A method, apparatus, equipment, and storage medium for detecting person-job matching

Abstract

This disclosure provides a method, apparatus, device, and storage medium for detecting person-job matching, relating to the field of computer technology, and particularly to the field of artificial intelligence technology. The specific implementation scheme is as follows: content analysis is performed on target description information to obtain ability representation information corresponding to a specified personnel type; target test questions are output; the answers of personnel of the specified personnel type to the target test questions are obtained; based on the answers, the ability representation information corresponding to the specified personnel type is updated to obtain target ability representation information; based on the target ability representation information and the requirement representation information, the matching result of the specified personnel type for the specified job is determined. Therefore, this scheme can improve the efficiency of person-job matching detection.

Description

Technical Field

[0001] This disclosure relates to the field of computer technology, and more particularly to the field of artificial intelligence technology, specifically to a method, apparatus, device, and storage medium for detecting human-job matching. Background Technology

[0002] Currently, companies typically need to conduct person-job matching tests to obtain matching results, and then determine whether a specific type of personnel meets the skill requirements of the job based on the matching results.

[0003] In related technologies, manual analysis is usually used to determine the capabilities of a specific personnel type and then to determine the matching result of that personnel type for the job based on those capabilities. Summary of the Invention

[0004] This disclosure provides a method, apparatus, equipment, and storage medium for detecting person-job matching.

[0005] According to one aspect of this disclosure, a method for detecting person-job matching is provided, the method comprising:

[0006] Content analysis is performed on the target description information to obtain the ability representation information corresponding to the specified personnel type; wherein, the target description information is a description of the skills mastered by the personnel of the specified personnel type, and the ability representation information is used to represent the degree of mastery of the personnel of the specified personnel type for each target skill;

[0007] Output target test questions; wherein, the target test questions are test questions from a question bank used for job matching analysis;

[0008] Obtain the answers of the specified personnel type to the target test questions;

[0009] Based on the response results, update the ability representation information corresponding to the specified personnel type to obtain the target ability representation information;

[0010] Based on the target capability representation information and demand representation information, the matching result of the specified personnel type for the specified position is determined; wherein, the demand representation information represents the degree of demand for each target skill in the specified position.

[0011] According to another aspect of this disclosure, a person-job matching detection device is provided, comprising:

[0012] The analysis module is used to perform content analysis on the target description information to obtain the ability representation information corresponding to the specified personnel type; wherein, the target description information is a description of the skills mastered by the personnel of the specified personnel type, and the ability representation information is used to represent the degree of mastery of the personnel of the specified personnel type for each target skill;

[0013] The output module is used to output target test questions; wherein, the target test questions are test questions from a question bank used for job matching analysis;

[0014] The acquisition module is used to acquire the answer results of the specified personnel type for the target test questions;

[0015] The update module is used to update the ability representation information corresponding to the specified personnel type based on the answer result, so as to obtain the target ability representation information;

[0016] The determination module is used to determine the matching result of the specified personnel type for the specified position based on the target capability representation information and the demand representation information; wherein, the demand representation information represents the degree of demand of the specified position for each target skill.

[0017] According to another aspect of this disclosure, an electronic device is provided, comprising:

[0018] At least one processor; and

[0019] A memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to enable the at least one processor to perform the steps of the above-described person-job matching detection method.

[0020] According to another aspect of this disclosure, a computer-readable storage medium is provided, which stores a computer program that, when executed by a processor, implements the steps of any of the above-described person-job matching detection methods.

[0021] According to another aspect of this disclosure, a computer program product containing instructions is provided that, when run on a computer, causes the computer to perform the steps of any of the person-job matching detection methods described in the above embodiments.

[0022] Beneficial effects of the embodiments disclosed herein:

[0023] This solution intelligently assesses the capabilities of designated personnel types based on target description information and outputs test questions to these personnel. The solution updates the capability representation information corresponding to each personnel type based on their responses to the test questions, thus obtaining target capability representation information. Then, it performs job-person matching detection based on the target capability representation information and the requirement representation information corresponding to each personnel type, obtaining the job-person matching result. Therefore, this solution can intelligently perform job-person matching detection without relying on manual detection, thereby improving the efficiency of job-person matching detection.

