Personnel comprehensive ability digital think tank system, method and device and medium

Abstract

The invention provides a personnel comprehensive ability digital think tank system, method and device and a medium, and relates to the field of personnel and post matching. The system comprises a preprocessing module, a feature extraction module, a post demand module, a matching degree module and an interaction module. The pre-processing module is configured to determine candidate person information and a plurality of historical enrollment information. The feature extraction module is configured to determine first features of candidate members based on the candidate member information; and determining a second feature corresponding to the plurality of pieces of historical recording information based on the plurality of pieces of historical recording information. The post demand module is configured to determine a third feature based on the second feature, the post information of the to-be-recruited post and the historical post information. The matching degree module is configured to determine a post matching degree between the candidate member and the to-be-recruited post. The interaction module is configured to display the post matching degree. According to the invention, talents with high matching degree with posts can be screened out, and more accurate and efficient recruitment or post allocation can be realized.

Description

Technical Field

[0001] The present invention relates to the field of person-job matching, and in particular to a digital think tank system, method, device and medium for comprehensive personnel capabilities. Background Art

[0002] At present, modern enterprises have increasingly higher requirements for the accuracy of matching people with jobs during recruitment or job allocation. Traditional recruitment or job allocation methods have problems such as low efficiency and low matching degree, which makes it difficult to meet the rapid and accurate needs of modern enterprises for talents.

[0003] Therefore, it is necessary to provide a digital think tank system, method, device and medium for comprehensive personnel capabilities to screen out talents with a high degree of match with the position and achieve more accurate and efficient recruitment or job allocation. Summary of the invention

[0004] The invention content includes a digital think tank system for comprehensive personnel capabilities. The system includes: a preprocessing module, a feature extraction module, a job requirement module, a matching degree module and an interaction module; the preprocessing module is configured to determine the candidate information based on the candidate's resume data; based on multiple historical resume data of the positions to be recruited, multiple historical recruitment information is determined; the feature extraction module is configured to determine the first feature of the candidate based on the candidate information; based on the multiple historical recruitment information, the second feature corresponding to the multiple historical recruitment information is determined; the job requirement module is configured to determine the third feature based on the second feature, the job information of the position to be recruited and the historical job information; the matching degree module is configured to determine the job matching degree between the candidate and the position to be recruited based on the first feature and the third feature, and send the job matching degree to the interaction module; the interaction module is configured to display the job matching degree.

[0005] The invention content includes providing a digital think tank method for comprehensive personnel capabilities. The method is executed by a processor, and the method includes: determining candidate information based on the resume data of the candidate; determining multiple historical recruitment information based on multiple historical resume data of the position to be recruited; determining a first feature of the candidate based on the candidate information; determining a second feature corresponding to the multiple historical recruitment information based on the multiple historical recruitment information; determining a third feature based on the second feature, the position information of the position to be recruited, and the historical position information; determining the position matching degree between the candidate and the position to be recruited based on the first feature and the third feature; and displaying the position matching degree.

[0006] The invention content includes a digital think tank device for comprehensive capabilities of personnel, which includes at least one processor and at least one memory; the at least one memory is used to store computer instructions; the at least one processor is used to execute at least part of the computer instructions to implement the digital think tank method for comprehensive capabilities of personnel described in this specification.

[0007] The invention content includes a computer-readable storage medium, which stores computer instructions. When a computer reads the computer instructions in the storage medium, the computer executes the digital think tank method for comprehensive personnel capabilities described in this specification.

[0008] The beneficial effects of the present invention include but are not limited to: (1) through the relevant information of historically hired personnel, more objective job requirements can be determined, and then the job matching degree between candidate personnel and the job to be recruited can be more accurately evaluated, providing an accurate reference for recruitment; (2) based on the job performance information, the second feature can be screened out based on the second feature of the historically hired personnel, which is conducive to determining a more accurate third feature; (3) when determining the job matching degree, the hard matching degree and soft matching degree of the candidate personnel are comprehensively considered, so as to obtain a more reference job matching degree. BRIEF DESCRIPTION OF THE DRAWINGS

[0009] This specification will be further described in the form of exemplary embodiments, which will be described in detail by the accompanying drawings. These embodiments are not restrictive, and in these embodiments, the same number represents the same structure, wherein: Figure 1 is a system structure diagram of a digital think tank system for comprehensive human capabilities according to some embodiments of this specification; Figure 2 is an exemplary flow chart of a digital think tank method for comprehensive human capabilities according to some embodiments of this specification; Figure 3 is an exemplary schematic diagram of a job requirement model according to some embodiments of this specification; Figure 4 This is an exemplary flow chart for determining job matching according to some embodiments of this specification. DETAILED DESCRIPTION

[0010] In order to more clearly illustrate the technical solutions of the embodiments of this specification, the following is a brief introduction to the drawings required for the description of the embodiments. Obviously, the drawings described below are only some examples or embodiments of this specification. For ordinary technicians in this field, this specification can also be applied to other similar scenarios based on these drawings without creative work. Unless it is obvious from the language environment or otherwise explained, the same reference numerals in the figures represent the same structure or operation.

[0011] Unless the context clearly indicates an exception, the words "a", "an", "an" and / or "the" do not refer to the singular and may also include the plural. Generally speaking, the terms "include" and "comprise" only indicate the inclusion of the steps and elements that have been clearly identified, and these steps and elements do not constitute an exclusive list. The method or device may also include other steps or elements.

