Enterprise recruitment management system based on talent portraits

Abstract

The invention provides an enterprise recruitment management system based on a talent portrait. The enterprise recruitment management system comprises a portrait construction module, a post matching module, an interview generation module and a result management module, the portrait construction module is used for acquiring talent feature data of the candidate and constructing a talent portrait of the candidate; the post matching module is used for acquiring the published post demand features to obtain post demand items and corresponding post demand feature vectors; extracting talent feature vectors according to the talent portraits; performing matching analysis according to the obtained post demand feature vector and the talent feature vector to obtain a candidate post matching analysis result; the interview generation module is used for matching corresponding interview questions from a question library according to the acquired candidate post matching analysis result, and generating an interview question list; and the result management module is used for acquiring interview information of the candidate for the interview problem and generating a candidate post interview evaluation result. According to the invention, the objectivity and the intelligent level of enterprise post recruitment can be improved.

Description

Technical Field

[0001] The present invention relates to the technical field of enterprise recruitment management, and in particular to an enterprise recruitment management system based on talent portrait. Background Art

[0002] The healthy development of an enterprise cannot be separated from the support of excellent talents. At present, enterprises have an increasing demand for and attention to talents. Traditional corporate recruitment usually relies on third-party talent platforms or talent pools as a medium to select suitable talents for the positions posted by the enterprise through resume keyword matching or manual screening. However, the above methods have problems such as low efficiency, strong subjectivity, and low job suitability of selected talents. This makes it impossible for enterprises to effectively identify talents quickly, truly, and effectively. Summary of the invention

[0003] In view of the above problems, the present invention aims to provide an enterprise recruitment management system based on talent portrait.

[0004] The purpose of the present invention is achieved by the following technical solutions: The present invention proposes an enterprise recruitment management system based on talent portrait, including a portrait construction module, a job matching module, an interview generation module and a result management module; wherein, The portrait construction module is used to obtain the talent characteristic data of the candidate and construct the talent portrait of the candidate based on the obtained talent characteristic data; wherein the talent portrait includes the characteristic quantity of each talent characteristic of the candidate and the corresponding evaluation factor; The job matching module is used to obtain the characteristics of the published job requirements, obtain the job requirement items and the corresponding job requirement feature vectors; and extract the talent feature vectors corresponding to the job requirement items according to the talent portrait; perform matching analysis based on the obtained job requirement feature vectors and talent feature vectors to obtain the candidate job matching analysis results; Among them, matching analysis is performed based on the obtained job requirement feature vector and talent feature vector, specifically including: For each job requirement, obtain the candidate's matching degree for each job requirement. P(i)=sim (VC i ,VJ i ) ,in VC i Indicates that the candidate has a high level of talent i The talent feature vector under VJ i Indicates that the recruitment position is targeted at the job demand project i The job requirement feature vector, sim(VC i ,VJ i ) express VC i and VJ i The similarity is calculated using the cosine function; The matching degree between the candidates and the recruitment positions is further analyzed, and the matching analysis function used is: ； In the formula, MatchP Indicates the candidate's job matching factor, where the larger the matching factor, the higher the matching degree between the candidate and the current recruitment position; P(i) Indicates the candidate's requirements for the position i The matching degree, N Indicates the total number of items; S i Indicates job requirements i The attention factor of the position is calculated by counting the project requirements of the position. i Obtained from the frequency of occurrence in the recruitment position characteristic data; j Represents a variable, S j Indicates job requirements j The attention factor λ i Indicates the preset job requirements i The time adjustment factor, tC i Indicates the candidate's talent characteristics project i Experience time difference, based on the candidate's talent characteristics project i The time difference between the latest experience time node and the current time node is calculated in months; α and β Respectively represent the set weight adjustment factors, δ represents the influence factor of the set condition, Num(c) Indicates the number of candidates that do not meet the minimum criteria, calculated based on the results of the candidate condition comparison analysis; NumT Indicates the number of standard conditions set; Output the candidate's matching degree for each job requirement and the candidate's job matching factor MatchP As the result of candidate job matching analysis; The interview generation module is used to match the corresponding interview questions from the question library according to the obtained candidate job matching analysis results, and generate an interview question list; The result management module is used to obtain the interview information of the candidate in response to the interview questions and generate the candidate's job interview evaluation results.

[0005] Preferably, the portrait construction module includes a resume data acquisition unit, a project data acquisition unit, a test data acquisition unit and a talent portrait construction unit; wherein, The resume data acquisition unit is used to acquire the resume data of the candidate and extract the corresponding skill feature data according to the resume data, wherein the skill feature data includes the skill items mastered, the mastery degree and experience corresponding to the skill items, and the qualification certificates obtained; The project data acquisition unit is used to acquire the project feature data corresponding to the candidate from the project release source, wherein the project feature data includes the number of projects, the time of project occurrence, and the skill features associated with the project, etc.; the test data acquisition unit is used to acquire the test feature data of the candidate, wherein the test feature data includes the results of logic tests, the results of personality tests, etc.; The talent profile building unit is used to build a talent profile of a candidate based on the candidate's skill feature data, project feature data and test feature data.