[0024] It should be understood that the description in this section is not intended to identify key or essential features of the embodiments of this disclosure, nor is it intended to limit the scope of this disclosure. Other features of this disclosure will become readily apparent from the following description. Attached Figure Description

[0025] The accompanying drawings are provided to better understand this solution and do not constitute a limitation of this disclosure. Wherein:

[0026] Figure 1 This is a flowchart illustrating the first method for detecting the matching of personnel to positions as disclosed in this publication.

[0027] Figure 2 This is a schematic diagram illustrating the principle of updating capability representation information corresponding to a specified personnel type, as provided in this disclosure.

[0028] Figure 3 This is a flowchart illustrating the second method for detecting job-person matching according to this disclosure;

[0029] Figure 4 This is a flowchart illustrating the third method for detecting the matching of personnel to positions as disclosed in this publication.

[0030] Figure 5 This is a flowchart illustrating the fourth method for detecting job-person matching in this disclosure;

[0031] Figure 6 This is a schematic diagram illustrating the principle of a topic selection strategy provided in this public disclosure;

[0032] Figure 7 This is a schematic diagram illustrating the principle of calculating the matching result of person-job matching provided in this disclosure;

[0033] Figure 8 This is a schematic diagram of a personnel-job matching detection device according to the present disclosure;

[0034] Figure 9 This is a block diagram of an electronic device used to implement the human-job matching detection method of the embodiments of this disclosure. Detailed Implementation

[0035] The exemplary embodiments of this disclosure are described below with reference to the accompanying drawings, including various details of the embodiments to aid understanding, and should be considered merely exemplary. Therefore, those skilled in the art will recognize that various changes and modifications can be made to the embodiments described herein without departing from the scope and spirit of this disclosure. Similarly, for clarity and brevity, descriptions of well-known functions and structures are omitted in the following description.

[0036] In related technologies, manual analysis is usually used to detect the capabilities of a specified personnel type and determine the matching result of the specified personnel type for the job based on the capabilities of the specified personnel type.

[0037] However, the following problems exist in the process of manual analysis:

[0038] (1) Manual analysis is inefficient and cannot enable staff to perform analysis on a large number of specified personnel types at the same time.

[0039] (2) The analysis is closely related to the historical experience of the staff who perform the analysis, and they have different analytical preferences. At the same time, the analysis of the specified personnel types is not comprehensive and detailed enough.

[0040] Based on the above, in order to improve the detection efficiency of person-job matching, this disclosure provides a method, apparatus, device, and storage medium for detecting person-job matching.

[0041] The following section first introduces a method for detecting person-job matching provided in the embodiments of this disclosure.

[0042] The person-job matching detection method provided in this disclosure can be applied to electronic devices. In specific applications, for example, the electronic device can be a terminal device or a server; the terminal device can be a smartphone, tablet computer, desktop computer, etc.; this disclosure does not limit the specific form of the electronic device.

[0043] Specifically, the executing entity of this job-person matching detection method can be a job-person matching detection device. For example, when the job-person matching detection method is applied to a terminal device, the job-person matching detection device can be a client running on the terminal device for implementing job-person matching detection. For example, when the job-person matching detection method is applied to a server, the job-person matching detection device can be a computer program running on the server, which can be used to perform job-person matching detection.

[0044] Furthermore, the person-job matching detection method provided in this disclosure can be used in any scenario where person-job matching detection is required. For example, when any company has a need for person-job matching detection during recruitment, it can use the method provided in this disclosure to perform person-job matching detection; or, when any person of a specified type has a need for person-job matching detection for a certain position, it can use the method provided in this disclosure to perform person-job matching detection, and so on.

[0045] In the technical solution disclosed herein, all operations involving the acquisition, storage, use, processing, transmission, provision, and disclosure of user personal information are carried out with the user's authorization.

[0046] The person-job matching detection method provided in this embodiment may include:

[0047] Content analysis is performed on the target description information to obtain the ability representation information corresponding to the specified personnel type; wherein, the target description information is a description of the skills mastered by the personnel of the specified personnel type, and the ability representation information is used to represent the degree of mastery of the personnel of the specified personnel type for each target skill;

[0048] Output target test questions; wherein, the target test questions are test questions from a question bank used for job matching analysis;

[0049] Obtain the answers of the specified personnel type to the target test questions;

[0050] Based on the response results, update the ability representation information corresponding to the specified personnel type to obtain the target ability representation information;

[0051] Based on the target capability representation information and demand representation information, the matching result of the specified personnel type for the specified position is determined; wherein, the demand representation information represents the degree of demand for each target skill in the specified position.

