generate industry distribution pie charts and salary histograms.

[0011] Secondly, the present invention provides a recruitment information acquisition system, comprising: The receiving module is used to receive the question text and extract the resume text associated with the question text; The recognition module is used to identify user intent based on the question text and resume text; The query module is used to retrieve relevant recruitment data from a pre-built knowledge graph based on the user's intent; The statistics module is used to perform statistical processing on the recruitment data based on the user's intent to obtain recruitment information; The entities in the knowledge graph include job titles, companies, skills, industries, geographical regions, and salary standards.

[0012] Thirdly, a device is provided, comprising: The memory is used to store the recruitment information retrieval program; A processor is configured to implement the steps of the recruitment information acquisition method as provided in the first aspect when executing the recruitment information acquisition program.

[0013] Fourthly, a computer-readable storage medium is provided, on which a recruitment information acquisition program is stored, and when executed by a processor, the recruitment information acquisition program implements the steps of the recruitment information acquisition method provided in the first aspect.

[0014] The beneficial effects of this invention are that the recruitment information acquisition method, system, device, and storage medium provided by this invention effectively overcome the limitations of traditional keyword matching and Boolean models by receiving question text combined with resumes and accurately identifying user intent. Based on a constructed knowledge graph containing multi-dimensional entities such as positions, companies, and skills, the system can intelligently associate and retrieve highly relevant recruitment information, significantly improving the matching degree between results and user needs and reducing invalid browsing by job seekers. More importantly, the system can perform in-depth statistical processing on recruitment data in the knowledge graph according to user intent (such as industry salary distribution, skill demand intensity, etc.), generating and feeding back valuable industry trend indicator data, breaking through the bottleneck of the original system only providing isolated job information, providing decision support for users, and improving the overall efficiency of the recruitment market.

[0015] Furthermore, the design principle of this invention is reliable, the structure is simple, and it has a very wide range of application prospects. Attached Figure Description

[0016] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, for those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0017] Figure 1 This is a schematic flowchart of a method according to an embodiment of the present invention.

[0018] Figure 2 This is a schematic block diagram of a system according to an embodiment of the present invention.

[0019] Figure 3 This is a schematic diagram of the structure of a device provided in an embodiment of the present invention. Detailed Implementation

[0020] To enable those skilled in the art to better understand the technical solutions of this invention, the technical solutions of the embodiments of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this invention, and not all embodiments. Based on the embodiments of this invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of this invention.

[0021] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains. The terminology used herein in the description of the invention is for the purpose of describing particular embodiments only and is not intended to be limiting of the invention.

[0022] The recruitment information acquisition method provided in this embodiment of the invention is executed by a computer device, and correspondingly, the recruitment information acquisition system runs on the computer device.

[0023] Figure 1 This is a schematic flowchart illustrating a method according to an embodiment of the present invention. Wherein, Figure 1 The implementing entity can be a recruitment information acquisition system. Depending on different needs, the order of the steps in this flowchart can be changed, and some can be omitted.

[0024] like Figure 1 As shown, the method includes: S1. Receive the question text and extract the resume text associated with the question text; S2. Identify user intent based on the question text and resume text; S3. Obtain relevant recruitment data from a pre-built knowledge graph based on the user's intent; the entities in the knowledge graph include job titles, companies, skills, industries, geographical regions, and salary standards; S4. Based on the user's intent and by performing statistical processing on the recruitment data, recruitment information is obtained.

[0025] In one embodiment of the present invention, based on step S1, the following will provide a possible embodiment and describe its specific implementation in a non-limiting manner.

[0026] Problem text received: Supports multiple input channels: Users can submit questions (such as "recommend suitable positions for me") through the recruitment platform's chat window, resume submission system, or a separate form.

[0027] Text preprocessing: Clean the problem text (remove punctuation and special symbols), segment the text (using NLP tools such as Jieba) and perform part-of-speech tagging, and extract core keywords (such as "front-end development", "Shanghai", "salary").

[0028] Linked user resume text: Resume parsing: Extracting structured information from user resumes using OCR technology or resume parsing APIs (such as ParseResume), including: Basic information: Name, contact information, address (geographical region).

[0029] Educational background: University, major, and degree (related to industry / skills).

