Efficient resume intelligent recommendation algorithm

Abstract

The invention relates to the related field of resume algorithms, discloses an efficient intelligent resume recommendation algorithm, and provides an intelligent resume recommendation system based on a collaborative filtering algorithm and fusing a deep learning model and a data driving technology, the system can accurately and efficiently resume matched with enterprise post requirements by integrating the advantages of different algorithms, and meanwhile, the resume recommendation efficiency is improved. The self-learning and dynamic optimization capabilities of the system ensure that the system can continuously adapt to changing recruitment requirements, meanwhile, the recruitment process is simplified, the labor cost is reduced, and the quality and speed of recruitment work are improved.

Description

Technical Field

[0001] The present invention relates to resume algorithm related fields, and in particular to an efficient resume intelligent recommendation algorithm. Background Art

[0002] In the current talent recruitment market, due to the large number of job seekers, companies and recruitment agencies are faced with a large amount of resume screening during recruitment. In existing specific scenarios, most of them are carried out by manual screening, and in some cases where online material submission is used, preliminary screening of demand restrictions will be carried out through recommendation procedures to reduce the resume screening workload of relevant staff.

[0003] Traditional manual screening methods are inefficient and easily affected by subjective factors. Some existing resume recommendation programs are mostly based on keyword matching and a small number of rule specifications. They have problems such as low accuracy, insufficient personalization of recommendation results, and inability to accurately recommend complex job requirements. They lack a deep understanding and prediction of user behavior, resulting in low recommendation quality and an inability to dynamically adapt to changes in corporate job requirements. Summary of the invention

[0004] The purpose of the present invention is to provide an efficient resume intelligent recommendation algorithm to solve the problems raised in the above background technology.

[0005] To achieve the above object, the present invention provides the following technical solutions:

[0006] An efficient resume intelligent recommendation algorithm, including:

[0007] Collaborative filtering based on filtering both parties is used to obtain resume features and analyze job requirements;

[0008] Based on the deep learning model, feature text extraction is performed on the unstructured text information of resume content and job requirements to obtain deep implicit features in the text information;

[0009] Obtain market change data and enterprise demand data, and adjust the weight of the recommendation algorithm based on data-driven data to achieve dynamic recommendation strategy adjustment;

[0010] According to the adjusted recommendation strategy, the resume requirements are matched, a number of matching degree data corresponding to the resumes are obtained, and resume rankings are generated in descending order based on the matching degrees and output visually. The matching degrees are used to characterize the degree of conformity to the positions.

[0011] As a further solution of the present invention: the steps of collaborative filtering processing based on filtering both parties specifically include:

[0012] Based on user collaborative filtering, the company's historical recruitment records are obtained, and the positions are classified according to their similarity. Data matching is performed on similar positions under the same category to obtain the corresponding recruitment preferences, and resume screening is performed based on the recruitment preferences.

[0013] Based on project collaborative filtering, according to each job project, the job demand information is analyzed to obtain the feature matrix of the corresponding job. The feature matrix is ​​used for similarity matching with resumes. The demand information includes job description, skill requirements and past recruitment characteristics.

[0014] As a further solution of the present invention: the step of extracting feature text from the unstructured text information of resume content and job requirements based on the deep learning model to obtain deep implicit features in the text information specifically includes:

[0015] Obtaining unformatted information in the resume, and performing semantic feature extraction on text data of the unformatted information based on a deep learning model to generate a deep feature vector corresponding to the resume, wherein the unformatted information includes work experience, educational background, and skill description;

[0016] Unformatted information in the job description is obtained, and features are extracted from the unformatted information based on a deep learning model to generate a deep feature vector corresponding to the job. The unformatted information includes job requirements, skill requirements, and experience requirements.