[0052] In the solution provided in this disclosure, when performing person-job matching detection, the content analysis of the target description information can be performed to obtain the ability representation information corresponding to the specified personnel type; target test questions can be output; the answer results of the personnel of the specified personnel type on the target test questions can be obtained; based on the answer results, the ability representation information corresponding to the specified personnel type can be updated to obtain target ability representation information; based on the target ability representation information and the requirement representation information, the matching result of the specified personnel type for the specified position can be determined, thus completing the person-job matching detection for the specified personnel type.

[0053] As can be seen, this solution intelligently assesses the capabilities of specified personnel types based on target description information, outputs test questions to these personnel, updates the capability representation information corresponding to each personnel type based on their responses to the test questions, thus obtaining target capability representation information. Then, based on the target capability representation information and the requirement representation information, a person-job matching test is performed to obtain the person-job matching result. Therefore, this solution can intelligently perform person-job matching testing without relying on manual testing, thereby improving the efficiency of person-job matching testing.

[0054] Furthermore, since it eliminates the need for manual checks on whether there is a person-job fit, and instead obtains precise target ability representation information for the specified personnel type for person-job fit analysis based on the target description information and the answers to the target test questions, this solution can avoid detection bias caused by subjectivity and improve the accuracy of person-job fit detection.

[0055] The following describes a method for detecting person-job matching provided by an embodiment of this disclosure, with reference to the accompanying drawings.

[0056] like Figure 1 As shown in the embodiments of this disclosure, a person-job matching detection method may include the following steps:

[0057] Step S101: Perform content analysis on the target description information to obtain the ability representation information corresponding to the specified personnel type; wherein, the target description information is a description of the skills mastered by the personnel of the specified personnel type, and the ability representation information is used to represent the degree of mastery of the personnel of the specified personnel type for each target skill;

[0058] In this embodiment of the disclosure, the designated personnel type can be any personnel type that has a need for person-job matching detection. The personnel of this any personnel type can be a group of personnel with similar characteristics. For example, the designated personnel type can be personnel who have participated in any training course. It is understood that personnel who have participated in such training courses usually possess similar or identical skills. Therefore, when it is necessary to perform job matching analysis on personnel who have participated in such training courses, these personnel can be identified as the same personnel type, and this same personnel type can be used as the designated personnel type. Of course, personnel with equivalent educational qualifications can also be identified as the same personnel type. If person-job matching analysis is required, the personnel type to which personnel with equivalent educational qualifications belong can be used as the designated personnel type.

[0059] In addition, the target description information can be the description information corresponding to any specified personnel type that has a need for person-job matching detection; and the form of the target description information can be text, video, etc. For example, the target description information can be a resume representing the skills mastered by personnel of the specified personnel type, and this disclosure does not limit it in this regard.

[0060] Since the target description information records content that reflects the skills and abilities of a specified personnel type, content analysis can be performed on the target description information during person-job matching detection to obtain the ability representation information corresponding to the specified personnel type. Understandably, the ability representation information obtained by analyzing the target description information can serve as the initial ability representation information for the specified personnel type. This information can then be updated by combining it with responses to test questions to obtain the target ability representation information, thereby ensuring the accuracy of the ability analysis for the specified personnel type.

[0061] Furthermore, different content analysis methods can be adopted for different forms of target description information. For example, if the target description information is in text form, text sentences can be extracted from the target description information, and then content analysis can be performed based on the text sentences to obtain the ability representation information of the specified personnel type. If the target description information is in video form, audio information can be extracted from the target description information, and then content recognition can be performed based on the audio information to obtain the ability representation information corresponding to the specified personnel type. This disclosure does not limit the specific content analysis method for target description information. Provided that the ability representation information corresponding to the specified personnel type can be obtained, any content analysis method can be applied to this disclosure. For clarity of solution and layout, the content analysis method for text-based target description information will be illustrated by another embodiment below.

