Employment post recommendation method based on human resource database

Abstract

The invention relates to the technical field of human resource management and information, and discloses an employment post recommendation method based on a human resource database, and the method comprises the following steps: S1, data collection; s2, data processing; s3, keyword extraction; s4, performing similarity calculation; s5, post filtering and sorting; s6, post recommendation; and S7, effect evaluation: through automatic data processing and similarity calculation, resumes most matched with post demands can be quickly screened out, the workload of recruiters is reduced, accurate matching degree evaluation is beneficial to realizing optimal configuration of talents and posts, employee loss and recruitment cost waste caused by mismatching are reduced, and the efficiency of recruitment is improved. The recruitment will and the matching degree of the job seekers can be improved through accurate recommendation, so that the recruitment success rate and efficiency are improved, more high-quality job seekers are attracted through efficient post recommendation and accurate matching, and the talent reserve and competitiveness of enterprises are enhanced.

Description

Technical Field

[0001] The present invention relates to the field of human resource management and information technology, and in particular to a method for recommending job positions based on a human resource database. Background Art

[0002] Human resource management is an upgrade of personnel management. It refers to a series of activities that are carried out under the guidance of economics and humanistic thinking, through recruitment, selection, training, compensation and other management forms to effectively use relevant human resources inside and outside the organization to meet the current and future development needs of the organization and ensure the realization of organizational goals and the maximization of member development. It is the whole process of predicting the human resource needs of the organization and making human resource demand plans, recruiting and selecting personnel and conducting effective organization, assessing performance, paying remuneration and conducting effective incentives, and combining organizational and individual needs for effective development in order to achieve optimal organizational performance.

[0003] With the rapid development of the Internet, the amount of recruitment information is increasing day by day. Job seekers are faced with the difficulty of selecting suitable job opportunities from the massive job information, and corporate recruiters are also faced with the challenge of how to efficiently and accurately find talents that meet job requirements. Traditional human resource recruitment methods often rely on job seekers to search for job information on their own or job seekers to submit resumes after companies publish recruitment information. This method has problems such as information asymmetry and low matching efficiency. Therefore, a job position recommendation method based on human resource database is proposed. Summary of the invention

[0004] The purpose of the present invention is to provide a job recommendation method based on a human resources database to solve the problems raised in the above background technology.

[0005] To achieve the above object, the present invention provides the following technical solution: a method for recommending job positions based on a human resources database, comprising the following steps:

[0006] S1. Data collection;

[0007] S2, data processing;

[0008] S3, keyword extraction;

[0009] S4, similarity calculation;

[0010] S5. Position filtering and sorting;

[0011] S6. Job recommendation;

[0012] S7. Effect evaluation.

[0013] Preferably, in S1, the data collection is collecting job seekers' resume data and job position information from a human resources database;

[0014] The resume data includes education background, work experience, skills and certifications;

[0015] The job information includes job requirements, skill requirements, work location and salary range.

[0016] Preferably, in S2 above, the data processing includes removing duplicate data, processing missing values ​​and standardizing data formats.

[0017] Preferably, in S3, the keyword extraction is to extract keywords of the job information and resume information using natural language processing technology, and assign weights according to the importance and relevance of the keywords to form a job vector and a resume vector.

[0018] Preferably, in S4, the similarity calculation is to calculate the similarity between the resume vector and the position vector, and select the resume that best matches the position information. The similarity (T) calculation formula is:

[0019]

[0020] Among them, weight i is the weight of the ith keyword, match i is the matching degree between the resume and the position on the ith keyword, and n is the total number of keywords.

[0021] Preferably, in S5, the job filtering and sorting is to sort the matching resumes according to the similarity calculation results, and can selectively filter out resumes that do not meet specific conditions.

[0022] As a preferred method, the above-mentioned formula for calculating the matching degree between individuals and positions is:

[0023]

[0024] Where n is the total number of features, individual features i is the value of the ith individual characteristic, and the weight of the position characteristic i is the corresponding job characteristic weight.

[0025] Preferably, in S6, the job recommendation is to sort all the jobs according to the calculated similarities and recommend the job with the highest similarity to the job seeker.

[0026] Preferably, in S7, the effect evaluation uses recommendation accuracy (P), mean reciprocal rank (MRR), and recommendation satisfaction to evaluate the job recommendation results;

[0027] The calculation formula for recommendation accuracy (P) is:

[0028]

[0029] The calculation formula for mean reciprocal rank (MRR) is:

[0030]

[0031] Among them, U is the set of job seekers, M(u) is the recommendation list of user u, J(u) is the actual contracted position of user u, and r(u) is the position number of J(u) in M(u).