[0030] Work experience: Company name, position, years of work experience, project description (extract skill keywords, such as "Java development" and "data analysis").

[0031] Skills list: clearly marked technology stack or soft skills (such as "Python" or "team collaboration").

[0032] Job objective: The user's target position, location, and salary expectations (if any).

[0033] Data mapping: Match entities in the resume (such as skills, industries, geographical regions) with entity types in the knowledge graph and create an index (such as the user skill "React" corresponding to the "skills" entity in the knowledge graph).

[0034] In one embodiment of the present invention, based on step S2, the following will provide a possible embodiment and describe its specific implementation in a non-limiting manner.

[0035] S201. A supervised learning algorithm is used to train an intent classifier, and the trained intent classifier is combined with a predefined intent labeling system to identify the intent of the question text.

[0036] A deep learning-based multi-classification model is constructed, which realizes semantic parsing of user questions through a predefined intent labeling system. The model architecture adopts a bidirectional long short-term memory network (Bi-LSTM) combined with an attention mechanism and is trained using a large-scale labeled corpus.

[0037] The model includes: Input layer processing: A 768-dimensional word embedding space is constructed using a fusion method of Word2Vec and GloVe pre-trained word vectors. After word segmentation of the question text, a fixed-length sequence vector representation of 50 is generated.

[0038] Feature extraction layer: A two-layer Bi-LSTM network is used to capture temporal dependencies in the text, with each layer containing 128 hidden units. Contextual semantic information is also obtained through a bidirectional propagation mechanism.

[0039] Attention mechanism: A multi-head self-attention module is introduced to calculate the importance weight of each position in the text sequence, as shown in the following formula:

[0040] Where Attention(Q, K, V) represents the output calculated by the attention mechanism based on the query, key, and value, where Q, K, and V are the query, key, and value matrices, respectively. k The dimension of the key vector.

[0041] Classification layer: Maps to a predefined intent label space through a fully connected layer, and uses the Softmax function to output the probability distribution.

[0042] Where exp is the exponential function, z i Let be the score for the i-th class, and C be the total number of intent categories.

[0043] Training process: Dataset: 50,000 labeled samples are used, divided into training, validation and test sets in an 8:1:1 ratio.

[0044] Optimization strategy: The Adam optimizer is used with a learning rate of 0.001 and a batch size of 64. An early stopping strategy is used to prevent overfitting.

[0045] Evaluation metrics: Achieved an accuracy of 89.7% on the test set and an F1-score (F1 score, representing the harmonic mean of precision and recall) of 88.3%, with a recall of 92.5% for the "job search" intent.

[0046] S202. Use named entity recognition technology to extract key entities from the question text and resume text.

[0047] This paper describes a BERT-CRF-based entity recognition model combined with rule matching to achieve cross-domain entity extraction. By constructing an industry-specific dictionary, it aims to improve the accuracy of entity recognition in the recruitment field.

[0048] Model architecture: BERT encoder: Employs HuggingFace's bert-base-chinese pre-trained model to extract deep semantic features from the text.

[0049] Feature fusion layer: The word vectors output by BERT are concatenated with handcrafted features (part-of-speech tags, word frequency statistics) to form an enhanced feature representation.

[0050] CRF decoding layer: Uses conditional random fields for sequence labeling, learns the transition probabilities between labels, and defines the joint probability.

[0051] in y For the label sequence, x Given the input sequence, Z(x) is the normalization factor, ensuring that the sum of probabilities is 1, t k and s l These are the transition feature and the state feature function, respectively, λ k and μ k , where n is the length of the dynamic sequence, K is the number of transition features, and L is the number of state features.

[0052] Entity type system: Core entities: {Job title, company, skills, industry, geographic region, salary range, years of work experience}; Extended entities: {Educational requirements, Benefits, Programming language, Tool name}.

[0053] Rule matching mechanism: Build a domain dictionary containing 10,000+ entity names, covering common technology stacks (such as "SpringBoot") and industry terms (such as "blockchain").

[0054] Design regular expression templates to recognize structured entities, for example: Salary Entity: 10,000-30,000 or more; Work experience: 1-2 years | Fresh graduates | No experience required.

[0055] S203. Modify the intent using the key entity.

[0056] Introduce user profile similarity scoring to improve the ability to understand personalized intent.