[0006] Preferably, the resume data acquisition unit includes: Get the candidate's resume text; Perform text vectorization processing on the acquired resume text, extract the text feature vector of the resume data, and extract key feature words based on the text feature vector to obtain the corresponding skill feature text; The acquired skill feature text is further subjected to text mapping processing to obtain the skill feature text and the mapping text as skill feature data.

[0007] Preferably, the system further comprises a job requirement extraction module; The job requirement extraction module is used to obtain the published recruitment job feature data, and extract the job requirement characteristics based on the obtained recruitment job feature data; the job requirement characteristics also include minimum standard requirement characteristics.

[0008] Preferably, the job matching module includes a feature integration unit, a condition analysis unit and a matching analysis unit; wherein, The feature integration unit is used to integrate the job requirement features under the same job requirement project according to the acquired job requirement features, and obtain the job requirement feature vectors of each job requirement project; and to integrate the talent feature data under the talent feature project corresponding to the job requirement project according to the talent profile of the candidate, and obtain the talent feature vector under the corresponding talent feature project; wherein the job requirement feature vector and the talent feature vector include the feature quantity or evaluation factor of the corresponding project; The condition analysis unit is used to perform condition comparison analysis based on the minimum standard items set in the job requirement characteristics and the corresponding talent characteristic items in the candidate talent profile, and obtain the condition comparison analysis results of the candidate corresponding to each minimum standard characteristic; The matching analysis unit is used to perform matching analysis based on the acquired job requirement characteristics and talent portraits to obtain candidate job matching analysis results.

[0009] Preferably, the interview generation module includes a question library unit, a question recommendation unit and a generation unit; wherein, The question library unit is used to store and manage interview questions for recruitment positions, where each interview question contains a corresponding question feature item label and a corresponding question feature vector label; The question recommendation unit is used to calculate the priority of each interview question in the question bank unit according to the candidate job matching analysis results; The generation unit is used to select appropriate interview questions from the question library unit according to the priority of each interview question and based on preset topic selection rules, and generate a corresponding interview question list.

[0010] Preferably, the result management module further includes: obtaining corresponding interview questions according to the generated interview question list and pushing them to the candidates, obtaining the interview results of the candidates for the interview questions, and displaying the interview results.

[0011] The beneficial effects of the present invention are as follows: the present invention proposes an enterprise recruitment management system based on talent portrait, wherein the talent characteristics of the candidate are extracted according to the obtained talent characteristic data of the candidate, and a talent portrait for the candidate is constructed. Based on the constructed talent portrait, the ability and degree of the candidate for different talent characteristic items can be intuitively and accurately reflected, and the situation of the candidate can be accurately fed back and evaluated. When performing job matching, the characteristic items and characteristic vectors corresponding to the recruitment position can be adaptively extracted based on the talent portrait of the candidate, and the matching analysis of the candidate and the job adaptability can be completed based on the talent characteristic vector. Through the matching analysis based on the talent characteristic vector, the adaptability of the candidate and the recruitment position can be accurately fed back; the adaptability of the candidate and the recruitment position can be objectively and comprehensively evaluated. At the same time, based on the matching analysis results of the candidate and the recruitment position, interview questions adapted to the current situation of the candidate are further matched to assist the manager to further complete further interviews or written tests for the candidate. Through the adapted interview questions, it is helpful to further assess and test the real situation of the candidate, thereby improving the objectivity and authenticity of the enterprise job recruitment. At the same time, this system can assist in completing corporate recruitment management, greatly saving time and manpower in data integration, matching and targeted data matching and screening, and further improving the efficiency of corporate recruitment management. BRIEF DESCRIPTION OF THE DRAWINGS

[0012] The present invention is further described using the accompanying drawings, but the embodiments in the accompanying drawings do not constitute any limitation to the present invention. A person skilled in the art can obtain other drawings based on the following drawings without creative work.

[0013] Figure 1 This is a framework diagram of an enterprise recruitment management system based on talent portraits shown in an embodiment of the present invention; Figure 2 for Figure 1 Schematic diagram of the module structure in the embodiment. DETAILED DESCRIPTION

[0014] The present invention is further described in conjunction with the following application scenarios.

[0015] See also Figure 1 As shown in the embodiment, it shows an enterprise recruitment management system based on talent portrait, including a portrait construction module, a job matching module, an interview generation module and a result management module; wherein, The portrait construction module is used to obtain the talent characteristic data of the candidate and construct the talent portrait of the candidate based on the obtained talent characteristic data; wherein the talent portrait includes the characteristic quantity of each talent characteristic of the candidate and the corresponding evaluation factor; The job matching module is used to obtain the characteristics of the published job requirements, obtain the job requirement items and the corresponding job requirement feature vectors; and extract the talent feature vectors corresponding to the job requirement items according to the talent portrait; perform matching analysis based on the obtained job requirement feature vectors and talent feature vectors to obtain the candidate job matching analysis results; the interview generation module is used to match the corresponding interview questions from the question library according to the obtained candidate job matching analysis results, and generate an interview question list; The result management module is used to obtain the interview information of the candidate in response to the interview questions and generate the candidate's job interview evaluation results.