[0062] Of course, to ensure skill universality—that is, regardless of the specific job position, the same skills can be used for analysis during person-job matching detection, thereby improving the convenience of the detection—each target skill can include not only the skills required for the specified job but also other skills. For example, for a company, if the specified job requires skills a, b, and c, and other jobs also require skills d and e, then each target skill can be three skills: a, b, and c. To ensure skill universality, it can also be five skills: a, b, c, d, and e. It is understood that in specific applications, the number of target skills involved in the target description information can be less than the number of pre-set target skills. For any target skill that is not present in the target description information, the mastery level of that target skill in the ability representation information can be set to represent the absence of the skill. Furthermore, the number and specific skill types of each target skill can be limited according to actual needs in specific applications; this embodiment does not impose such limitations.

[0063] It should be noted that, assuming it can characterize the degree of mastery of each target skill by a specified personnel type, the specific form of the ability representation information can be varied. Optionally, in one implementation, the ability representation information corresponding to a specified personnel type can be represented in the form of a vector, where each component of the vector corresponds to a target skill. Furthermore, the value of each component can characterize the degree of mastery of the target skill corresponding to that component by the personnel of that specified personnel type; a higher value indicates a higher degree of mastery. For example, the ability representation information corresponding to a specified personnel type is a three-dimensional row vector (0.6, 0.5, 0.8), where each component corresponds to a target skill, and the value of each component represents the score of the corresponding target skill. For example, the target skills corresponding to the components in the above three-dimensional row vector from left to right are target skill A, target skill B, and target skill C, respectively. Through this three-dimensional row vector, it can be seen that the score of target skill A is 0.6, the score of target skill B is 0.5, and the score of target skill C is 0.8. The score range of the skill is from 0 to 1. The higher the score of the skill, the higher the degree of mastery of the skill by the personnel of the specified personnel type. Of course, the form of ability representation information is not limited to this, and the representation of ability representation information is not specifically limited in this embodiment.

[0064] Step S102: Output target test questions; wherein, the target test questions are test questions from the question bank used for job matching analysis;

[0065] The question bank used for job matching analysis stores test questions related to the skills required for the job. Considering that the skill mastery levels recorded in the target description information are not accurate enough, to ensure the accuracy of the mastery levels of each target skill for the identified personnel type, not only can the target description information be analyzed, but target test questions can also be output to the personnel of the designated personnel type. Subsequently, based on the answers to the target test questions, the corresponding ability representation information for the designated personnel type is updated. The number of target test questions can be one or more.

[0066] It is understandable that the question bank mentioned above could be designed for various positions or for specific positions; theoretically, both are feasible.

[0067] Furthermore, there can be multiple ways to output target test questions to personnel of a specified personnel type.

[0068] Optionally, each test question in the question bank can be associated with the target skill it examines, reflecting the skill coverage of each test question. This allows for the output of target test questions that meet specified conditions to designated personnel types based on the number of target skills associated with each test question. These conditions could include the number of skills exceeding a predetermined threshold. Alternatively, each test question in the question bank can be associated with the target skill it considers and the question's difficulty level. This allows for the selection of test questions that simultaneously meet predetermined screening conditions based on both the number of target skills associated with each test question and its difficulty level, serving as target test questions for output to designated personnel types.

[0069] The above-described implementation of outputting target test questions to personnel of a specified personnel type is merely an example and should not be construed as limiting the embodiments of this disclosure. In specific applications, target test questions can also be personalized for personnel of a specified personnel type based on the ability representation information corresponding to that type, thereby making the ability analysis of personnel of that specified personnel type more targeted and comprehensive. For clarity and layout, the following describes, in conjunction with another embodiment, a specific implementation of personalized outputting target test questions for personnel of a specified personnel type based on the ability representation information corresponding to that type.

[0070] Furthermore, after outputting the target test questions to the specified personnel type, the answers of the specified personnel type to the target test questions can be obtained; the answers can include whether they are correct or incorrect. Moreover, the methods for generating the answers of the specified personnel type to the target test questions can include, but are not limited to: obtaining the answers of the specified personnel type to the target test questions, and performing intelligent comparison and analysis between the answers and the reference answers to obtain the answers of the specified personnel type to the target test questions.

[0071] Step S103: Obtain the answer results of the specified personnel type for the target test questions;

[0072] There are multiple ways to obtain the answers of the specified personnel type to the target test questions. For example, the answer can be submitted directly by the personnel of the specified personnel type, and the answer can be uploaded by the staff, who can then upload the answer to obtain the answer. This solution does not limit the method of obtaining the answer.