[0057] Algorithm correction process: Entity-Intent Relationship Calculation: Constructing an entity-intent association matrix M e,i The strength of the association between entity e and intent i is represented by the following formula:

[0058] Where n e,i Let n be the number of times entity e appears in the intention sample i. i Let N be the total number of samples for intention i, and n be the total number of samples. e This represents the number of samples containing entity e.

[0059] Intent confidence correction: Original Intent Classifier Output Probability Distribution Entity set The corrected intention probability is:

[0060] in p i This indicates the original intent classifier for the first... i The predicted probability of each intent category, where α is a correction coefficient, set to 0.3 in the experiment.

[0061] User profile similarity score: Calculate the cosine similarity between the question entity and the user resume entity:

[0062] Where E q and E r These are the question entity vector and the resume entity vector, respectively.

[0063] The ultimate intention is determined: By combining the corrected intent probability and user profile similarity, the intent category with the highest score is selected:

[0064] in The system ultimately predicts the user intent category, and β is the balancing coefficient, which is set to 0.2 in the experiment.

[0065] In one example, the following steps are included: Intent Classification: The system first analyzes the user's question text. "What jobs are available?" indicates that the user is looking for a job; "high salary" indicates that they are concerned about salary levels; and "Shanghai, Python, AI, Fintech" are specific filtering criteria. A well-trained intent classification model will determine the intent of this question as "finding a list of jobs with specific criteria," possibly with additional sub-intents such as "salary comparison" or "skills requirement analysis."

[0066] Key entity extraction: Named entity recognition (NER) models scan both the question text and the resume text simultaneously. The question text extracts the following keywords: location Shanghai, skills Python and AI, industry attribute Fintech, and high salary.

[0067] The resume text includes: Skills: Python (3 years experience), Skills: Machine Learning (1 year experience - this may correspond to AI), Education: Bachelor of Finance, and possibly expected salary.

[0068] Intent Refinement: Additional details in the resume (such as 3 years of experience) may make the system more inclined to search for "mid-to-senior level" positions, and machine learning skills mentioned in the resume further confirm that the user possesses AI-related skills. This information will be used to refine and enrich the initial intent understanding, clarifying that the search is for mid-to-senior level, high-paying positions in Shanghai, the fintech industry, and that require Python and AI skills.

[0069] In one embodiment of the present invention, based on step S3, the following will provide a possible embodiment and describe its specific implementation in a non-limiting manner.

[0070] Methods for constructing knowledge graphs include: Collect publicly available job postings and extract entities from them. Entity types include job title, company, skills, industry, geographic region, and salary standards. Extract the relationships between entities from the recruitment text, including the relationship between position and skills, position and company, company and industry, company and geographic location, skills and skills, industry and skills, and salary standards and entity combinations, where entity combinations include industry, geographic location and position; Add attributes to entities and relations, construct triples based on entities and relations, and build a knowledge graph based on triples.

[0071] Specifically, entities and attributes include: 1. Job Posting Unique identifier: job_id (primary key); Core attributes: title: Job title (e.g., "Senior Java Engineer"); description: Detailed job description text; min_experience / max_experience: Minimum / maximum experience required (years); education_level: Educational level required (high school / associate degree / bachelor's / master's / doctorate); employment_type: Employment type (full-time / part-time / internship / contract); publish_date / expiration_date: Job posting date and deadline.

[0072] 2. Company Unique identifier: company_id (primary key); Core attributes: name: Full company name; company_type: Type of company (state-owned enterprise / foreign-owned enterprise / private enterprise / startup); founding_year: Year of establishment; employee_count: Number of employees; website: Official website link.

[0073] 3. Skills Unique identifier: skill_id (primary key); Core attributes: name: standardized skill name (e.g., "Python"); category: skill category (e.g., "programming language"); popularity: market popularity index (0-1).

[0074] 4. Industry Unique identifier: industry_id (primary key); Core attributes: name: industry name (e.g., "Artificial Intelligence"); parent_industry: parent industry ID (constructs an industry tree structure).

[0075] 5. Geographical region (Location) Unique identifier: location_id (primary key); Core attributes: city / province / country: city / province / country; gdp_per_capita: GDP per capita (regional economic indicator).