[0016] The above-mentioned embodiment of the present invention proposes an enterprise recruitment management system based on talent portrait, wherein the talent characteristics of the candidate are extracted according to the obtained talent characteristic data of the candidate, and a talent portrait for the candidate is constructed. Based on the constructed talent portrait, the ability and degree of the candidate for different talent characteristic items can be intuitively and accurately reflected, and the situation of the candidate can be accurately fed back and evaluated. When matching positions, the characteristic items and characteristic vectors corresponding to the recruitment position can be adaptively extracted based on the talent portrait of the candidate, and the matching analysis of the candidate and the position adaptability can be completed based on the talent characteristic vector. Through the matching analysis based on the talent characteristic vector, the adaptability of the candidate and the recruitment position can be accurately fed back; the adaptability of the candidate and the recruitment position can be objectively and comprehensively evaluated. At the same time, based on the matching analysis results of the candidate and the recruitment position, interview questions adapted to the current situation of the candidate are further matched to assist the manager to further complete further interviews or written tests for the candidate. Through the adapted interview questions, it is helpful to further assess and test the real situation of the candidate, thereby improving the objectivity and authenticity of the enterprise position recruitment. At the same time, this system can assist in completing corporate recruitment management, greatly saving time and manpower in data integration, matching and targeted data matching and screening, and further improving the efficiency of corporate recruitment management.

[0017] In one scenario, the system proposed in the present invention can be built based on a cloud server, a local server, etc., and the corresponding data can be entered or transmitted through a smart terminal or the like.

[0018] Preferably, see Figure 2 The portrait construction module includes a resume data acquisition unit, a project data acquisition unit, a test data acquisition unit and a talent portrait construction unit; wherein, The resume data acquisition unit is used to acquire the resume data of the candidate and extract the corresponding skill feature data according to the resume data, wherein the skill feature data includes the skill items mastered, the mastery degree and experience corresponding to the skill items, and the qualification certificates obtained; The project data acquisition unit is used to acquire project feature data corresponding to the candidate from the project publishing source, wherein the project feature data includes the number of projects, the time of project occurrence, and the skill features associated with the project; The test data acquisition unit is used to acquire the test feature data of the candidate, wherein the test feature data includes logic test results, personality test results, etc.; The talent profile building unit is used to build a talent profile of a candidate based on the candidate's skill feature data, project feature data and test feature data.

[0019] The extraction of candidate talent feature data is usually completed based on resume data, project data and test data. Therefore, the above-mentioned embodiment of the present invention particularly sets up a corresponding data acquisition unit, which can extract the required candidate talent feature data based on different data sources, laying the foundation for the subsequent construction of candidate talent portraits. Preferably, the resume data acquisition unit includes: Get the candidate's resume text; Perform text vectorization processing on the acquired resume text, extract the text feature vector of the resume data, and extract key feature words based on the text feature vector to obtain the corresponding skill feature text; The acquired skill feature text is further subjected to text mapping processing to obtain the skill feature text and the mapping text as skill feature data.

[0020] In one scenario, for skill feature text, a text classification model based on the BERT model can be used to implement it. By converting the resume text into a 768-dimensional feature vector, and performing text classification processing based on the feature vector and the trained classifier, the corresponding skill feature text is obtained, such as "proficient in Linux", "proficient in Python", etc.

[0021] Skill feature data is extracted from the acquired resume data, where the text describing the mastered skills on the resume is first extracted, and data cleaning and text mapping processing are performed based on the extracted text. On the one hand, the description of skill features can be accurately extracted from the resume text, and on the other hand, the text of skill features is further adaptively mapped, and the extracted skill text feature description can be further standardizedly mapped to adaptively extract the candidate's skill features.

[0022] In one scenario, for the skill feature text "Proficient in Linux" recorded in the candidate's resume, after standardized mapping, the mapping features related to Linux such as "Proficient in script writing", "Proficient in permission management", and "Proficient in log troubleshooting" can be further extracted according to the preset mapping rules; in another scenario, for the skill feature text "IELTS 7.5", after standardized mapping, the mapping features of "proficient in English listening, speaking, reading, and writing" can be further obtained according to the mapping rules; thereby completing the further adaptive expansion of the candidate's skill feature data, so that the skill feature data extracted based on the resume text is more comprehensive.

[0023] Preferably, the project data acquisition unit acquires the project feature data in a similar manner to the resume data acquisition unit, by acquiring the candidate's project description text or the works associated with the project to extract the corresponding project feature data. For example, text vectorization and key feature word extraction are performed based on the acquired project description text to obtain the project feature data text as the project feature data.

[0024] Wherein, for obtaining the project characteristic data of the candidate, if there is no corresponding project publishing source, the corresponding project description text can also be obtained from the project experience column of the candidate's resume.

[0025] In one scenario, based on the candidate's project experience description in the resume, the candidate's project feature data is extracted as "2 years of AI algorithm development experience", "2 years of Python experience", "5 GitHub works" and corresponding release time, "1 SCI paper corresponding to NLP natural language processing" and corresponding release time, etc.