[0073] Step S104: Based on the response result, update the ability representation information corresponding to the specified personnel type to obtain the target ability representation information;

[0074] After personnel of a specified personnel type answer the target test questions, the ability representation information corresponding to the specified personnel type can be updated based on their answers to the target test questions, thereby obtaining the target ability representation information. This improves the matching degree between the ability representation information and the actual ability of the specified personnel type, that is, it improves the accuracy of the ability representation information.

[0075] It should be noted that there are multiple ways to update the ability representation information corresponding to the specified personnel type based on the answers of the specified personnel type to the target test questions.

[0076] Optionally, in one implementation, considering that the target test questions are associated with target skills, the specified representation information in the ability representation information corresponding to the target skill associated with the target test questions can be adjusted according to a predetermined adjustment rule based on the answer results of the target test questions, i.e., whether the answer is correct or incorrect. For example, the predetermined adjustment rule may include increasing the value representing the specified representation information by a predetermined range if the answer is correct, and decreasing the value representing the specified representation information by a predetermined range if the answer is incorrect. For example, taking the aforementioned three-dimensional row vector (0.6, 0.5, 0.8) as an example, if the target skills associated with the target test question are A and B, after outputting the target test question to a person of a specified personnel type, if the answer is correct, the first and second components of the three-dimensional row vector can be increased by 0.05 to obtain the updated three-dimensional row vector (0.65, 0.55, 0.8); if the answer is incorrect, the first and second components of the three-dimensional row vector can be decreased by 0.05 to obtain the updated three-dimensional row vector (0.55, 0.45, 0.8).

[0077] Optionally, in one implementation, updating the ability representation information corresponding to the specified personnel type based on the personnel's answers to the target test questions to obtain the target ability representation information includes steps A1-A3:

[0078] Step A1: Based on the specified description parameters of the target test question and the ability representation information corresponding to the specified personnel type, construct the interaction information corresponding to the target test question; wherein, the interaction information corresponding to the target test question is used to characterize the ability advantages of the specified personnel type in relation to the target test question;

[0079] Each test question in the question bank can be pre-associated with specified description parameters, thus allowing for easy access to the specified description parameters of the target test question.

[0080] Optionally, the specified description parameters may include one or more of a first description parameter, a second description parameter, and a third description parameter; wherein, the first description parameter is used to characterize the difficulty of the exercise; the second description parameter is used to characterize the discrimination of the exercise, which reflects the degree of difference between the exercise and other exercises; and the third description parameter is used to characterize the associated knowledge vector, which is a vector used to characterize whether each target skill is included. The various parameters included in the specified description information can describe the target test exercise from different dimensions, and the required dimension of the description parameter can be selected according to actual needs to suit different detection scenarios.

[0081] For example, the first and second description parameters can be obtained from the exercise numbers through a trainable embedding matrix. This trainable embedding matrix contains the first and second description parameters for all target test exercises. Given the exercise number of any target test exercise, the first and second description parameters for that exercise can be obtained from the trainable embedding matrix. The method for determining the first and second description parameters is not limited to the above implementation. Furthermore, the third description parameter can be a multi-dimensional vector, where each component represents a target skill. By setting different values ​​for each component, it can be shown whether the skill involved in the test exercise includes the target skill corresponding to that component. For example, a value of 1 indicates inclusion, and a value of 0 indicates non-inclusion.

[0082] Additionally, for example, when the specified description parameters include the first description parameter, the second description parameter, and the third description parameter, the method for constructing the interactive information corresponding to the target test question may include:

[0083] The difference between the ability representation information corresponding to the specified personnel type and the first description parameter of the target test exercise is calculated. The result of the difference is multiplied digit-wise with the third description parameter of the target test exercise. The result of the multiplication is then multiplied with the second description parameter of the target test exercise to obtain the interaction information corresponding to the target test exercise.

[0084] Based on the above method of determining the interactive information corresponding to the target test questions, it can be seen that...

[0085] Suppose a test question has a first descriptive parameter, a second descriptive parameter, and a third descriptive parameter, which are: question difficulty parameter, etc. Exercise Discrimination Parameter Knowledge vector Q associated with exercises j Given the calculated ability representation information θ corresponding to the specified personnel type, which is in vector form, an interactive formula can be used. To determine the ability advantages of a specific personnel type for the target test questions, the result determined by the interaction formula is the aforementioned interaction information.