[0012] Figure 1 It is a system structure diagram of the digital think tank system for comprehensive personnel capabilities shown in some embodiments of this specification.

[0013] In some embodiments, the digital think tank system 100 for comprehensive personnel capabilities may include a preprocessing module 110 , a feature extraction module 120 , a job requirement module 130 , a matching degree module 140 and an interaction module 150 .

[0014] In some embodiments, the pre-processing module 110 is configured to determine candidate information based on the resume data of the candidate; and determine multiple historical recruitment information based on multiple historical resume data of the positions to be recruited.

[0015] In some embodiments, the feature extraction module 120 is configured to determine a first feature of a candidate based on the candidate information; and to determine a second feature corresponding to multiple historical recruitment information based on multiple historical recruitment information.

[0016] In some embodiments, the job requirement module 130 is configured to determine the third feature based on the plurality of second features, the job information of the job to be recruited, and the historical job information.

[0017] In some embodiments, the matching module 140 is configured to determine the job matching degree between the candidate and the job to be recruited based on the first feature and the third feature, and send the job matching degree to the interaction module.

[0018] In some embodiments, the interaction module 150 is configured to display the job matching degree. In some embodiments, the interaction module 150 may include a display device or a terminal interface.

[0019] In some embodiments, the job requirement module 130 can be further configured to screen the second feature based on the job performance information of historical recruits corresponding to the second feature to obtain the target second feature; and determine the third feature through the job requirement model based on the target second feature, job information and historical job information.

[0020] In some embodiments, the matching module 140 can be further configured to determine the candidate's historical employment information based on the candidate information; determine the candidate's hard matching degree based on the first feature and the third feature; determine the candidate's soft matching degree based on the historical employment information and job information; and determine the job matching degree based on the hard matching degree and the soft matching degree.

[0021] In some embodiments, the matching module 140 can be further configured to predict the estimated employment information of the candidate through an information prediction model based on historical employment information and job information; determine reference employment information based on the employment performance information of historically hired personnel corresponding to the second feature; and determine the soft matching degree based on the estimated employment information and the reference employment information.

[0022] In some embodiments, the personnel comprehensive capability digital think tank system 100 may further include a processor and a memory. The processor is configured to process data from at least one module of the personnel comprehensive capability digital think tank system 100 or an external data source. The processor includes a central processing unit (CPU), an application specific integrated circuit (ASIC), a controller, a microcontroller unit, a microprocessor, etc. or any combination thereof. In some embodiments, the preprocessing module, the feature extraction module, the job requirement module, the matching module, and the interaction module may be components of the processor.

[0023] The memory is configured to store data, instructions and / or information related to the personnel comprehensive capability digital think tank system 100. For example, the memory stores historical job information, etc. In some embodiments, the memory can be integrated into the processor, and can also be set in the cloud, etc.

[0024] For more information about the Digital Think Tank System 100 for Comprehensive Human Resources Capabilities, see Figure 2-Figure 4 and its related description.

[0025] In some embodiments of this specification, through the digital think tank system of comprehensive personnel capabilities, the job matching degree between candidate personnel and the positions to be recruited can be quickly determined, thereby assisting in determining whether to hire the candidate personnel, reducing recruitment costs and improving recruitment efficiency.

[0026] Figure 2 It is an exemplary flow chart of the digital think tank method for comprehensive personnel capabilities shown in some embodiments of this specification.

[0027] like Figure 2 As shown, the process 200 includes the following steps. In some embodiments, the process 200 may be executed by a processor.

[0028] Step 210, determine the candidate information based on the resume data of the candidate; determine multiple historical recruitment information based on multiple historical resume data of the positions to be recruited.

[0029] Candidates refer to persons who are yet to be confirmed for employment.

[0030] The positions to be recruited refer to positions that require recruitment. It is understandable that candidates correspond to the positions to be recruited, and candidates submit resume data for the positions to be recruited in the hope of being hired.

[0031] Resume data refers to files or files that record relevant information about candidates.

[0032] Historical resume data refers to the resume data of historically hired personnel corresponding to the position to be recruited. Historically hired personnel refer to candidates who were determined to be hired at a historical time. In some embodiments, there may be multiple historical recruitments for the position to be recruited at a historical time, and the resume data of historically hired personnel corresponding to each historical recruitment is a set of historical resume data.

[0033] In some embodiments, the processor may obtain resume data and historical resume data from a memory or user input.

[0034] The candidate information refers to information related to the candidate's own situation. In some embodiments, the candidate information includes at least one of the candidate's age, education, major, work experience, work skills, etc.

[0035] In some embodiments, the processor may determine the candidate information based on the candidate's resume data in a variety of ways. For example, the processor may pre-process the candidate's resume data to obtain standardized resume data, and determine the candidate information based on the standardized resume data.

[0036] Standardized resume data refers to resume data represented by a preset representation method. In some embodiments, the preset representation method includes a label representation method, etc. The preprocessing process includes: the processor identifies whether there is information corresponding to the label in the resume data by means of text recognition, etc. If there is information corresponding to the label in the resume data, the label is used to represent the corresponding information. Among them, there can be multiple labels, which are preset by the processor or determined by user input. Each label corresponds to a piece of text or information, etc.

[0037] For example, the resume data describes work skills as "proficient in multiple programming languages, able to independently complete projects using Java, Python, etc.", and the processor recognizes that there is information corresponding to the tags "Java" and "Python" in the resume data, and uses the tags "Java" and "Python" to represent the work skills in the resume data.