[0026] For the extraction of test data, the corresponding test results can be obtained through questionnaires based on the test questionnaires customized by the enterprise or the authoritative questionnaires in the industry. The corresponding test data of the candidates can be entered or automatically obtained through the test data acquisition unit, which helps to improve the convenience of test data acquisition.

[0027] In one scenario, the test data uses direct test results as corresponding test feature data.

[0028] Preferably, the talent portrait construction unit specifically includes: Performing feature quantification processing on the obtained candidate talent feature data to extract feature quantities of the candidate feature data; wherein the candidate talent feature data includes the candidate's skill feature data, project feature data and test feature data; The evaluation factor of the corresponding talent characteristic item is calculated based on the obtained characteristic quantity, and the evaluation factor calculation function used is: ; in, Mark(i) Indicates the corresponding talent characteristics i The evaluation factor of x(i) Indicates the talent characteristics of the candidate i The characteristic quantity of μ(i) and σ(i) Respectively represent the talent characteristics items in the talent pool i The average feature quantity and standard deviation of the feature quantity; represents the preset non-zero adjustment item, where ; Based on the obtained talent characteristic items and corresponding evaluation factors of the candidate, a talent profile of the candidate is constructed, wherein the talent profile of the candidate includes the characteristic quantities of various talent characteristics of the candidate and the corresponding evaluation factors.

[0029] In one scenario, when quantifying talent characteristic data, the characteristic data corresponding to the talent characteristic items are quantified, for example, "proficiency" and "mastery" are quantified as "4 points" and "3 points" respectively; and for characteristic quantities expressed in numerical values, such as "5 published works", "3 years of experience", etc., the corresponding numbers "5" and "3" are directly used as the characteristic quantities of the corresponding talent characteristic items "number of works" and "experience" respectively.

[0030] When constructing a talent portrait for an individual candidate, the candidate's talent feature data is first numerically represented through feature quantification processing. An evaluation factor calculation function is proposed, which can use the same type of talent feature data in the talent pool as a benchmark, so as to accurately quantify the candidate's talent feature data after horizontal comparison, and objectively and accurately represent the level of a certain talent feature of the candidate. At the same time, based on the same quantitative standard, it can lay the foundation for further feature matching and analysis.

[0031] Preferably, the system further comprises a job requirement extraction module; The job requirement extraction module is used to obtain the published recruitment job feature data, and extract the job requirement characteristics based on the obtained recruitment job feature data; the job requirement characteristics also include minimum standard requirement characteristics.

[0032] In one scenario, the job requirement extraction module obtains the description text of the recruitment position, and extracts key features based on the obtained description text to obtain the job requirement features corresponding to the recruitment position, such as "proficient in Java", "5 years of Java experience", "3 years of Python", "e-commerce operation experience", "less than 6 months of idle time", etc. At the same time, according to the actual situation, the minimum standard requirement features can also be set to stipulate the minimum standard situation, such as setting the minimum standard requirement feature "Java not less than 1 year".

[0033] The method of obtaining the job demand data of an enterprise is usually to extract it based on the recruitment information text released by the enterprise; at the same time, the enterprise job demand features can also be directly entered into the system based on a special feature entry framework.

[0034] In one scenario, the method of extracting job requirement features from enterprise job information text can be implemented by using a text feature extraction model that has been trained in the prior art, such as an enterprise recruitment feature extraction model built based on TF-IDF or other artificial intelligence models. The present invention does not make specific limitations here.

[0035] Preferably, the job matching module includes a feature integration unit, a condition analysis unit and a matching analysis unit; wherein, The feature integration unit is used to integrate the job demand features under the same job demand project according to the acquired job demand features, and obtain the job demand feature vector of each job demand project. VJ i ; and according to the candidate's talent portrait, the talent feature data under the talent feature project corresponding to the job requirement project is integrated to obtain the talent feature vector under the corresponding talent feature project VC i ; Among them, the job requirement feature vector and the talent feature vector include the feature quantity or evaluation factor of the corresponding project; The condition analysis unit is used to perform condition comparison analysis based on the minimum standard items set in the job requirement characteristics and the corresponding talent characteristic items in the candidate talent profile, and obtain the condition comparison analysis results of the candidate corresponding to each minimum standard characteristic; The matching analysis unit is used to perform matching analysis based on the obtained job requirement feature vector and talent feature vector to obtain candidate job matching analysis results.

[0036] Preferably, the matching analysis unit specifically includes: For each job requirement, obtain the candidate's matching degree for each job requirement. P(i)=sim (VC i ,VJ i ) ,in VC i Indicates that the candidate has a high level of talent i The talent feature vector under VJ i Indicates that the recruitment position is targeted at the job demand project i The job requirement feature vector, sim(VC i ,VJ i ) express VC i and VJ i The similarity is calculated using the cosine function; The matching degree between the candidates and the recruitment positions is further analyzed, and the matching analysis function used is: ； In the formula, MatchP Indicates the candidate's job matching factor, where the larger the matching factor, the higher the matching degree between the candidate and the current recruitment position; P(i) Indicates the candidate's requirements for the position i The matching degree, N Indicates the total number of items; S i Indicates job requirements i The attention factor of the position is calculated by counting the project requirements of the position. i Obtained from the frequency of occurrence in the recruitment position characteristic data; j Represents a variable, S j Indicates job requirements j The attention factor λ i Indicates the preset job requirements i The time adjustment factor, tC i Indicates the candidate's talent characteristics project i Experience time difference, based on the candidate's talent characteristics project i The time difference between the latest experience time node and the current time node is calculated in months; α and β Respectively represent the set weight adjustment factors, δ represents the influence factor of the set condition, Num(c) Indicates the number of candidates that do not meet the minimum criteria, calculated based on the results of the candidate condition comparison analysis; NumT Indicates the number of standard conditions set; Output the candidate's matching degree for each job requirement and the candidate's job matching factor MatchP As the result of candidate job matching analysis.