[0032] Preferably, the above recommendation satisfaction is collected through a questionnaire survey or user feedback.

[0033] Compared with the prior art, the present invention adopts the above technical solution and has the following technical effects:

[0034] 1. Collect data through multiple channels such as human resources databases, social media, and professional recruitment websites to increase the diversity and richness of data, improve the comprehensiveness and accuracy of recommendations, clean, process, and standardize data to ensure that data from different sources and types are comparable in similarity calculations, accurately match job seekers' skills and interests with job requirements, reduce the time and energy job seekers spend sifting through massive recruitment information, provide job seekers with personalized job recommendations, and improve job search satisfaction and efficiency.

[0035] 2. Through automated data processing and similarity calculation, it is possible to quickly screen out resumes that best match job requirements, reducing the workload of recruiters. Accurate matching assessment helps to achieve optimal allocation of talents and positions, reducing employee turnover and waste of recruitment costs due to mismatch. Accurate recommendations can increase job seekers' willingness to apply and matching, thereby improving the success rate and efficiency of recruitment. Efficient job recommendations and accurate matching can attract more high-quality job seekers and enhance the company's talent pool and competitiveness.

[0036] 3. Through quantitative indicators such as recommendation accuracy and average ranking inverse, we can objectively evaluate the recommendation effect and provide data support for recruitment decisions. The effect evaluation can reveal potential problems in the recommendation algorithm or process, make targeted improvements and optimizations, establish a feedback mechanism to collect user opinions, continuously optimize and iterate the recommendation algorithm, and improve the recommendation effect and service quality. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0038] Figure 1 is a flow chart of the method of the present invention. DETAILED DESCRIPTION

[0039] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0040] It should be noted that the structures, proportions, sizes, etc. illustrated in the drawings of this specification are only used to match the contents disclosed in the specification for people familiar with this technology to understand and read, and are not used to limit the conditions under which this application can be implemented. Therefore, they have no substantive technical significance. Any structural modification, change in proportional relationship or adjustment of size should still fall within the scope of the technical content disclosed in this application without affecting the effects and purposes that can be achieved by this application.

[0041] Example

[0042] See also Figure 1 The present invention provides a technical solution: a method for recommending job positions based on a human resources database, comprising the following steps:

[0043] S1. Data collection;

[0044] Data collection involves collecting applicants' resume data and job information from human resource databases. In addition to human resource databases, data can also be collected from social media, professional recruitment websites, applicants' personal websites, and other channels to increase data diversity and richness, implement a real-time data update mechanism, and ensure the latest status of job information and applicants' resumes.

[0045] As the main data source, the human resources database regularly extracts job applicant resume data and position information from the company's internal human resources database. The data usually includes standardized fields and content that has undergone preliminary review.

[0046] Use API interfaces or crawler technology to capture job seekers' professional background information, skill tags, educational experience, etc. from social media platforms, while paying attention to industry trends and job postings.

[0047] Establish cooperation with mainstream recruitment websites or use public APIs to obtain the latest job information and job seekers' resumes in real time. Websites usually provide rich screening and search functions to help accurately locate target data.

[0048] Scheduled tasks: Set up scheduled tasks, such as automatically pulling the latest data from various channels every morning to ensure the timeliness of the data;

[0049] Incremental update: For data sources that support incremental update, incremental data pulling is preferred to reduce the data processing burden;

[0050] Change detection: For data sources that do not support direct incremental updates, compare historical data with current data to identify the changed parts and update them;

[0051] Resume data includes education background, work experience, skills and certificates. Resume data mainly includes the following parts:

[0052] Educational background: including academic qualifications, degree, graduation school, major, etc.

[0053] Work experience: previous companies, positions, job responsibilities, work achievements, etc.

[0054] Skills: professional skills, language skills, soft skills (such as teamwork, leadership, etc.).

[0055] Certificates: various professional qualification certificates, skill certificates, etc.

[0056] Self-evaluation and job-seeking intention: the applicant's evaluation of himself and his expected position type, work location, salary requirements, etc.;

[0057] Job information includes job requirements, skill requirements, work location and salary range. Job information mainly includes the following parts:

[0058] Job requirements: basic requirements for the position, such as education, work experience, professional skills, etc.

[0059] Skill requirements: A list of skills required for a specific position, which may include both essential and premium skills.

[0060] Work Location: The specific work location or acceptable work area.

[0061] Salary range: The salary range or specific salary level for a position.

[0062] Company information: including company profile, corporate culture, development prospects, etc. This information helps job seekers understand the company background.