[0038] In some embodiments, the processor extracts multiple tags from the standardized resume data and combines the multiple tags to obtain candidate information.

[0039] In some embodiments, the processor may also determine candidate personnel information through an information extraction model.

[0040] The information extraction model refers to a model used to extract candidate information. In some embodiments, the information extraction model is a machine learning model, such as a convolutional neural network (CNN) model or any combination of other custom model structures.

[0041] In some embodiments, the information extraction model can be obtained by training based on a large number of first training samples with first labels through a gradient descent method or the like. The first training samples include sample resume data of sample candidates, and the first labels include actual candidate information of the sample candidates. In some embodiments, the first training samples can be obtained based on historical data, and the first labels can be obtained by manual annotation.

[0042] In some embodiments, the training process of the information extraction model includes: inputting a plurality of first training samples with first labels into the initial information extraction model, constructing a loss function through the training labels and the prediction results of the initial information extraction model, iteratively updating the initial information extraction model based on the loss function, and completing the information extraction model training when the loss function of the initial information extraction model meets a preset end condition. The preset end condition may include the convergence of the loss function or the number of iterations reaching a set value, etc.

[0043] The historical recruitment information refers to information related to the personal situation of the historically recruited personnel. In some embodiments, the historical recruitment information is similar to the candidate information, and the acquisition method of the historical recruitment information is similar to the acquisition method of the candidate information.

[0044] Step 220, based on the candidate information, determine the first feature of the candidate; based on the multiple historical recruitment information, determine the second features corresponding to the multiple historical recruitment information.

[0045] The first feature refers to information related to the position to be recruited in the candidate information. In some embodiments, the first feature may include the candidate's education, major, and work experience and / or work skills related to the position to be recruited.

[0046] In some embodiments, the processor can extract the candidate's education, major, and work experience and / or work skills related to the position to be recruited from the candidate information by text extraction and other methods to obtain the first feature. For example, if the position to be recruited is a front-end programmer, and the work experience in the candidate information includes "restaurant internship" and "software company internship", the processor can extract information related to "software company internship" and include it in the first feature. The work experience and / or work skills related to the position to be recruited can be preset based on historical experience.

[0047] It is understandable that the candidate information may contain information that is irrelevant to the job to be recruited. By eliminating the information that is irrelevant to the job to be recruited from the candidate information, it is helpful to determine a more accurate job matching degree later.

[0048] The second feature refers to information related to the job to be recruited in the historical recruitment information. In some embodiments, the second feature is similar to the first feature, and the method for obtaining the second feature is similar to the method for obtaining the first feature.

[0049] Step 230, determining a third feature based on the second feature, the job information of the job to be recruited, and the historical job information.

[0050] Position information refers to information related to the position to be recruited. In some embodiments, the position information may include at least one of the age requirement, education requirement, work experience requirement, position type, position responsibilities, industry to which the position belongs, and work skill requirements of the position to be recruited.

[0051] Historical job information refers to information related to the job to be recruited at a historical time. It is understandable that the requirements for the same job to be recruited may be different in multiple historical recruitments at different times. Each historical recruitment corresponds to a historical job information.

[0052] In some embodiments, the processor may obtain the job information and historical job information through a memory or user input.

[0053] The third feature refers to key information in the job information. In some embodiments, the type of content included in the third feature is similar to the type of content included in the first feature.

[0054] In some embodiments, the processor can determine the third feature in a variety of ways based on multiple second features, job information of the job to be recruited, and historical job information. For example, the processor can select historical job information from multiple historical job information that meets similarity conditions with the job information as the target job information, and determine one or more historical recruits corresponding to the target job information as the target personnel, filter out the second features corresponding to the target personnel from multiple second features, and combine the minimum requirements of each data in the second features corresponding to the target personnel into the third feature. For example, the range of academic qualifications in the second feature corresponding to the target personnel is junior college or undergraduate, then the minimum requirement for academic qualifications in the third feature is junior college. For another example, the work skills in the second feature corresponding to the target personnel include Java, C++, Python, etc., then the minimum requirement for work skills in the third feature is at least one of Java, C++, Python, etc.

[0055] In some embodiments, the processor may construct feature vectors based on the historical job information and the job information, and represent the similarity between the historical job information and the job information through the similarity between the feature vectors. The similarity between the feature vectors is negatively correlated to the distance between the feature vectors. The vector distance includes the Euclidean distance, etc.

[0056] In some embodiments, the processor may also determine the third feature through a job requirement model, see Figure 3 and related instructions.

[0057] It is understandable that the current job information is determined subjectively by the user and does not necessarily conform to the actual situation of the job to be recruited. Through the relevant information of the people hired in multiple historical recruitments, the key requirements for being competent for the job to be recruited can be determined, which is conducive to the subsequent determination of a job match that is more in line with the actual situation.

[0058] Step 240: Determine the job matching degree between the candidate and the job to be recruited based on the first feature and the third feature.

[0059] Job matching refers to the data used to characterize the degree of matching between candidates and the jobs to be recruited.

[0060] In some embodiments, the processor may determine the job matching degree between the candidate and the job to be recruited in a variety of ways based on the first feature and the third feature. For example, the processor may determine the individual score of each data item according to a preset scoring rule based on the difference between each data item in the first feature and the third feature, and use the sum of multiple individual scores as the job matching degree.

[0061] In some embodiments, the preset scoring rules may be preset by the processor or obtained by user input. The preset scoring rules include basic scores and addition and subtraction rules corresponding to multiple data in the third feature.