[0017] As a further solution of the present invention: the step of obtaining market change data and enterprise demand data and adjusting the weight of the recommendation algorithm based on the data to achieve dynamic recommendation strategy adjustment specifically includes:

[0018] Access the recruitment data platform to obtain job information and resume data in real time, evaluate and update its feature data and calculate the matching degree according to the preset update cycle;

[0019] Obtain matching feedback data, and optimize the recommendation algorithm based on the matching feedback data, wherein the matching feedback data is used to characterize the recruitment status of the corresponding position, and the recruitment status includes resume information of the recruitment object;

[0020] Use the preset evaluation cycle to obtain corporate preferences and market trends within the historical period, obtain the changes in demand weights of different companies and different positions, and based on this, perform multi-level recommendation weight optimization to achieve recommendation strategy adjustment.

[0021] As a further solution of the present invention: the multi-level recommendation weights are specifically used to implement the following steps:

[0022] Recommend commonly used resumes for similar positions based on collaborative filtering algorithms, based on HR's historical behavior and job similarity;

[0023] By deeply matching resumes with job characteristics, we recommend resumes that are highly consistent with job descriptions;

[0024] Based on HR's operational behavior and feedback information, the recommendation strategy is dynamically adjusted to optimize the recommendation results.

[0025] As a further solution of the present invention: it also includes an output management step, specifically including:

[0026] Selecting a specified number of resumes that meet the current job requirements based on the matching ranking of the resumes, marking the resumes based on the matching scores, and visually outputting them;

[0027] The corresponding resume information is obtained, and the information text matching the position is displayed and marked according to the matching degree. The display mark is used to facilitate quick positioning when the enterprise views it.

[0028] As a further solution of the present invention: it also includes the steps of:

[0029] When the recommended resumes output in a visual manner do not obtain selection feedback information from the enterprise, the system optimizes the recommendation algorithm based on the enterprise screening and issues an alarm, which is used to inform the need for adjustment of job requirements.

[0030] Compared with the prior art, the beneficial effects of the present invention are: it provides a resume intelligent recommendation system based on collaborative filtering algorithm, fusion deep learning model and data-driven technology. By integrating the advantages of different algorithms, the system can accurately and efficiently recommend resumes that match the company's job requirements. At the same time, its self-learning and dynamic optimization capabilities ensure that the system can continuously adapt to changing recruitment needs, while simplifying the recruitment process, reducing labor costs, and improving the quality and speed of recruitment work. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] Figure 1 This is a flowchart of an efficient resume intelligent recommendation algorithm.

[0032] Figure 2 A flowchart of the collaborative filtering process in an efficient resume intelligent recommendation algorithm. DETAILED DESCRIPTION

[0033] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0034] The specific implementation of the present invention is described in detail below in conjunction with specific embodiments.

[0035] like Figure 1 The invention provides an efficient resume intelligent recommendation algorithm according to an embodiment of the present invention, comprising the following steps:

[0036] S10, based on collaborative filtering processing between filtering parties, to obtain resume features and analyze job requirements;

[0037] S20, extracting feature text from unstructured text information of resume content and job requirements based on a deep learning model to obtain deep implicit features in the text information;

[0038] S30, obtaining market change data and enterprise demand data, and based on this, adjusting the weight of the data-driven recommendation algorithm to achieve dynamic adjustment of the recommendation strategy;

[0039] S40, matching resumes to requirements according to the adjusted recommendation strategy, obtaining a number of matching degree data corresponding to the resumes, generating resume rankings in descending order based on the matching degrees and performing visual output, wherein the matching degrees are used to characterize the degree of conformity to the position.