[0076] 6. Salary Benchmark Unique identifier: benchmark_id (primary key); Core attributes: percentile_25 / percentile_50 / percentile_75: 25 / 50 / 75 percentile salary values.

[0077] The core relationship between entities: 1. Relationship between job title and skills Relationship type: Job Requirements Skills (Job → Skills); Meaning: Describes the skills required of a job applicant for a specific position; Key attribute: proficiency_level (proficiency requirement): clearly distinguishes between three levels: beginner, intermediate, and advanced; Application example: A "Senior Java Development Engineer" position requires job seekers to have "Java" skills, with a proficiency level requirement of advanced.

[0078] 2. Relationship between position and company Relationship type: Company to which the job belongs (Job → Company); Meaning: Indicates which company posted the job posting; Feature: This relationship does not contain any additional attributes; Application example: A "Product Manager" position belongs to "Tencent".

[0079] 3. Company's Relationship with the Industry Relationship type: Company's industry (Company → Industry); Meaning: Indicates the business area in which the company operates; Features: Direct association without additional attributes; Application example: Alibaba is a company that belongs to the "e-commerce" industry.

[0080] 4. Company / Position and Geographical Relationship Relationship type: Company / Job location (Company or Job → Geographic region); Meaning: Indicates the location of the company headquarters / branch or the office location of a specific position; Key attribute: office_type (office type): distinguishes different types such as headquarters, branches, and offices; Application example: A position located in a "branch office" in "Shanghai".

[0081] 5. Interrelationships between skills Relationship type: Skill association (skill → skill); Meaning: Indicates the correlation between different skills; Key attribute: correlation (correlation index): a small value between 0 and 1, where 1 indicates perfect correlation; Application example: The correlation between "Python" and "data analysis" skills is 0.85.

[0082] 6. Relationship between Industry and Skills Relationship type: Industry-required skills (Industry → Skills); Meaning: Indicates the degree of demand for specific skills in a particular industry; Key attribute: demand_index: a decimal value between 0 and 1, reflecting the importance of skills in the industry; Application example: The demand index for "machine learning" skills in the "artificial intelligence" industry is 0.92 (extremely high).

[0083] 7. Salary Standards and Relationships Relationship type: Applicable objects of salary standards (salary standards → composite entity); Meaning: Salary standards are tailored to specific industry, region, and job combination. Composite target: A target group jointly defined by three entities: industry, region, and job title; Key attribute: effective_date: indicates the timeliness of this salary standard; Application Example: Salary Standards for Data Scientists in the Fintech Industry in Shanghai in 2023 (P25: 35k, P50: 50k, P75: 75k).

[0084] S301. Parse the user intention using a pre-built rule engine and generate parameters required for graph queries based on the parsing results.

[0085] The rule engine includes: intention processing rules for mapping user intentions to parameters required for graph queries; for example: Urgency Judgment: Logic: If it is detected that the user uses words such as "urgently looking for" or "starting work immediately" ➔ then preferentially find the positions that can be hired the fastest and skip companies that require complex processes; Example: The user says "urgently seeking a Python job that can start next week" ➔ automatically filter out positions that require "three rounds of interviews" or "background checks"; Salary Expectation Translation: Logic: When the user says "high salary" and the resume writes the expected salary ➔ only find positions with a salary ≥ 90% of the expected value; When the user says "high salary" but does not write the expectation ➔ find positions that exceed 75% of the salary level of the same position in this region; Example: The resume writes an expected salary of 30K ➔ only show positions above 27K; only say "high salary" without writing a specific number ➔ show positions with a monthly salary > 25K in Java development positions in Shanghai.

[0086] Entity extension rules for extending synonyms for entities in the user intention; for example: Skill synonym connection: Logic: When seeing the resume write "proficient in TensorFlow" ➔ automatically supplement the search for positions that require "deep learning frameworks" and "neural networks"; Example: The user searches for "jobs that can write PyTorch" ➔ simultaneously search for positions marked as "deep learning engineer" and "AI R & D"; Location alias conversion: Logic: Recognize that "Magic City" = Shanghai, "Imperial City" = Beijing, and "Zhangjiang" is automatically associated with "Shanghai Pudong New Area"; Example: The user says "want to find an IT job in Zhangjiang" ➔ actually search for "Shanghai Pudong New Area + IT positions".