[0062] For example, if the educational requirement in the third feature is a bachelor's degree, the basic score corresponding to the educational requirement is 6. If the educational requirement in the first feature is one level higher than the educational requirement in the third feature (such as a master's degree is one level higher than a bachelor's degree), then the individual score of the educational requirement is 7. When the educational requirement in the first feature is one level lower than the educational requirement in the third feature (such as a junior college degree is one level lower than a bachelor's degree), then the individual score of the educational requirement is 5.

[0063] If the job skills required in the third feature include Java, C++, and Python, the basic score corresponding to the job skills is 6. If the job skills in the first feature include Java and C++ but not Python, the single score of the job skills is 5. If the job skills in the first feature include Java, C++, and Python, as well as other skills related to the position to be recruited, the single score of the job skills is the sum of the number of other skills related to the position to be recruited and the basic score.

[0064] If the time span of work experience in the third feature is three years, the basic score for work experience is 6 points. If the time span of work experience in the first feature is four years, the score for work experience is 7 points. If the time span of work experience in the first feature is two years, the score for work experience is 5 points.

[0065] In some embodiments, the processor may also determine the job matching degree based on the hard matching degree and the soft matching degree. For more information, see Figure 4 and related instructions.

[0066] Step 250, displaying the job matching degree.

[0067] In some embodiments, the processor may display the job matching degree through the interactive module. For example, the processor may send the job matching degree to the interactive module through a network, etc., to display the job matching degree.

[0068] In some embodiments of this specification, through the relevant information of historically hired personnel, more objective job requirements can be determined, and then the job matching degree between candidate personnel and the job to be recruited can be more accurately evaluated, providing an accurate reference for recruitment. Intuitively displaying the job matching degree to the user can improve the efficiency of recruitment.

[0069] Figure 3 It is an exemplary schematic diagram of a job requirement model according to some embodiments of this specification.

[0070] In some embodiments, the processor may screen the second feature based on the job performance information of historically employed personnel corresponding to the second feature to obtain the target second feature; and determine the third feature through the job requirement model 330 based on the target second feature 310 , job information 350 , and historical job information 320 .

[0071] For descriptions of the second feature, job information, historical job information, and the third feature, see Figure 2 .

[0072] Job performance information refers to information related to the performance of historically hired personnel after joining the company. In some embodiments, the job performance information includes the task completion rate during the probation period, the job grade promotion rate after regularization, and whether the employee has been dismissed. The job grade promotion rate of a person can be represented by the ratio of the number of job grades promoted by the person to the total working hours of the person. The task completion rate refers to the ratio of the number of tasks completed on time to the total number of tasks.

[0073] In some embodiments, the job performance information can be obtained by manual input by the user or from a memory. The user may include management personnel, personnel personnel, etc. of the unit where the historically hired personnel work. The processor automatically records the task completion rate, job grade and other work performance-related information of the historically hired personnel during their employment, and stores it in the memory.

[0074] The target second feature refers to the second feature corresponding to the historical employed personnel whose job performance information meets the screening conditions.

[0075] In some embodiments, the processor screens the plurality of second features based on the screening condition, removes the second features that do not meet the screening condition, and obtains the target second feature. There may be multiple target second features.

[0076] In some embodiments, the screening conditions may include that the task completion rate of historical recruits during the probation period is not less than a completion rate threshold and / or the job grade promotion rate after regularization is not less than a promotion rate threshold.

[0077] In some embodiments, the completion rate threshold and the rising rate threshold can be determined in a variety of ways. For example, the completion rate threshold and the rising rate threshold are system preset values ​​or set by the user.

[0078] For example, the completion rate threshold may be determined based on the assessment requirements of the unit where the historically employed personnel work. The promotion rate threshold may be 0, and if the promotion rate after regularization is not less than the promotion rate threshold, it means that the historically employed personnel have not been demoted during their employment.

[0079] In some embodiments, the processor may also determine a completion rate threshold and an increase rate threshold based on the change characteristics of historical job information.

[0080] The change characteristics of historical job information refer to information used to reflect the change direction and magnitude of historical job information.

[0081] In some embodiments, the change characteristics of historical job information can be represented by numerical values. A positive change characteristic indicates that the requirements of the job to be recruited have increased, and a negative change characteristic indicates that the requirements of the job to be recruited have decreased. The larger the absolute value of the change characteristic, the more the requirements of the job to be recruited have increased or decreased. The processor can determine the change characteristics corresponding to each data based on the change of each data in multiple historical job information, and determine the sum of the change characteristics corresponding to all data as the change characteristics of the historical job information. For example, the processor traverses multiple historical job information in chronological order. If the academic qualification requirements in the historical job information are increased from junior college to undergraduate, the change characteristics corresponding to the academic qualification requirements are +1. If the work skill requirements are reduced from 5 skills to 4 skills, the change characteristics corresponding to the work skill requirements are -1.

[0082] In some embodiments, the completion rate threshold and the rising rate threshold are positively correlated with the change characteristics of the historical job information. The processor can calculate the completion rate threshold and the rising rate threshold through the following formulas (1) and (2).

[0083] (1) (2) in, is the completion rate threshold, is the initial completion rate threshold, For the changing characteristics of historical job information, is the rising rate threshold, is the initial rising rate threshold, , is the coefficient. The initial completion rate threshold, initial rise rate threshold and and It is the system default value or set by the user.