[0037] In one scenario, in order to improve the accuracy of the calculation, VC i and VJ iWhen calculating the similarity of the job, first correct the features of each dimension in the talent feature vector. When the feature value of a certain dimension is greater than the feature value of the corresponding dimension in the job requirement feature vector, correct the value of the feature value to the same value as the corresponding dimension in the job requirement feature vector. Or when using the cosine function to calculate the cosine similarity, set the local maximum value to limit the similarity range between the two, so as to avoid the result callback caused by exceeding the maximum value during the similarity calculation process.

[0038] In one scenario, for example, the job requirement vector for the job requirement item "Logical test score" is "120+", and the candidate's logical test result is 130. When calculating the similarity between the job requirement vector and the talent feature vector, the talent feature item vector is corrected so that the similarity (matching degree) of the two is 1.

[0039] In one scenario, in the feature integration unit, the job requirement features "mastering Java", "more than 5 years of Java experience", "Java project experience as a main developer", "proficient in BUG troubleshooting", etc. can be integrated into a unified job requirement feature vector under the job requirement project "Java" VJ java ={2+ab,5+yrs,0.8+mmb,2.2+abb} , where 4+ and 1+ are the evaluation factor standards corresponding to the Java capability item and the Java experience item, respectively. Similarly, based on the candidate's talent profile, the candidate's characteristic quantities or evaluation factors related to the talent characteristic item "Java", such as "Java capability evaluation factor 2.3", "Java experience 4 years", "Java experience evaluation factor 2.5", and the mapped capability feature "bug troubleshooting evaluation factor 2.2" obtained based on his "proficiency in Python" are integrated into a talent characteristic vector VC java ={2.3ab,4yrs,2.5mmb,2.2abb} .

[0040] In another scenario, in the feature integration unit, for test-related items, the feature quantities or evaluation factors corresponding to the results of each test (such as personality test, logic test, etc.) can be integrated into a feature vector under a unified "test" item; the results of each test can also be subdivided, with different "personality tests" or "logic tests" as separate items, and the feature vectors (which may be one-dimensional or multi-dimensional feature vectors) under the corresponding items can be obtained based on the feature quantities or evaluation factors under each item.

[0041] It should be noted that the above-mentioned job requirement items and talent feature items can be collectively referred to as items. The only difference is that when the items appear in the relevant content of the recruitment position, they are called job requirement items; when the items appear in the relevant content of the candidates, they are called talent feature items. Therefore, in the actual data processing and calculation process, the job requirement feature vector and talent feature vector of the same item can be processed accordingly.

[0042] The above-mentioned implementation mode of the present invention takes into account that in an actual recruitment scenario, there will usually be one or more feature quantities for the same project. Therefore, a feature vector extraction based on the project is first performed based on the job requirement characteristics and talent characteristic data. One-dimensional or multi-dimensional feature vectors can be constructed for different projects, which is convenient for subsequent matching or analysis of the data.

[0043] Among them, when conducting condition analysis, firstly, the minimum condition comparison analysis of the talent characteristics of the candidates is carried out based on the minimum standard in the job requirements. Based on the principle of minimum standard, each talent characteristic item of the candidate can be verified to ensure that the candidate can meet the minimum standard of the recruitment position (the minimum standard can have a direct impact on the subsequent matching analysis results); When performing matching analysis, based on the idea of ​​single item analysis and comprehensive analysis, we first calculate the matching degree based on the proposed function P(i) Based on the job requirement feature vector and the talent feature vector, the candidate's corresponding talent feature items can be matched and analyzed one by one according to the job requirement items; based on the obtained project matching degree, further based on the proposed matching analysis function MatchP , to comprehensively calculate the matching degree of each project. At the same time, in the process of comprehensive statistics, the attention factor S is specially added to objectively adjust the demand (importance) and attention of the recruitment position for different projects to improve the accuracy of comprehensive statistics. At the same time, the timeliness of talent ability is further considered in the comprehensive statistics part. Therefore, when conducting comprehensive analysis, a time adjustment item is specially added to correct the timeliness of the talent ability of the candidate. Compared with the traditional method based on a single evaluation standard, it can further fit the changing characteristics of talent characteristics. At the same time, the matching analysis function further makes a final correction to the matching degree of the candidate to the job recruitment based on the conditional comparison analysis results, so that the obtained candidate position matching factor can accurately and objectively quantify the matching degree of the candidate to the recruitment position, improving the intelligence level of the candidate and recruitment position matching analysis.