[0040] In this embodiment, a resume intelligent recommendation system based on collaborative filtering algorithm, fusion deep learning model and data-driven technology is provided. By integrating the advantages of different algorithms, the system can accurately and efficiently recommend resumes that match the job requirements of the enterprise. At the same time, its self-learning and dynamic optimization capabilities ensure that the system can continuously adapt to changing recruitment needs, while simplifying the recruitment process, reducing labor costs, and improving the quality and speed of recruitment work; among them, collaborative filtering can mine implicit preferences based on the company's historical recruitment behavior, while the deep learning model can extract high-level semantic features from resumes and job descriptions. The two complement each other to ensure that the resumes recommended by the system are accurately matched with the company's job requirements, reducing the recommendation and screening time of irrelevant resumes ; Specifically, the technical architecture features required for the implementation of this embodiment are: Intelligent resume recommendation system architecture: the overall architecture of the resume recommendation system based on the combination of collaborative filtering algorithm and deep learning, including the entire process design of data collection, processing, recommendation and feedback optimization; Collaborative filtering algorithm module: This module is based on HR's recruitment behavior and historical operations, and is the core algorithm for resume recommendation through implicit preference mining mechanism; Deep learning model module: A deep learning model that extracts high-level semantic features from resumes and job descriptions, especially a processing method for unstructured data (such as text descriptions); Feedback learning mechanism: Through HR's operational feedback, the recommendation algorithm is dynamically optimized, so that the system can adaptively adjust the recommendation strategy of the functional module; Multi-level recommendation strategy : The recommendation strategy adopted by the present invention is based on the combination of collaborative filtering, deep learning and real-time feedback to improve the accuracy and diversity of recommendation results; Personalized recommendation mechanism: A mechanism for generating personalized recommendation lists based on HR's historical behavior and job requirements, including personalized parameter adjustment and dynamic optimization algorithms; Advanced semantic analysis technology for matching resumes with job descriptions: A method in which the system extracts semantic features from natural language texts and deeply matches the skills and experience in resumes with job requirements; The main protected technical feature points are: The combined application of collaborative filtering algorithms and deep learning models: This is the core technical breakthrough of the present invention. By combining the preference mining of collaborative filtering with the advanced semantic feature extraction of deep learning, the resume recommendation is improved. The accuracy and matching degree of the recommendation results can be improved by adjusting the recommendation results. Most of the existing technologies use a single algorithm for matching, while the present invention realizes an effective combination of the two. Recommendation optimization mechanism based on real-time feedback: The recommendation system in the present invention can perform real-time self-learning and dynamic optimization according to the feedback from HR. This is an important innovation that is different from the traditional static recommendation system. This mechanism can effectively improve the accuracy of the recommendation results and ensure that the system is continuously optimized as demand changes. Deep processing capability of unstructured data: The ability to process unstructured resume text and job descriptions through deep learning technology and extract high-level semantic features makes the system not only rely on keyword matching, but also perform deep semantic analysis through natural language processing technology. This is a key technical difficulty in resume recommendation.

[0041] like Figure 2 As shown, as another preferred embodiment of the present invention, the steps of collaborative filtering processing based on filtering both parties specifically include:

[0042] S11, based on collaborative filtering of users, obtain the historical recruitment records of the enterprise, classify them according to job similarities, perform data matching on similar jobs under the same classification, obtain corresponding recruitment preferences, and screen resumes based on the recruitment preferences;

[0043] S12, project-based collaborative filtering, analyzes the job demand information according to each job project, and obtains the feature matrix of the corresponding job. The feature matrix is ​​used for similarity matching with resumes. The demand information includes job description, skill requirements and past employment characteristics.

[0044] In this embodiment, collaborative filtering is a commonly used recommendation algorithm, which is particularly suitable for personalized recommendation. The present invention adopts a combination of user-based collaborative filtering and project-based collaborative filtering, which are used to recommend resumes and analyze job requirements respectively; user-based collaborative filtering: by analyzing the historical recruitment behavior of enterprise HR, the preferences when recruiting for similar positions are identified, and resumes similar to past recruitment preferences are recommended. Each enterprise HR will generate a user preference matrix based on its previous resume selection, interview feedback and recruitment status, and provide resume recommendations for the current position by analyzing the behavioral similarity with other HRs; project-based collaborative filtering: the system regards each position as a project, and analyzes the job description, skill requirements, etc. of the position. , the characteristics of previously hired personnel, generate a characteristic matrix of the position, and use the similarity between resumes and position characteristics to recommend resumes that are highly matched to the position; the advantage of collaborative filtering is that resumes can be recommended using historical data, and it continuously learns and adjusts based on the HR's selection behavior to make the recommendation results more in line with actual needs; in addition to being based on the company as a unit, recruitment preferences can also be differentiated based on the individual HR recruiter, and can provide personalized resume recommendations for specific positions or specific companies based on the personalized needs of HR. The system generates customized recommendation plans based on HR's historical operation records to ensure that different companies and different HRs can still obtain resume recommendations that meet their specific preferences when facing the same position.