[0084] The job requirement model is a model used to determine the third feature. In some embodiments, the job requirement model is a machine learning model, for example, any one or combination of a convolutional neural network (CNN) model, a deep neural network (DNN) model, or other custom model structures.

[0085] In some embodiments, the input of the job requirement model 330 includes the target second feature 310 , job information 350 , and historical job information 320 , and the output includes the third feature 360 ​​.

[0086] In some embodiments, the processor may obtain a job requirement model based on a plurality of second training samples with second labels by training using a gradient descent method or the like.

[0087] The second training sample includes the sample target second feature, sample job information, and sample historical job information corresponding to the sample job to be recruited. The second label includes the second feature of the actual recruited person corresponding to the sample job to be recruited. In some embodiments, the processor can use the historical job information and target second feature of the sample job to be recruited in the historical data at the first historical time as the sample historical job information and the sample target second feature, and use the job information corresponding to the second historical time as the sample job information.

[0088] In some embodiments, the processor can obtain the second features of multiple actual recruits corresponding to the sample positions to be recruited at the third historical time from the historical data, and combine the minimum requirements of each data item in the second features of the multiple actual recruits to obtain the second label corresponding to the sample positions to be recruited. The first historical time is earlier than the second historical time, and the second historical time is earlier than the third historical time. For an explanation of the minimum requirements, see Figure 2 and its related description.

[0089] In some embodiments, the training process of the job requirement model is similar to the training process of the information extraction model, and its implementation method can refer to the training process of the information extraction model.

[0090] In some embodiments, Figure 3 As shown, the job requirement model 330 includes a feature embedding layer 331 and a requirement prediction layer 332 .

[0091] The feature embedding layer is used to determine job matching features. In some embodiments, the feature embedding layer is a machine learning model, such as a convolutional neural network (CNN) model.

[0092] The input of the feature embedding layer 331 includes the target second feature 310 and the corresponding historical job information 320 , and the output includes the job matching feature 340 .

[0093] The position matching feature refers to data used to reflect the degree of matching between the target second feature and the corresponding historical position information. In some embodiments, if there are multiple target second features input to the feature embedding layer, the output of the feature embedding layer includes a position matching feature sequence. The position feature matching sequence refers to a sequence composed of multiple position matching features, where one position matching feature corresponds to one target second feature.

[0094] In some embodiments, the multiple job matching features in the job matching feature sequence are arranged in chronological order according to the corresponding target second features.

[0095] The demand prediction layer is used to determine the third feature. In some embodiments, the demand prediction layer is a machine learning model, such as a deep neural network (CNN) model.

[0096] The input of the demand prediction layer 332 includes a job matching feature sequence or a job matching feature 340 and job information 350 , and the output includes a third feature 360 ​​.

[0097] In some embodiments, the processor may obtain the feature embedding layer and the demand prediction layer through joint training based on the training data set. The training data set includes multiple training samples with training labels, wherein the training samples include a second training sample, and the training label includes a second label corresponding to the second training sample.

[0098] Exemplarily, multiple groups of sample target second features and sample historical job information are input into the initial feature embedding layer to obtain a job matching feature sequence output by the initial feature embedding layer, the job matching feature sequence and multiple sample job information are input into the initial demand prediction layer, a loss function is constructed based on the output of the initial demand prediction layer and the second label, the initial feature embedding layer and the initial demand prediction layer are updated based on the iteration of the loss function, and the training of the feature embedding layer and the demand prediction layer is completed when the loss function meets the preset end condition.

[0099] In some embodiments, the output of the demand prediction layer may further include multiple third features. The multiple third features include the first demand feature and multiple second demand features. When the output of the demand prediction layer includes multiple third features, the training labels of the training samples in the training data set also include training labels corresponding to the first demand feature and the multiple second demand features.

[0100] The first demand characteristic refers to the third characteristic that can reflect the general demand of the position to be recruited.

[0101] The second requirement feature refers to the third feature that can reflect the special requirements of the position to be recruited. For example, if the candidate's work experience spans more than ten years, the candidate's education or major can be ignored, and the second requirement feature includes the work experience spanning more than ten years.

[0102] In some embodiments, when calculating the job matching degree, the processor can also determine the job matching degree corresponding to the second requirement feature and / or the first requirement feature based on the first feature and the second requirement feature and / or the first requirement feature, and display the job matching degree corresponding to the second requirement feature and / or the first requirement feature.

[0103] In some embodiments, when the output of the demand prediction layer includes a first demand feature and multiple second demand features, the processor can obtain the second features of multiple actual recruits corresponding to the sample positions to be recruited at the third historical time from the historical data, and combine the frequent items of each data in the second features of the multiple actual recruits to obtain the training label corresponding to the first demand feature. Among them, the frequent item refers to the data with the most occurrences in multiple data of the same type. For example, the number of bachelor's degrees in the second features of multiple actual recruits is the largest, and the frequent item corresponding to the degree is bachelor's degree.

[0104] In some embodiments, the processor may count the individual sub-features in the historical second feature corresponding to multiple training samples, and use the historical second feature corresponding to the individual sub-feature that meets the training conditions as the training label corresponding to the second requirement feature. The historical second feature corresponding to the training sample refers to the second feature of the actual employed personnel corresponding to the training sample. The individual sub-feature refers to the individual data in the second feature. The training condition includes that the individual sub-feature is the minimum requirement among multiple individual sub-features.

[0105] Exemplarily, the processor counts the educational background in the historical second feature corresponding to multiple training samples, selects the lowest educational background (such as high school graduation) among the historical second features corresponding to multiple training samples, and uses the historical second feature corresponding to the lowest educational background as a training label corresponding to a second demand feature.