[0063] S2; data processing;

[0064] Data processing includes removing duplicate data, handling missing values ​​and standardizing data formats. In addition to removing duplicate data and handling missing values, outlier detection and correction can also be performed to ensure data accuracy and ensure that data from different sources and types are comparable in similarity calculations.

[0065] De-duplicate data: Identify and delete duplicate records using unique identifiers, such as ID numbers, email addresses, or internal IDs, to ensure the purity of the data set;

[0066] Handling missing values: For missing data fields, adopt appropriate filling strategies according to the nature and distribution of the data, such as mean filling, median filling, mode filling, filling based on model prediction, or retaining them as special values;

[0067] Statistical methods: Detect outliers using statistical principles, such as Z-score, IQR interquartile range method, etc. For detected outliers, decide whether to delete, replace with normal values, or retain and mark them based on business logic.

[0068] Model-based approaches: Using clustering algorithms (such as DBSCAN) or machine learning techniques such as isolation forests to identify anomalous data points can automatically identify unusual patterns in the data, rather than just based on the statistical distribution of a single variable.

[0069] Domain knowledge: Combined with professional knowledge in the field of human resources, manually review and correct certain obviously unreasonable data points, such as positions that are several times higher than the industry average salary level, unreasonable correspondence between academic qualifications and positions, etc.

[0070] Numerical data: For numerical data (such as salary, age, etc.), normalization or standardization methods are used to scale the data to the same scale for easy comparison and analysis.

[0071] Text data: For text data, such as educational background and work experience description, preprocessing operations such as word segmentation, stop word removal, and stem extraction are performed, and it may be converted into a numerical representation.

[0072] Date and time data: Dates and times are formatted uniformly and may be converted to the number of days or months since a base date to facilitate time series analysis.

[0073] Unified data format: Ensure that data from different channels follow the same format and naming conventions, such as unified currency units, time units, etc.

[0074] Standardized field naming: Create a data dictionary to clarify the meaning of each field and the allowed value range, ensuring consistency in field naming and meaning across data from different sources.

[0075] Data conversion: Based on analysis requirements, data from different sources and types are converted into a format suitable for analysis. For example, text-based skill labels can be converted into Boolean features, or different levels of education can be converted into numerical scores.

[0076] S3, keyword extraction;

[0077] Keyword extraction uses natural language processing technology to extract keywords from job information and resume information, and assigns weights based on the importance and relevance of the keywords to form job vectors and resume vectors. Combined with professional knowledge in the human resources field, a domain dictionary or thesaurus is constructed to guide keyword extraction and weight assignment.

[0078] Text preprocessing: First, preprocess the job description and resume text, including word segmentation, punctuation removal, stop word filtering, word form restoration or part-of-speech tagging, etc., to reduce noise and improve the accuracy of subsequent processing.

[0079] Keyword identification: Use NLP models or algorithms such as TF-IDF, TextRank, Word2Vec, BERT, etc. to identify keywords in text. TF-IDF focuses on counting word frequency and document frequency, while deep learning models such as Word2Vec and BERT can capture the semantic information of words, which helps to more accurately identify context-related keywords.

[0080] Statistical method: Preliminary weights are assigned based on the TF-IDF value or the frequency of occurrence of a word in the text. The higher the TF-IDF value, the more important the word is in the current text, but it is relatively rare in the entire dataset, so it is more likely to be a keyword for the text.

[0081] Semantic-based method: Use models such as Word2Vec and BERT to calculate the semantic similarity between keywords, regard keywords with high similarity as mutually reinforcing, and adjust their weights appropriately.

[0082] Involvement of domain experts: Combining professional knowledge in the field of human resources, domain experts will evaluate and adjust keywords and their weights.

[0083] Collection and organization: Collect professional terms and common words related to human resources from various channels such as professional books, industry reports, recruitment websites, internal corporate materials, etc. to form an initial vocabulary.

[0084] Classification and labeling: Classify and label the words in the vocabulary, such as job category, skill category, industry field, etc., to facilitate subsequent keyword extraction and weight allocation.

[0085] Dynamic updates: As the industry develops and changes, the field dictionary or thesaurus is regularly updated to add new professional terms and vocabulary and delete outdated or no longer commonly used vocabulary.

[0086] S4, similarity calculation;

[0087] The similarity calculation is to calculate the similarity between the resume vector and the job vector, and select the resume that best matches the job information. The similarity calculation formula (T) is:

[0088]

[0089] Among them, weight i is the weight of the ith keyword, match i is the matching degree between the resume and the position on the i-th keyword, n is the total number of keywords, if both the resume and the position appear on the i-th keyword, then match i =1, otherwise match i = 0, ignoring the specific position and contextual information of the keywords in the text, which is a simplified form of "binary similarity" or "Boolean similarity".