[0087] Intelligent extension rules for predicting implicit requirements; for example: If the resume shows that 3 consecutive jobs are < 1 year ➔ preferentially recommend enterprises with "high stability"; Programmers over 30 years old ➔ focus on recommending "technical management positions"; Example: A 28-year-old test engineer submits their resume ➔ The system automatically recommends "Automation Test Development" career transition positions.

[0088] S302. Use graph query language to perform graph query based on the parameters to obtain recruitment data.

[0089] Cypher is used as the core language of the knowledge graph query engine. Its declarative syntax supports traversal of complex graph structures and multi-condition filtering.

[0090] Generate a relationship chain based on the entity association logic in the intent, for example: The intent is to "recommend Java positions in Shanghai" → the path is (p:JobPosting)-[:requires]->(s:Skill{name:"Java"})-[:located_in]->(l:Location {city:"Shanghai"}).

[0091] The intent is "to analyze the demand for Python in the financial industry" → the path is (i:Industry{name:"finance"})-[:requires_skill]->(s:Skill {name:"Python"}).

[0092] For conditions combining multiple entities (such as industry + region + job title), use the logical operators AND / OR to concatenate the clauses.

[0093] To improve the query efficiency of large-scale knowledge graphs (node ​​size > 1 million), the following optimization techniques are adopted: Create composite indexes for frequently queried attributes, for example: CREATE INDEX FOR (p:JobPosting) ON (p.title, p.location.city,p.publish_date) CREATE INDEX FOR (s:Skill) ON (s.name, s.category).

[0094] Neo4j’s label index can be used to quickly locate entity types. For example, MATCH(p:JobPosting) can be filtered directly by the label index.

[0095] In one embodiment of the present invention, based on step S4, the following will provide a possible embodiment and describe its specific implementation in a non-limiting manner.

[0096] S401. Filter the recruitment data according to the key conditions in the user's intent.

[0097] Basic condition filtering: Structured parameters (such as location, salary, and skills) are extracted from user intent and mapped to entity attributes in the knowledge graph. For example, when a user queries "Java positions in Shanghai with a salary ≥ 20K", the system automatically generates filtering conditions: Geographic region: City = Shanghai; Salary standard: 25th percentile value ≥ 20000; Skill Requirements: Skill Name = Java; At the same time, filtering rules are supplemented by implicit conditions in user resumes (such as education level and years of work experience). For example, if "Master's degree" is found in a resume, the requirement of "Master's degree" will be automatically added.

[0098] Dynamic weighted filtering: A user profile weight matrix is ​​introduced (dimensions include skill matching, salary fit, and regional preference), and personalized filtering thresholds are calculated based on historical behavioral data (such as clicks and application records). For example, if a user has historically focused more on skill matching (weighting 40%), then positions with a skill matching score > 80% will be prioritized for retention.

[0099] Timeliness filtering: A time window is constructed based on the job posting date (publish_date) and expiration date (expiration_date) to filter expired jobs. For urgent needs containing keywords such as "urgently hiring", additional filtering is applied to positions with a recruitment process cycle > 7 days (judged by the company attribute "recruitment process days").

[0100] S402. Aggregate and statistically analyze the remaining data to obtain the number of job postings, salary distribution, and skill matching degree.

[0101] Display the job list in tabular form, including job title, company, location, salary, skill requirements, and match degree; generate industry distribution pie charts and salary histograms.

[0102] Job posting statistics: Grouped by industry, geographic region, company size, etc., using the HyperLogLog algorithm to estimate the cardinality to reduce memory consumption. Supports industry-level statistics (e.g., "Internet industry → E-commerce / Social media sub-sectors"), recursively expanded through the tree structure of industry entities (parent_industry attribute).

[0103] Salary distribution analysis: The 25th, 50th, and 75th quantiles of the standard salary entity were extracted, and a salary histogram was generated using kernel density estimation (KDE). The formula is as follows:

[0104] Where K is the Gaussian kernel function, and σ is optimized using the Silverman rule. Cross-regional salary comparisons are supported (e.g., Beijing vs. Shanghai), and Welch's t-test is used to test for statistical significance (p<0.05 is considered statistically significant).