[0106] In some embodiments of the present specification, by outputting multiple third features through the job requirement model, the job matching degree corresponding to the multiple third features can be obtained, thereby providing users with multi-angle references to avoid screening out candidates who do not meet some requirements but are qualified for the job to be recruited because of their outstanding advantages.

[0107] In some embodiments of the present specification, the second feature is screened based on the job performance information, and the second features corresponding to more outstanding historical recruits can be screened out. Based on the target second feature obtained by the screening, a more accurate third feature can be determined through the demand prediction model, which is conducive to improving the accuracy of subsequent determination of job matching.

[0108] Figure 4 This is an exemplary flow chart for determining job matching according to some embodiments of this specification.

[0109] like Figure 4 As shown, the process 400 includes the following steps. In some embodiments, the process 400 may be executed by a processor.

[0110] Step 410, based on the candidate information, determine the candidate's historical employment information.

[0111] For information about candidate information, see Figure 2 .

[0112] Historical employment information refers to information related to the candidate's work experience. In some embodiments, historical employment information includes the work unit attributes, work content, work skills, job level promotion rate, and time span corresponding to the work experience. Among them, the work unit attributes include business operations, unit size, etc.

[0113] In some embodiments, if the candidate has multiple work experiences, the processor uses the work unit attributes, work content, work skills, and time span corresponding to each work experience as a historical employment information.

[0114] In some embodiments, the processor extracts information related to different work experiences from the candidate information by extracting keywords and tags, etc., based on the candidate information, to obtain historical employment information. For example, the candidate information includes two tags, front-end development and back-end development, and the processor extracts the front-end development work experience and the back-end development work experience through the tags.

[0115] Step 420: Determine the hard matching degree of the candidate based on the first feature and the third feature.

[0116] For the description of the first and third characteristics, see Figure 2 .

[0117] Hard matching refers to the degree of match between the hard qualities of the candidate and the position to be recruited. Hard qualities include the work content or work skills that the candidate has completed.

[0118] In some embodiments, the processor may determine the degree of hard match in a variety of ways.

[0119] In some embodiments, the processor may Figure 2 The job matching degree determined in is used as the hard matching degree.

[0120] In some embodiments, when there are multiple job matching degrees, the processor may use the job matching degree with the highest value among the multiple job matching degrees as the hard matching degree.

[0121] In some embodiments, the processor may determine the experience matching degree of the candidate based on the historical employment information and the position information; and determine the hard matching degree based on the experience matching degree, the first feature, and the third feature.

[0122] Experience matching refers to the degree of match between the candidate's work experience and the position to be recruited.

[0123] In some embodiments, the processor may feature vectorize the historical employment information and the job information, calculate the similarity between the two feature vectors, and represent the similarity between the feature vectors. If there are multiple historical employment information, the processor may calculate the similarity corresponding to the multiple historical employment information, and determine the experience matching degree based on the historical employment information with the greatest similarity.

[0124] In some embodiments, the experience matching degree is negatively correlated with the time interval between the end time of the time span in the historical employment information and the current time, and is positively correlated with the similarity. For example, the processor can calculate the experience matching degree using the following formula (3).

[0125] (3) in, is the experience matching degree, is the similarity, is the time interval between the end time of the time span and the current time, is the coefficient, The value can be preset for the system.

[0126] In some embodiments, the processor may, based on the first feature and the third feature, Figure 2 The hard matching degree is determined by the method of determining the job matching degree in , and the sum of the hard matching degree and the experience matching degree is determined as the new hard matching degree.

[0127] In some embodiments of this specification, since the candidate may have less work experience and may not be able to perform the job well, focusing on the matching degree of work experience when determining the hard matching degree can provide users with a more accurate reference.

[0128] Step 430, determining the soft matching degree of the candidate based on the historical employment information and position information.

[0129] For job information and candidate information, see Figure 2 .

[0130] Soft matching refers to the degree of match between the soft qualities of the candidate and the position to be recruited. Soft qualities include the adaptability of personnel to the position to be recruited and the ability to transfer experience. The ability to transfer experience refers to the ability to apply past experience to the position to be recruited. Since soft qualities cannot be quantified, they can be represented by the time span of historical employment information. The larger the time span, the more adaptable the candidate is to the position to be recruited and the stronger the ability to transfer experience.

[0131] In some embodiments, the processor may determine the weight corresponding to the one or more historical employment information based on the similarity between the one or more historical employment information and the job information, and perform weighted summation on the time span corresponding to the one or more historical employment information to obtain the soft matching degree. The greater the similarity between the historical employment information and the job information, the greater the weight of the time span corresponding to the historical employment information.

[0132] In some embodiments, the processor can predict the estimated employment information of the candidate based on historical employment information and position information through an information prediction model; determine reference employment information based on the employment performance information of historically hired personnel corresponding to multiple second features; and determine the soft matching degree based on the estimated employment information and the reference employment information.

[0133] For a description of the second feature, see Figure 2 .

[0134] The estimated employment information refers to the estimated information related to the candidate's work performance after joining the company. In some embodiments, the estimated employment information includes the estimated working hours, job promotion rate, etc. of the candidate after joining the company.

[0135] The information prediction model is a model used to determine the estimated employment information. In some embodiments, the information prediction model is a machine learning model, such as a recurrent neural network (RNN) model.