[0044] In one scenario, according to the obtained candidate job matching factor, a corresponding standard threshold can be set to conduct preliminary screening of the candidates, thereby achieving accurate screening results based on the objective conditions of the candidates.

[0045] Considering that only the candidate's job matching factor or other comprehensive evaluation results can usually only provide feedback on the candidate's situation from the perspective of hardware conditions, but cannot fully reflect the candidate's true situation (for example, there are subjective evaluations, false reporting of the situation (falsely high or falsely low), etc.), therefore, it is usually necessary to combine further interview questions to consider the candidate. However, the current interview question selection method is to use fixed interview questions to test candidates. This method can usually only consider whether the interviewee meets a certain condition (such as whether he is qualified), but cannot further consider outstanding talents, and the gradient level is insufficient; the other is to ask interview questions based on the interviewer's subjective thoughts. This method is prone to strong subjectivity or bias, and cannot further consider the candidate's comprehensive situation.

[0046] Therefore, the present invention further proposes a method for matching and generating interview questions, based on the recruitment position and the characteristics of the candidate, matching appropriate questions to assist in further consideration of the candidate, thereby providing true feedback on the candidate's true situation.

[0047] Preferably, the interview generation module includes a question library unit, a question recommendation unit and a generation unit; wherein, The question library unit is used to store and manage interview questions for recruitment positions, where each interview question contains the corresponding question feature item label i and the corresponding problem feature vector label VT i ; The question recommendation unit is used to calculate the priority of each interview question in the question library unit according to the candidate job matching analysis results. The priority calculation function used is: ; In the formula, PR t (i,VT i )surface Interview Questions t Priority of interview questions t The corresponding problem feature items are i , the corresponding style feature vector is VT i , P(i) Indicates the candidate's requirements for the position i The matching degree of P(i)=sim (VC i ,VJ i ) , VC i Indicates that the candidate has a high level of talent i The talent feature vector under VJ i Indicates that the recruitment position is targeted at the job demand project i The job requirement feature vector, sim(VC i ,VJ i ) express VC i and VJ i The similarity is calculated using the cosine function; MatchP represents the candidate's job matching factor, θP represents the preset standard matching factor threshold, k represents the set sensitivity factor, where k∈[0.1,1000] , sim(VT i ,VC i ) express VT i and VC i The similarity of sim(VT i ,VJ i ) express VT i and VJ i similarity; The generation unit is used to select appropriate interview questions from the question library unit according to the priority of each interview question and based on preset topic selection rules, and generate a corresponding interview question list.

[0048] Among them, the question library unit stores interview questions for different ability projects (for example, Java, logical ability, Python, etc.), which carry corresponding question feature project labels and corresponding question feature vector labels. The above-mentioned question feature projects correspond to the job requirement projects of the above-mentioned recruitment positions and the talent feature projects of the candidates. For example, the question feature project labels carried in the interview questions are "Java", "Project Management", etc.; and based on the same idea, the question feature vector labels also correspond to the job requirement feature vectors of the job requirement projects and the talent feature vectors under the talent feature projects. For example, the feature vector labels carried in the interview questions are VT java = {3ab,2abb} wait.

[0049] In one scenario, the generation unit selects appropriate interview questions from the question library unit based on a preset topic selection rule, wherein the preset topic selection rule may be to select the N interview questions with the highest priority from the questions according to the priority of each question to generate an interview question list. In another scenario, as an optimization, more conditions may be added when selecting interview questions according to the priority of the interview questions. For example, for each recruitment demand project, the top n interview questions of the corresponding project priority are obtained to generate an interview question list to ensure that the questions can fully cover the job requirements.

[0050] Based on the generated interview question list, the interviewer / manager can obtain the corresponding interview questions to conduct further personalized and targeted assessments of the candidates, and then further evaluate whether to hire the candidates or tap the candidates' talent capabilities based on the candidates' responses to the interview questions.

[0051] The above-mentioned implementation mode of the present invention sets corresponding assessment items (projects) and difficulty (reflected by feature vectors) for interview questions in the question bank based on the constructed question bank; when matching interview questions for designated recruitment positions and candidates, based on the proposed priority calculation function, the priority of the interview questions can be matched based on the similarity between the interview question feature vector and the job requirement vector of the recruitment position and the talent feature vector of the candidate as a benchmark, wherein a judgment factor based on the candidate's job matching factor is specially added to adaptively adjust and match appropriate interview questions based on the candidate's situation, so that the final matched interview questions can adapt to the job requirements while their difficulty can be as close to the candidate's actual level as possible, thereby reflecting the ability gradient between candidates and helping to improve the intelligence level of interview question matching.

[0052] Among them, for the setting of the above priority calculation function, when the candidate's ability is slightly lower than the job requirement standard, the interview focuses more on whether the candidate can meet the job requirements. Therefore, through the calculation of priority, the difficulty and conditions of the interview questions can be made closer to the level corresponding to the job requirements, so as to further consider the candidate's situation based on the interview questions, which is helpful for further discovery and mining of potential candidates or candidates with low conditions, and avoid waste of talents. When the candidate's ability is obviously higher than the job requirement, the focus of the interview is to find the accurate positioning of the interviewee's ability. Therefore, matching the corresponding interview questions based on the candidate's characteristics as a benchmark can help reflect the candidate's true situation, screen candidates with high conditions and mine the candidate's true ability, and improve the intelligence level and accuracy of the evaluation of the candidate's true level.