[0045] As another preferred embodiment of the present invention, the step of extracting feature text from unstructured text information of resume content and job requirements based on a deep learning model to obtain deep implicit features in the text information specifically includes:

[0046] Obtaining unformatted information in the resume, and performing semantic feature extraction on text data of the unformatted information based on a deep learning model to generate a deep feature vector corresponding to the resume, wherein the unformatted information includes work experience, educational background, and skill description;

[0047] Unformatted information in the job description is obtained, and features are extracted from the unformatted information based on a deep learning model to generate a deep feature vector corresponding to the job. The unformatted information includes job requirements, skill requirements, and experience requirements.

[0048] In this embodiment, in order to make up for the shortcomings of collaborative filtering in processing complex text data, the present invention further introduces a deep learning model to process unstructured text information in resumes and job descriptions. The deep learning model can extract more accurate and implicit features from text data by self-learning, thereby improving the accuracy of resume-job matching. The system's multi-level recommendation mechanism can quickly screen out the optimal resume, significantly reducing the time cost of HR screening through massive resumes. Through the automatic recommendation mechanism, HR does not need to manually match job requirements with resumes one by one. The recommendation list already includes highly matched candidates, greatly improving recruitment efficiency, especially suitable for large-scale recruitment or emergency recruitment scenarios; the data-driven mechanism can dynamically adjust the parameters of the recommendation algorithm according to real-time market changes and HR's operational feedback, ensuring that the system can continuously optimize the recommendation results as demand changes. Through real-time data updates and feedback learning, the system can adaptively adjust the matching strategy to ensure the continuous accuracy and applicability of the recommendation results to meet recruitment needs in different time periods.

[0049] As another preferred embodiment of the present invention, the step of obtaining market change data and enterprise demand data, and adjusting the weight of the recommendation algorithm driven by data based thereon to achieve dynamic recommendation strategy adjustment specifically includes:

[0050] Access the recruitment data platform to obtain job information and resume data in real time, evaluate and update its feature data and calculate the matching degree according to the preset update cycle;

[0051] Obtain matching feedback data, and optimize the recommendation algorithm based on the matching feedback data, wherein the matching feedback data is used to characterize the recruitment status of the corresponding position, and the recruitment status includes resume information of the recruitment object;

[0052] Use the preset evaluation cycle to obtain corporate preferences and market trends within the historical period, obtain the changes in demand weights of different companies and different positions, and based on this, perform multi-level recommendation weight optimization to achieve recommendation strategy adjustment.

[0053] Furthermore, the multi-level recommendation weights are specifically used to implement the following steps:

[0054] Recommend commonly used resumes for similar positions based on collaborative filtering algorithms, based on HR's historical behavior and job similarity;

[0055] By deeply matching resumes with job characteristics, we recommend resumes that are highly consistent with job descriptions;

[0056] Based on HR's operational behavior and feedback information, the recommendation strategy is dynamically adjusted to optimize the recommendation results.

[0057] In this embodiment, in order to improve the flexibility and adaptability of the resume recommendation system, the present invention designs a data-driven dynamic adjustment mechanism, so that the system can automatically adjust the recommendation strategy according to changes in the recruitment market and different corporate needs; real-time data update: by accessing real-time data such as the recruitment platform, resume database, HR operation behavior, etc., the system can obtain the latest job information and resume data at any time. For each job and resume, the system will continuously update its feature data and recalculate the matching degree based on the latest information; feedback learning mechanism: HR feedback (such as interview arrangement, recruitment, resume screening, etc.) is recorded in the system as feedback data for the recommendation results. The system optimizes the recommendation algorithm through this feedback information and continuously improves the accuracy of resume matching. For example, if a certain type of resume is always selected by HR and passes the interview, the system will increase the recommendation weight of this type of resume for similar positions. Dynamic weight adjustment: The system adopts a multi-level recommendation mechanism, which not only combines the recommendation results of collaborative filtering and deep learning, but also dynamically adjusts the weight of the recommendation algorithm according to the historical preferences of the company's HR and market trends. This dynamic adjustment enables the system to flexibly adapt and improve the accuracy of recommendations when facing the needs of different companies and different positions; The recommendation results of the system are generated by a multi-level mechanism to ensure the comprehensiveness and accuracy of the recommendation results. The first layer: Collaborative filtering recommendation: The system first recommends commonly used resumes for similar positions based on the collaborative filtering algorithm, mainly based on the historical behavior of HR and job similarity recommendations; The second layer: Deep learning recommendation: Through deep matching of resumes and job features, recommend resumes that are highly consistent with the job description; The third layer: Dynamic feedback optimization: Combined with HR's operational behavior and feedback information, dynamically adjust the recommendation strategy and optimize the recommendation results.