[0136] The input of the information prediction model includes historical employment information and job information, and the output is estimated employment information.

[0137] In some embodiments, if a candidate has multiple historical employment information, the processor can arrange the multiple historical employment information in chronological order to obtain a historical employment information sequence, and input the historical employment information sequence and position information into an information prediction model to obtain estimated employment information.

[0138] In some embodiments, the processor may train the information prediction model based on a plurality of third training samples with third labels by using a gradient descent method or the like. The training process of the information prediction model is similar to that of the information extraction model. For instructions on how to train the information extraction model, see Figure 2 .

[0139] The third training sample includes sample historical employment information and sample job information of the sample candidate. The third label includes the actual working hours and actual job grade promotion rate of the sample candidate after joining the company. In some embodiments, the processor can use the historically hired personnel in the historical data as the sample candidate, determine the historical employment information of the sample candidate as the sample historical employment information, determine the job information corresponding to the sample candidate as the sample job information, and use the actual working hours and actual job grade promotion rate of the sample candidate after joining the company as the third label. The actual working hours and actual job grade promotion rate of the sample candidate after joining the company are obtained through the memory or obtained by user input.

[0140] In some embodiments, the input to the information prediction model includes an empirical match.

[0141] In some embodiments, when the input of the information prediction model includes experience matching, the third training sample also includes sample experience matching. The processor can calculate the sample experience matching based on the sample historical employment information and the sample position information.

[0142] In some embodiments of the present specification, if the experience match is low, the candidate may leave the job in a short period of time or have a slow promotion in rank due to being unable to adapt to the job to be recruited. When determining the estimated employment information, taking the experience match into consideration can obtain more accurate estimated employment information.

[0143] Reference employment information refers to employment performance information used for comparison with estimated employment information.

[0144] In some embodiments, the processor may determine reference employment information based on the employment performance information corresponding to the historically hired personnel corresponding to the plurality of second features. For example, the processor counts the employment performance information corresponding to the historically hired personnel corresponding to the plurality of second features, calculates the average working hours and the average job grade promotion rate, and uses the average working hours and the average job grade promotion rate as reference employment information.

[0145] In some embodiments, the processor may cluster the employment performance information corresponding to historically hired personnel corresponding to multiple target second features to obtain multiple clusters; and determine reference employment information based on the multiple clusters.

[0146] In some embodiments, the processor may construct multiple clustering vectors based on the job performance information corresponding to the historically employed personnel corresponding to the multiple target second features, and perform clustering based on the multiple clustering vectors to obtain multiple clustering clusters. The number of clustering clusters may be a system preset value. Clustering methods include but are not limited to K-Means algorithm, etc.

[0147] In some embodiments, the number of clusters is also related to the variance of the distances between the plurality of cluster vectors. The smaller the variance, the smaller the number of clusters.

[0148] In some embodiments, the processor performs weighted summation on the cluster centers of the plurality of clusters to obtain the reference employment information. The weight of the cluster is positively correlated to the number of cluster vectors in the cluster.

[0149] In some embodiments of the present specification, multiple employment performance information is clustered to obtain multiple representative clusters, and weighted summation is performed based on cluster centers of the clusters to obtain reference employment information with more reference value.

[0150] In some embodiments, the processor may construct feature vectors based on the estimated employment information and the reference employment information, respectively, and represent the similarity between the feature vectors to determine the soft matching degree based on the similarity, wherein the soft matching degree is positively correlated with the similarity.

[0151] It is understandable that if the estimated employment information is less similar to the reference employment information, it means that the candidate may not be able to adapt to the job to be recruited or the job to be recruited cannot meet the job needs of the candidate.

[0152] In some embodiments of this specification, the estimated employment information can be quickly obtained by determining the estimated employment information through the information prediction model. By comparing the estimated employment information with the reference employment information, a soft match that better meets the actual situation of the candidate can be determined to avoid the situation where the recruited candidate cannot adapt to the job after joining the company.

[0153] Step 440, determining the job matching degree based on the hard matching degree and the soft matching degree.

[0154] In some embodiments, the processor may determine the sum of the hard matching degree and the soft matching degree as the position matching degree.

[0155] In some embodiments, the processor may also perform weighted summation of the hard matching degree and the soft matching degree, and determine the result of the weighted summation as the job matching degree. The weight of the weighted summation is related to the correlation coefficient between the job grade promotion rate of the historically hired personnel corresponding to the job to be recruited in the historical data and the first feature corresponding to the historically hired personnel. In some embodiments, the processor may calculate the correlation coefficient between the job grade promotion rate and the first feature based on the job grade promotion rate corresponding to the historically hired personnel and the first feature through a preset algorithm. The preset algorithm includes the Pearson correlation coefficient method, etc. If there are multiple historically hired personnel, the processor calculates the correlation coefficients corresponding to the multiple historically hired personnel, and takes the average of the multiple correlation coefficients as the current correlation coefficient.

[0156] In some embodiments, if the correlation coefficient is not less than a preset correlation threshold, the weight of the hard match is greater than the weight of the soft match, and if the correlation coefficient is less than the preset correlation threshold, the weight of the hard match is less than the weight of the soft match. The preset correlation threshold may be a system default value, for example, the preset correlation threshold is 0.5, etc.

[0157] In some embodiments of the present specification, both hard qualities and soft qualities will affect the work performance of the recruited candidates after joining the company. When determining the job matching degree, comprehensively considering the hard matching degree and soft matching degree of the candidates can obtain a more reference job matching degree.