[0053] Preferably, the result management module further includes: According to the generated interview question list, the corresponding interview questions are obtained and pushed to the candidates, and the interview results of the candidates for the interview questions are obtained and displayed.

[0054] Among them, after generating the list of interview questions, you can choose to test the candidates according to the interview questions in the form of written test or interview according to the actual situation, and get the corresponding recruitment results based on the candidates' answers to the interview questions. The recruitment results of the candidates can be managed as a whole through the result management module.

[0055] It should be noted that each functional unit / module in each embodiment of the present invention may be integrated into one processing unit / module, or each unit / module may exist physically separately, or two or more units / modules may be integrated into one unit / module. The above-mentioned integrated unit / module may be implemented in the form of hardware or in the form of software functional unit / module.

[0056] Through the description of the above implementation modes, it can be clearly understood by those skilled in the art that the embodiments described herein can be implemented in hardware, software, firmware, middleware, code or any appropriate combination thereof. For hardware implementation, the processor can be implemented in one or more of the following units: an application specific integrated circuit (ASIC), a digital signal processor (DSP), a digital signal processing device (DSPD), a programmable logic device (PLD), a field programmable gate array (FPGA), a processor, a controller, a microcontroller, a microprocessor, other electronic units designed to implement the functions described herein or a combination thereof. For software implementation, part or all of the processes of the embodiment can be completed by instructing the relevant hardware through a computer program. When implemented, the above program can be stored in a computer-readable medium or transmitted as one or more instructions or codes on a computer-readable medium. Computer-readable media include computer storage media and communication media, wherein the communication media include any medium that facilitates the transmission of computer programs from one place to another. The storage medium can be any available medium that can be accessed by a computer. Computer-readable media can include, but are not limited to, RAM, ROM, EEPROM, CD-ROM or other optical disk storage, disk storage media or other magnetic storage devices, or any other medium that can be used to carry or store the desired program code in the form of instructions or data structures and can be accessed by a computer.

[0057] Finally, it should be noted that the above embodiments are only used to illustrate the technical solution of the present invention, rather than to limit the scope of protection of the present invention. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should analyze that the technical solution of the present invention can be modified or replaced by equivalents without departing from the essence and scope of the technical solution of the present invention.

Claims

1. An enterprise recruitment management system based on talent portrait, characterized in that: It includes portrait construction module, job matching module, interview generation module and result management module; among them, The portrait construction module is used to obtain the talent characteristic data of the candidate and construct the talent portrait of the candidate based on the obtained talent characteristic data; wherein the talent portrait includes the characteristic quantity of each talent characteristic of the candidate and the corresponding evaluation factor; The job matching module is used to obtain the characteristics of the published job requirements, obtain the job requirement items and the corresponding job requirement feature vectors; and extract the talent feature vectors corresponding to the job requirement items according to the talent portrait; perform matching analysis based on the obtained job requirement feature vectors and talent feature vectors to obtain the candidate job matching analysis results; Among them, matching analysis is performed based on the obtained job requirement feature vector and talent feature vector, specifically including: Get the matching degree of candidates for each job requirement item separately P(i)=sim(VC i ,VJ i ) ,in VC i Indicates that the candidate has a high level of talent i The talent feature vector under VJ i Indicates that the recruitment position is targeted at the job demand project i The job requirement feature vector, sim(VC i ,VJ i ) express VC i and VJ i The similarity is calculated using the cosine function; The matching degree between the candidates and the recruitment positions is further analyzed, and the matching analysis function used is: ； In the formula, MatchP Indicates the candidate's job matching factor, where the larger the matching factor, the higher the matching degree between the candidate and the current recruitment position; P(i) Indicates the candidate's requirements for the position i The matching degree, N Indicates the total number of items; S i Indicates job requirements i The attention factor of the position is calculated by counting the project requirements of the position. i Obtained from the frequency of occurrence in the recruitment position characteristic data; j Represents a variable, S j Indicates job requirements j The attention factor λ i Indicates the preset job requirements i The time adjustment factor, tC i Indicates the candidate's talent characteristics project i Experience time difference, based on the candidate's talent characteristics project i The time difference between the latest experience time node and the current time node is obtained by statistics; α and β Respectively represent the set weight adjustment factors, δ represents the influence factor of the set condition, Num(c) Indicates the number of candidates that do not meet the minimum criteria, calculated based on the results of the candidate condition comparison analysis; NumT Indicates the number of standard conditions set; Output the candidate's matching degree for each job requirement and the candidate's job matching factor MatchP As the result of candidate job matching analysis; The interview generation module is used to match the corresponding interview questions from the question library according to the obtained candidate job matching analysis results, and generate an interview question list; The result management module is used to obtain the interview information of the candidate in response to the interview questions and generate the candidate's job interview evaluation results.