[0058] Specifically, resumes and job descriptions usually contain a large amount of unstructured data, such as text descriptions, skill tags, etc. The deep learning technology adopted in the present invention can effectively process these complex data types and extract deep text features, so that the recommendation of resumes is not limited to keyword matching, but is recommended through more in-depth semantic analysis, which greatly improves the matching degree between complex job requirements and resumes, and is especially suitable for technical positions, management positions and other positions with more complex descriptions; at the same time, the system has a built-in feedback learning mechanism, which can automatically optimize the algorithm according to HR's resume screening, interview and recruitment behavior. By accumulating HR's operation data, the system can gradually optimize the recommendation effect and make it more and more in line with the company's recruitment preferences. This self-learning mechanism not only reduces the workload of manually adjusting the recommendation algorithm, but also ensures the gradual optimization and intelligent development of the recommendation system.

[0059] As another preferred embodiment of the present invention, it also includes an output management step, specifically including:

[0060] Selecting a specified number of resumes that meet the current job requirements based on the matching ranking of the resumes, marking the resumes based on the matching scores, and visually outputting them;

[0061] The corresponding resume information is obtained, and the information text matching the position is displayed and marked according to the matching degree. The display mark is used to facilitate quick positioning when the enterprise views it.

[0062] Furthermore, the method further comprises the steps of:

[0063] When the recommended resumes output in a visual manner do not obtain selection feedback information from the enterprise, the system optimizes the recommendation algorithm based on the enterprise screening and issues an alarm, which is used to inform the need for adjustment of job requirements.

[0064] In this embodiment, the system will eventually generate a resume matching ranking, recommend the resume that best meets the current job requirements, and present it to HR in a visual interface; among them, resume recommendation list: the system displays a list of recommended resumes with matching scores, which is convenient for HR to quickly screen resumes; recommendation details: HR can view the matching details of each resume with the job requirements, and the correspondence between the important features in the resume and the job description will be highlighted to help HR better understand the basis for the recommendation. When the system detects that HR is not satisfied with the recommended resume, it will automatically issue an alarm to prompt HR to adjust the job requirements or resume screening conditions and provide optimization suggestions. This function can help HR adjust the recruitment strategy in time, avoid wasting time due to inaccurate demand descriptions or poor recommendation results, and at the same time improve the success rate of recruitment; furthermore, the system provides a simple and visual recommendation interface, and HR can easily view the matching scores and detailed matching basis of the recommended resumes. Through the intuitive interface display, HR can quickly understand why the recommended resume is suitable for the position, which improves the user's operating experience and trust in the recommendation results.

[0065] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the program can be stored in a non-volatile computer-readable storage medium. When the program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. As an illustration and not limitation, RAM is available in many forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).

[0066] Those skilled in the art will readily appreciate other embodiments of the present disclosure after considering the disclosure in the specification and examples. This application is intended to cover any variations, uses or adaptations of the present disclosure, which follow the general principles of the present disclosure and include common knowledge or customary techniques in the art that are not disclosed in the present disclosure. The specification and examples are intended to be exemplary only, and the true scope and spirit of the present disclosure are indicated by the claims.

[0067] It should be understood that the present disclosure is not limited to the exact structures that have been described above and shown in the drawings, and that various modifications and changes may be made without departing from the scope thereof. The scope of the present disclosure is limited only by the appended claims.