[0090] S5. Position filtering and sorting;

[0091] Job filtering and sorting is to sort the matching resumes based on the similarity calculation results, and selectively filter out resumes that do not meet specific conditions.

[0092] The formula for calculating the matching degree between individuals and positions is:

[0093]

[0094] Where n is the total number of features, individual features i is the value of the ith individual characteristic, and the weight of the position characteristic i is the corresponding job characteristic weight.

[0095] Hard condition filtering: Set hard filtering conditions according to job requirements, such as education, years of work experience, specific skills certificates, etc. Only resumes that meet these conditions can enter the next round of screening.

[0096] Soft condition filtering: In addition to hard conditions, you can also set some soft filtering conditions, such as the similarity between the job seeker's work experience and the job requirements, the match between the job seeker's salary expectations and the job salary range, etc., which can be evaluated by setting thresholds or scoring mechanisms.

[0097] Customized filtering conditions: allows recruiters to customize filtering conditions based on actual needs, such as company culture match, job seekers' personality traits, etc.

[0098] Multi-feature fusion: Match personal features with job features in a many-to-one or many-to-many manner, considering not only the matching degree of a single feature but also the comprehensive matching degree of multiple features. For example, the matching degree between the applicant’s education, work experience, skills and other features and the job requirements can be considered at the same time.

[0099] Dynamic weight adjustment: The weight of job characteristics can be dynamically adjusted based on factors such as market demand, corporate strategy, and job urgency.

[0100] Nonlinear calculation: The matching degree calculation formula can introduce nonlinear factors to better reflect the actual matching situation. For example, machine learning models such as neural networks or decision trees can be used to calculate the matching degree, which can capture the complex relationships and nonlinear effects between features.

[0101] Recommendations based on user portraits: Build user portraits based on job seekers’ historical behaviors, interest preferences, career plans, and other information, and recommend positions that are highly matched with their personal characteristics.

[0102] Intelligent filtering and prediction: Use machine learning algorithms to analyze a large amount of resume and job data to predict job seekers’ future career development and job demand trends. Based on these prediction results, intelligently filter out resumes and positions that do not meet job seekers’ career plans or future job requirements.

[0103] Feedback mechanism: Establish a feedback mechanism to collect feedback from recruiters and job seekers, and continuously optimize the filtering and sorting algorithms based on the feedback results. Improve the accuracy and efficiency of job filtering and sorting through continuous learning and iterative improvement.

[0104] S6. Job recommendation;

[0105] Job recommendation is to sort all jobs according to the calculated similarity and recommend the job with the highest similarity to the job seeker.

[0106] Through accurate job recommendations, job seekers can find positions that match their skills and interests more quickly, reduce the time and energy spent on sifting through massive amounts of recruitment information, and improve job search satisfaction and efficiency.

[0107] For companies, job recommendations can accurately identify potential suitable candidates, reduce the screening and processing of invalid resumes, and enable recruiters to focus more on communicating with and evaluating high-quality candidates, thereby improving overall recruitment efficiency.

[0108] Through algorithm matching, we can more accurately judge the fit between job seekers and positions, which will help achieve optimal allocation of talents and reduce employee turnover and waste of recruitment costs due to job mismatch.

[0109] S7. Effect evaluation.

[0110] The effectiveness evaluation uses recommendation accuracy (P), mean reciprocal rank (MRR), and recommendation satisfaction to evaluate the job recommendation results;

[0111] The calculation formula for recommendation accuracy (P) is:

[0112]

[0113] The calculation formula for mean reciprocal rank (MRR) is:

[0114]

[0115] Among them, U is the set of job seekers, M(u) is the recommendation list of user u, J(u) is the actual contracted position of user u, and r(u) is the position number of J(u) in M(u).

[0116] Recommendation satisfaction is collected through questionnaires or user feedback. Quantitative indicators such as recommendation accuracy (P) and mean rank reciprocal (MRR) can be used to objectively evaluate the effectiveness of job recommendations and provide data support for subsequent algorithm optimization and strategy adjustment. Effect evaluation can reveal potential problems in recommendation algorithms or processes, such as poor recommendation effects for certain types of jobs and under-utilization of certain key features. Collecting recommendation satisfaction data through questionnaires or user feedback can directly understand the feelings and opinions of job seekers and recruiters on job recommendation services, and provide a reference for further improving service quality. Effect evaluation is an ongoing process. With the accumulation of data and optimization of algorithms, the recommendation effect will continue to improve.