[0158] It should be noted that the above description of the process 200 and the process 400 is only for example and explanation, and does not limit the scope of application of this specification. For those skilled in the art, various modifications and changes can be made to the process under the guidance of this specification. However, these modifications and changes are still within the scope of this specification.

[0159] The present invention also provides a digital think tank device for comprehensive personnel capabilities, the device comprising at least one processor and at least one memory. The at least one memory is used to store computer instructions. The at least one processor is used to execute at least part of the computer instructions to implement a digital think tank method for comprehensive personnel capabilities described in any of the above embodiments.

[0160] The present invention also provides a computer-readable storage medium, which stores computer instructions. When a computer reads the computer instructions in the storage medium, the computer executes a digital think tank method for comprehensive personnel capabilities as described in any of the above embodiments.

[0161] Furthermore, certain features, structures or characteristics in one or more embodiments of this specification may be appropriately combined.

[0162] In some embodiments, numbers describing the quantity of components and attributes are used. It should be understood that such numbers used in the description of the embodiments are modified by the modifiers "about", "approximately" or "substantially" in some examples. Unless otherwise specified, "about", "approximately" or "substantially" indicates that the numbers allow a variation of ±20%.

[0163] If there is any inconsistency or conflict between the descriptions, definitions, and / or usage of terms in the materials cited in this specification and the contents described in this specification, the descriptions, definitions, and / or usage of terms in this specification shall prevail.

Claims

1. A digital think tank system for comprehensive personnel capabilities, characterized in that: The system includes a preprocessing module, a feature extraction module, a job requirement module, a matching degree module and an interaction module; The preprocessing module is configured to determine candidate information based on the resume data of the candidate; determine multiple historical recruitment information based on multiple historical resume data of the positions to be recruited; The feature extraction module is configured to determine a first feature of the candidate person based on the candidate person information; Based on the multiple pieces of historical recruitment information, determining a second feature corresponding to the multiple pieces of historical recruitment information; The position requirement module is configured to determine a third feature based on the second feature, the position information of the position to be recruited, and historical position information; The matching degree module is configured to determine the job matching degree between the candidate and the job to be recruited based on the first feature and the third feature, and send the job matching degree to the interaction module; The interaction module is configured to display the job matching degree.

2. The system according to claim 1, characterized in that The job requirement module is further configured as follows: Based on the employment performance information of historically employed personnel corresponding to the second feature, the second feature is screened to obtain a target second feature; Based on the target second feature, the job information and the historical job information, the third feature is determined through a job requirement model, and the job requirement model is a machine learning model.

3. The system according to claim 1, characterized in that The matching module is further configured to: Based on the candidate information, determine the candidate's historical employment information; Determining a hard matching degree of the candidate based on the first feature and the third feature; Determining the soft matching degree of the candidate based on the historical employment information and the position information; The position matching degree is determined based on the hard matching degree and the soft matching degree.

4. The system according to claim 3, characterized in that The matching module is further configured to: Based on the historical employment information and the position information, predicting the estimated employment information of the candidate through an information prediction model, wherein the information prediction model is a machine learning model; Determining reference employment information based on employment performance information corresponding to historically employed personnel corresponding to the second feature; The soft matching degree is determined based on the estimated employment information and the reference employment information.

5. A digital think tank method for comprehensive personnel capabilities, characterized in that: The method is executed by a processor, and the method includes: Determine candidate information based on candidate resume data; determine multiple historical recruitment information based on multiple historical resume data of the positions to be recruited; Based on the candidate information, determine a first feature of the candidate; based on the multiple historical recruitment information, determine a second feature corresponding to the multiple historical recruitment information; Determine a third feature based on the second feature, the job information of the job to be recruited, and historical job information; Based on the first feature and the third feature, determining the job matching degree between the candidate and the job to be recruited; Demonstrate the matching degree of the said position.

6. The method according to claim 5, characterized in that The determining of the third feature based on the second feature, the job information of the job to be recruited, and the historical job information includes: Based on the employment performance information of historically employed personnel corresponding to the second feature, the second feature is screened to obtain a target second feature; Based on the target second feature, the job information and the historical job information, the third feature is determined through a job requirement model, and the job requirement model is a machine learning model.

7. The method according to claim 5, characterized in that The determining of the job matching degree between the candidate and the job to be recruited based on the first feature and the third feature includes: Based on the candidate information, determine the candidate's historical employment information; Determining a hard matching degree of the candidate based on the first feature and the third feature; Determining the soft matching degree of the candidate based on the historical employment information and the position information; The position matching degree is determined based on the hard matching degree and the soft matching degree.

8. The method according to claim 7, characterized in that The determining of the soft matching degree of the candidate based on the historical employment information and the position information includes: Based on the historical employment information and the position information, predicting the estimated employment information of the candidate through an information prediction model, wherein the information prediction model is a machine learning model; Determining reference employment information based on employment performance information corresponding to historically employed personnel corresponding to the second feature; The soft matching degree is determined based on the estimated employment information and the reference employment information.

9. A digital think tank device for comprehensive personnel capabilities, characterized in that: The apparatus comprises at least one processor and at least one memory; The at least one memory is used to store computer instructions; The at least one processor is configured to execute at least part of the computer instructions to implement the method of any one of claims 5 to 8.

10. A computer-readable storage medium, characterized in that: The storage medium stores computer instructions. When a computer reads the computer instructions in the storage medium, the computer executes the method as claimed in any one of claims 5 to 8.

Images