2. According to claim 1, the enterprise recruitment management system based on talent portrait is characterized in that: The portrait construction module includes a resume data acquisition unit, a project data acquisition unit, a test data acquisition unit and a talent portrait construction unit; among which, The resume data acquisition unit is used to acquire the resume data of the candidate and extract the corresponding skill feature data according to the resume data, wherein the skill feature data includes the skill items mastered, the mastery degree and experience corresponding to the skill items, and the qualification certificates obtained; The project data acquisition unit is used to acquire project feature data corresponding to the candidate from the project publishing source, wherein the project feature data includes the number of projects, the time when the projects occurred, and the skill features associated with the projects; The test data acquisition unit is used to acquire the test feature data of the candidate, wherein the test feature data includes a logic test result and a personality test result; The talent profile building unit is used to build a talent profile of a candidate based on the candidate's skill feature data, project feature data and test feature data.

3. According to claim 2, the enterprise recruitment management system based on talent portrait is characterized in that: The resume data acquisition unit includes: Get the candidate's resume text; Perform text vectorization processing on the acquired resume text, extract the text feature vector of the resume data, and extract key feature words based on the text feature vector to obtain the corresponding skill feature text; The acquired skill feature text is further subjected to text mapping processing to obtain the skill feature text and the mapping text as skill feature data.

4. According to claim 2, the enterprise recruitment management system based on talent portrait is characterized in that: The talent portrait construction unit specifically includes: performing feature quantification processing according to the obtained candidate talent feature data, and extracting feature quantities of the candidate feature data; wherein the candidate talent feature data includes the candidate's skill feature data, project feature data and test feature data; The evaluation factor of the corresponding talent characteristic item is calculated based on the obtained characteristic quantity, and the evaluation factor calculation function used is: ； in, Mark(i) Indicates the corresponding talent characteristics i The evaluation factor, x(i) Indicates the talent characteristics of the candidate i The characteristic quantity of μ(i) and σ(i) Respectively represent the talent characteristics items in the talent pool i The average feature quantity and standard deviation of the feature quantity; represents the preset non-zero adjustment item, where ; Based on the obtained talent characteristic items and corresponding evaluation factors of the candidate, a talent profile of the candidate is constructed, wherein the talent profile of the candidate includes the characteristic quantities of various talent characteristics of the candidate and the corresponding evaluation factors.

5. According to claim 1, the enterprise recruitment management system based on talent portrait is characterized in that: It also includes a job requirements extraction module; The job requirement extraction module is used to obtain the published recruitment job feature data and extract job requirement features based on the obtained recruitment job feature data; The job requirement characteristics also include minimum standard requirement characteristics.

6. The enterprise recruitment management system based on talent portrait according to claim 1 is characterized in that: The job matching module includes a feature integration unit, a condition analysis unit and a matching analysis unit; among them, The feature integration unit is used to integrate the job demand features under the same job demand project according to the acquired job demand features, and obtain the job demand feature vector of each job demand project. VJ i ; and according to the candidate's talent portrait, the talent feature data under the talent feature project corresponding to the job requirement project is integrated to obtain the talent feature vector under the corresponding talent feature project VC i ; Among them, the job requirement feature vector and the talent feature vector include the feature quantity or evaluation factor of the corresponding project; The condition analysis unit is used to perform condition comparison analysis based on the minimum standard items set in the job requirement characteristics and the corresponding talent characteristic items in the candidate talent profile, and obtain the condition comparison analysis results of the candidate corresponding to each minimum standard characteristic; The matching analysis unit is used to perform matching analysis based on the acquired job requirement characteristics and talent portraits to obtain candidate job matching analysis results.

7. The enterprise recruitment management system based on talent portrait according to claim 1 is characterized in that: The interview generation module includes a question library unit, a question recommendation unit and a generation unit; among them, The question library unit is used to store and manage interview questions for recruitment positions, where each interview question contains the corresponding question feature item label i and the corresponding problem feature vector label VT i ; The question recommendation unit is used to calculate the priority of each interview question in the question library unit according to the candidate job matching analysis results. The priority calculation function used is: ； In the formula, PR t (i,VT i )surface Interview Questions t Priority of interview questions t The corresponding problem feature items are i , the corresponding style feature vector is VT i , P(i) Indicates the candidate's requirements for the position i The matching degree of P(i)=sim(VC i , VJ i ) , VC i Indicates that the candidate has a high level of talent i The talent feature vector under VJ i Indicates that the recruitment position is targeted at the job demand project i The job requirement feature vector, sim(VC i ,VJ i ) express VC i and VJ i The similarity is calculated using the cosine function; MatchP represents the candidate's job matching factor, θP represents the preset standard matching factor threshold, k represents the set sensitivity factor, where k∈[0.1,1000] , sim(VT i ,VC i ) express VT i and VC i The similarity of sim(VT i ,VJ i ) express VT i and VJ i similarity; the generating unit is used to select appropriate interview questions from the question library unit according to the priority of each interview question and based on the preset topic selection rules, and generate a corresponding interview question list.

8. The enterprise recruitment management system based on talent portrait according to claim 1 is characterized in that: The result management module further includes: According to the generated interview question list, the corresponding interview questions are obtained and pushed to the candidates, and the interview results of the candidates for the interview questions are obtained and displayed.

Images