[0105] Basic matching degree: Calculated using the Jaccard coefficient based on the user's resume skills set (R) and the job requirement skills set (P).

[0106] Weighted Adjustment: A skill weight matrix is ​​introduced (combining industry demand index and skill popularity), and weights are applied according to skill proficiency (beginner = 0.5, intermediate = 0.8, advanced = 1.0). The formula is as follows:

[0107] The core information is displayed in a responsive table format, plotted using the ECharts library, showing the percentage of job openings in each industry and supporting drill-down to sub-industries. Based on kernel density estimation results, different colors are used to distinguish regions (e.g., blue = Beijing, orange = Shanghai), with the horizontal axis representing salary range and the vertical axis representing job density.

[0108] In some embodiments, the recruitment information acquisition system may include multiple functional modules composed of computer program segments. The computer programs for each program segment in the recruitment information acquisition system may be stored in the memory of a computer device and executed by at least one processor to perform (see details). Figure 1 (Description) Function for obtaining recruitment information.

[0109] In this embodiment, the recruitment information acquisition system can be divided into multiple functional modules based on its functions, such as... Figure 2 As shown. The module referred to in this invention is a series of computer program segments that can be executed by at least one processor and perform a fixed function, and is stored in memory. In this embodiment, the functions of each module will be described in detail in subsequent embodiments.

[0110] The receiving module is used to receive the question text and extract the resume text associated with the question text; The recognition module is used to identify user intent based on the question text and resume text; The query module is used to retrieve relevant recruitment data from a pre-built knowledge graph based on the user's intent; The statistics module is used to perform statistical processing on the recruitment data based on the user's intent to obtain recruitment information; The entities in the knowledge graph include job titles, companies, skills, industries, geographical regions, and salary standards.

[0111] Figure 3 The recruitment information acquisition method provided in the embodiments of this application can be applied to a device. Those skilled in the art will understand that the device structure involved in the embodiments of this invention does not constitute a limitation on the device. A device may include more or fewer components than illustrated, or combine certain components, or have different component arrangements. In the embodiments of this invention, the device includes, but is not limited to, laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The device may also represent various forms of mobile devices, such as personal digital processors, cellular phones, smartphones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the embodiments of this application described and / or claimed herein.

[0112] The device 300 may include a processor 310, a memory 320, and a communication unit 330. These components communicate via one or more buses. Those skilled in the art will understand that the server structure shown in the figure does not constitute a limitation of the present invention. It may be a bus topology or a star topology, and may include more or fewer components than shown, or combine certain components, or have different component arrangements.

[0113] The memory 320 can be used to store execution instructions of the processor 310. The memory 320 can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk. When the execution instructions in the memory 320 are executed by the processor 310, the device 300 is able to perform some or all of the steps in the above method embodiments.

[0114] The processor 310 serves as the control center of the storage device, connecting various parts of the electronic device via various interfaces and lines. It executes software programs and / or modules stored in the memory 320, and calls data stored in the memory to perform various functions of the electronic device and / or process data. The processor can be composed of integrated circuits (ICs), such as a single packaged IC or multiple packaged ICs with the same or different functions connected together. For example, the processor 310 may consist only of a central processing unit (CPU). In this embodiment of the invention, the CPU may have a single processing core or include multiple processing cores.

[0115] The communication unit 330 is used to establish a communication channel, enabling the storage device to communicate with other devices. It can receive user data sent by other devices or send user data to other devices.

[0116] The present invention also provides a computer storage medium, wherein the computer storage medium may store a program, which, when executed, may include some or all of the steps provided in the embodiments of the present invention. The storage medium may be a magnetic disk, an optical disk, read-only memory (ROM), or random access memory (RAM), etc.

[0117] Those skilled in the art will clearly understand that the techniques in the embodiments of the present invention can be implemented using software plus necessary general-purpose hardware platforms. Based on this understanding, the technical solutions in the embodiments of the present invention, or the parts that contribute to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium such as a USB flash drive, a portable hard drive, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk, or any other medium capable of storing program code. It includes several instructions to cause a computer device (which may be a personal computer, a server, or a second device, network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention.