[0117] In summary, data is collected through multiple channels such as human resources databases, social media, professional recruitment websites, etc., to increase the diversity and richness of data, improve the comprehensiveness and accuracy of recommendations, clean, process and standardize data, ensure that data from different sources and types are comparable in similarity calculations, accurately match job seekers' skills and interests with job requirements, reduce the time and energy job seekers spend sifting through massive recruitment information, provide job seekers with personalized job recommendations, and improve job search satisfaction and efficiency.

[0118] Through automated data processing and similarity calculation, we can quickly screen out resumes that best match job requirements, reducing the workload of recruiters. Accurate matching assessment helps to achieve optimal allocation of talents and positions, reducing employee turnover and waste of recruitment costs due to mismatch. Precise recommendations can increase job seekers' willingness to apply and matching, thereby improving the success rate and efficiency of recruitment. Efficient job recommendations and accurate matching can attract more high-quality job seekers and enhance the company's talent pool and competitiveness.

[0119] Through quantitative indicators such as recommendation accuracy and average ranking inverse, the recommendation effect can be objectively evaluated to provide data support for recruitment decisions. The effect evaluation can reveal potential problems in the recommendation algorithm or process, make targeted improvements and optimizations, establish a feedback mechanism to collect user opinions, continuously optimize and iterate the recommendation algorithm, and improve the recommendation effect and service quality.

[0120] It will be appreciated by those skilled in the art that the features described in the various embodiments and / or claims of the present invention may be combined and / or combined in various ways, even if such combinations and / or combinations are not explicitly described in the present invention. In particular, the features described in the various embodiments and / or claims of the present invention may be combined and / or combined in various ways without departing from the spirit and teachings of the present invention. All of these combinations and / or combinations fall within the scope of the present invention.

Claims

1. A job recommendation method based on a human resources database, characterized in that: The following steps are involved: S1. Data collection; S2, data processing; S3, keyword extraction; S4, similarity calculation; S5. Position filtering and sorting; S6. Job recommendation; S7. Effect evaluation.

2. The method for recommending job positions based on a human resources database according to claim 1, characterized in that: In S1, the data collection is collecting job seeker resume data and position information from a human resources database; The resume data includes education background, work experience, skills and certifications; The job information includes job requirements, skill requirements, work location and salary range.

3. The method for recommending job positions based on a human resources database according to claim 1, characterized in that: In S2, the data processing includes removing duplicate data, handling missing values ​​and standardizing data formats.

4. The method for recommending job positions based on a human resources database according to claim 1, characterized in that: In S3, the keyword extraction is to extract keywords of the job information and resume information using natural language processing technology, and assign weights according to the importance and relevance of the keywords to form a job vector and a resume vector.

5. The method for recommending job positions based on a human resources database according to claim 1, characterized in that: In S4, the similarity calculation is to calculate the similarity between the resume vector and the position vector, and select the resume that best matches the position information. The similarity (T) calculation formula is: Among them, weight i is the weight of the ith keyword, match i is the matching degree between resume and position on the ith keyword, and n is the total number of keywords.

6. The method for recommending job positions based on a human resources database according to claim 5, characterized in that: In S5, the job filtering and sorting is to sort the matching resumes according to the similarity calculation results, and selectively filter out resumes that do not meet specific conditions.

7. The method for recommending job positions based on a human resources database according to claim 6, characterized in that: The formula for calculating the matching degree between the individual and the position is: Where n is the total number of features, individual features i is the value of the ith individual characteristic, and the position characteristic weight i is the corresponding job characteristic weight.

8. The method for recommending job positions based on a human resources database according to claim 7, characterized in that: In S6, the job recommendation is to sort all the jobs according to the calculated similarities and recommend the job with the highest similarity to the job seeker.

9. The method for recommending job positions based on a human resources database according to claim 1, characterized in that: In S7, the effect evaluation uses recommendation accuracy (P), mean reciprocal rank (MRR), and recommendation satisfaction to evaluate the job recommendation results; The calculation formula for recommendation accuracy (P) is: The calculation formula for mean reciprocal rank (MRR) is: Among them, U is the set of job seekers, M(u) is the recommendation list of user u, J(u) is the actual contracted position of user u, and r(u) is the position number of J(u) in M(u).

10. The method for recommending job positions based on a human resources database according to claim 9, characterized in that: The recommendation satisfaction is collected through questionnaires or user feedback.

Images