Claims

1. An efficient resume intelligent recommendation algorithm, characterized in that: Include: Collaborative filtering based on filtering both parties is used to obtain resume features and analyze job requirements; Based on the deep learning model, feature text extraction is performed on the unstructured text information of resume content and job requirements to obtain deep implicit features in the text information; Obtain market change data and enterprise demand data, and adjust the weight of the recommendation algorithm based on data-driven data to achieve dynamic recommendation strategy adjustment; According to the adjusted recommendation strategy, the resume requirements are matched, a number of matching degree data corresponding to the resumes are obtained, and resume rankings are generated in descending order based on the matching degrees and output visually. The matching degrees are used to characterize the degree of conformity to the positions.

2. According to claim 1, an efficient resume intelligent recommendation algorithm is characterized in that: The steps of collaborative filtering processing based on filtering both parties specifically include: Based on user collaborative filtering, the company's historical recruitment records are obtained, and the positions are classified according to their similarity. Data matching is performed on similar positions under the same category to obtain the corresponding recruitment preferences, and resume screening is performed based on the recruitment preferences. Based on project collaborative filtering, according to each job project, the job demand information is analyzed to obtain the feature matrix of the corresponding job. The feature matrix is ​​used for similarity matching with resumes. The demand information includes job description, skill requirements and past recruitment characteristics.

3. According to claim 2, an efficient resume intelligent recommendation algorithm is characterized in that: The step of extracting feature text from the unstructured text information of resume content and job requirements based on the deep learning model to obtain deep implicit features in the text information specifically includes: Obtaining unformatted information in the resume, and performing semantic feature extraction on text data of the unformatted information based on a deep learning model to generate a deep feature vector corresponding to the resume, wherein the unformatted information includes work experience, educational background, and skill description; Unformatted information in the job description is obtained, and features are extracted from the unformatted information based on a deep learning model to generate a deep feature vector corresponding to the job. The unformatted information includes job requirements, skill requirements, and experience requirements.

4. According to claim 3, an efficient resume intelligent recommendation algorithm is characterized in that: The steps of obtaining market change data and enterprise demand data and adjusting the weight of the recommendation algorithm based on the data to achieve dynamic recommendation strategy adjustment specifically include: Access the recruitment data platform to obtain job information and resume data in real time, evaluate and update its feature data and calculate the matching degree according to the preset update cycle; Obtain matching feedback data, and optimize the recommendation algorithm based on the matching feedback data, wherein the matching feedback data is used to characterize the recruitment status of the corresponding position, and the recruitment status includes resume information of the recruitment object; Use the preset evaluation cycle to obtain corporate preferences and market trends within the historical period, obtain the changes in demand weights of different companies and different positions, and based on this, perform multi-level recommendation weight optimization to achieve recommendation strategy adjustment.

5. According to claim 4, an efficient resume intelligent recommendation algorithm is characterized in that: The multi-level recommendation weights are specifically used to implement the following steps: Recommend commonly used resumes for similar positions based on collaborative filtering algorithms, based on HR's historical behavior and job similarity; By deeply matching resumes with job characteristics, we recommend resumes that are highly consistent with job descriptions; Based on HR's operational behavior and feedback information, the recommendation strategy is dynamically adjusted to optimize the recommendation results.

6. According to claim 5, an efficient resume intelligent recommendation algorithm is characterized in that: It also includes output management steps, including: Selecting a specified number of resumes that meet the current job requirements based on the matching ranking of the resumes, marking the resumes based on the matching scores, and visually outputting them; The corresponding resume information is obtained, and the information text matching the position is displayed and marked according to the matching degree. The display mark is used to facilitate quick positioning when the enterprise views it.

7. The efficient resume intelligent recommendation algorithm according to claim 6 is characterized in that: Also includes the steps: When the recommended resumes output in a visual manner do not obtain selection feedback information from the enterprise, the system optimizes the recommendation algorithm based on the enterprise screening and issues an alarm, which is used to inform the need for adjustment of job requirements.

Images