[0118] The same or similar parts between the various embodiments in this specification can be referred to mutually. In particular, the device embodiments are basically similar to the method embodiments, so the description is relatively simple, and the relevant parts can be referred to the description in the method embodiments.

[0119] In the embodiments provided by this invention, it should be understood that the disclosed systems and methods can be implemented in other ways. For example, the system embodiments described above are merely illustrative. For instance, the division of modules is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple modules 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 modules may be electrical, mechanical, or other forms.

[0120] The modules described as separate components may or may not be physically separate. The components shown as modules may or may not be physical modules; that is, they may be located in one place or distributed across multiple network modules. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs.

[0121] In addition, the functional modules in the various embodiments of the present invention can be integrated into one processing module, or each module can exist physically separately, or two or more modules can be integrated into one module.

[0122] Although the present invention has been described in detail with reference to the accompanying drawings and preferred embodiments, the present invention is not limited thereto. Various equivalent modifications or substitutions can be made to the embodiments of the present invention by those skilled in the art without departing from the spirit and essence of the invention, and such modifications or substitutions should all be within the scope of the present invention. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should also be covered within the protection scope of the present invention.

Claims

1. A method for obtaining recruitment information, characterized in that, include: Receive the question text and extract the resume text associated with the question text; Identify user intent based on question text and resume text; Recruit relevant recruitment data from a pre-built knowledge graph based on the user's intent; Based on the user's intent and by statistically processing the recruitment data, recruitment information is obtained; The entities in the knowledge graph include job titles, companies, skills, industries, geographical regions, and salary standards.

2. The method according to claim 1, characterized in that, Identify user intent based on question text and resume text, including: A supervised learning algorithm is used to train an intent classifier, and the trained intent classifier is combined with a predefined intent labeling system to identify the intent of the question text. Use named entity recognition technology to extract key entities from question text and resume text; The intent is modified using the key entity.

3. The method according to claim 1, characterized in that, Methods for constructing knowledge graphs include: Collect publicly available job postings and extract entities from them. Entity types include job title, company, skills, industry, geographic region, and salary standards. Extract the relationships between entities from the recruitment text, including the relationship between position and skills, position and company, company and industry, company and geographic location, skills and skills, industry and skills, and salary standards and entity combinations, where entity combinations include industry, geographic location and position; Add attributes to entities and relations, construct triples based on entities and relations, and build a knowledge graph based on triples.

4. The method according to claim 3, characterized in that, Based on the user intent, relevant recruitment data is obtained from a pre-built knowledge graph, including: The user intent is parsed using a pre-built rules engine, and the parameters required for graph queries are generated based on the parsing results. The recruitment data is obtained by performing a graph query based on the parameters using a graph query language.

5. The method according to claim 4, characterized in that, The rule engine includes: Intent processing rules are used to map user intents to the parameters required for graph queries; Entity expansion rules are used to expand synonyms for entities in user intents; Intelligent expansion rules are used to predict implicit requirements.

6. The method according to claim 1, characterized in that, Based on the user intent, statistical processing of recruitment data is performed, including: Filter the recruitment data according to key conditions in the user's intent; The remaining data is aggregated and statistically analyzed to obtain the number of job postings, salary distribution, and skill matching degree.

7. The method according to claim 6, characterized in that, The method further includes: The job list is displayed in tabular form, including job title, company, location, salary, skill requirements, and match rate. Generate industry distribution pie charts and salary histograms.

8. A recruitment information acquisition system, characterized in that, include: The receiving module is used to receive the question text and extract the resume text associated with the question text; The recognition module is used to identify user intent based on the question text and resume text; The query module is used to retrieve relevant recruitment data from a pre-built knowledge graph based on the user's intent; The statistics module is used to perform statistical processing on the recruitment data based on the user's intent to obtain recruitment information; The entities in the knowledge graph include job titles, companies, skills, industries, geographical regions, and salary standards.

9. A device, characterized in that, include: The memory is used to store the recruitment information retrieval program; A processor is configured to implement the steps of the recruitment information acquisition method as described in any one of claims 1-7 when executing the recruitment information acquisition program.

10. A computer-readable storage medium storing a computer program, characterized in that, The readable storage medium stores a recruitment information acquisition program, which, when executed by a processor, implements the steps of the recruitment information acquisition method as described in any one of claims 1-7.

Images