[0086] For example, in one implementation, the interaction information corresponding to the target test question can be obtained simply by subtracting the ability representation information from the first descriptive parameter of the target test question. Based on this implementation, assuming the first descriptive parameter of a test question is the question difficulty parameter... The difficulty parameter of the exercise can be subtracted from the calculated ability representation information θ corresponding to the specified personnel type. We obtain vector y, which represents the interaction information described above, including the exercise difficulty parameter. Both the ability representation information θ and the problem difficulty parameter are vectors with the same dimension. Each component of the ability representation information θ corresponds to the same skill, wherein the aforementioned ability representation information θ and the exercise difficulty parameter The dimension is determined based on the number of skills in the target description information and the number of skills in the exercises. For generality, the exercise difficulty parameter... The ability representation information θ sets the non-existent skills to 0. Each component of the vector y represents the ability advantage of a person of a specified personnel type in the corresponding skill. The larger the value of each component, the greater the ability advantage of a person of a specified personnel type in the corresponding skill.

[0087] The interactive information corresponding to the target test questions constructed using the above method can comprehensively and accurately reflect the ability advantages of a specified personnel type regarding the target test questions. Of course, the above method of constructing the interactive information corresponding to the target test questions is merely an example and should not be construed as limiting the embodiments of this disclosure. Any method of constructing the interactive information corresponding to the target test questions can be applied to the embodiments of this disclosure.

[0088] Step A2: Input the interaction information corresponding to the target test question into a pre-trained second prediction model to obtain the predicted value of the answer result of the specified personnel type for the target test question; wherein, the second prediction model is a model trained with sample interaction information and the ground truth value corresponding to the sample interaction information, the sample interaction information is used to characterize the ability advantage of the sample personnel type for the sample question, and the ground truth value corresponding to the sample interaction information is the actual answer result of the sample personnel type for the sample question;

[0089] After obtaining the interaction information corresponding to the target test question, a pre-trained neural network model, specifically a pre-trained second prediction model, can be used to predict the answer results of the specified personnel type for the target test question. It should be noted that there can be multiple target test questions. Therefore, for each target test question, the second prediction model can be used to predict the answer result, obtaining the predicted answer result for the specified personnel type for each target test question. The actual answer result of the sample user for the sample question can be correct or incorrect, and the predicted answer result for the specified personnel type for the target test question can be the probability of a correct answer for the specified personnel type, i.e., the predicted probability of a correct answer.

[0090] The second prediction model can be called CDM (cognitive diagnosis model), and this disclosure does not limit the network structure and training method of the second prediction model.

[0091] Step A3: Based on the difference between the obtained predicted value and the answer results of the specified personnel type on the target test questions, update the ability representation information corresponding to the specified personnel type to obtain the target ability representation information.

[0092] Since the answers given by a person of a specified personnel type to the target test questions may differ from the predicted values ​​corresponding to the target test questions, and these differences can reflect the difference between the skill abilities recorded in the target description information and the actual skill abilities, the ability representation information corresponding to the specified personnel type can be updated by using the difference between the obtained predicted values ​​and the answers given by the person of the specified personnel type to the target test questions, thereby obtaining the target ability representation information.

[0093] For example, Figure 2 A schematic diagram illustrating the principle of updating the competency representation information corresponding to a specified personnel type using the CDM diagnostic module (i.e., the aforementioned CDM) when the specified descriptive parameters include a first descriptive parameter, a second descriptive parameter, and a third descriptive parameter is provided. Figure 2 As shown, it will be done through interactive formulas The obtained interactive information is input into CDM to obtain the probability of a person of a specified personnel type answering the question correctly. Then, the correct answer probability, i.e. the predicted value, is used to update the ability representation information corresponding to the specified personnel type to obtain the target ability representation information.

[0094] Optionally, considering that the model loss caused by the difference between the predicted value and the answer result of the target test question can reflect the ability differences of a specified personnel type to a certain extent, the model loss can be applied to the ability representation information update process to ensure the effectiveness of the update. Therefore, the step of updating the ability representation information corresponding to the specified personnel type based on the difference between the obtained predicted value and the answer result of the specified personnel type on the target test question, to obtain the target ability representation information, includes steps A31-A33:

[0095] Step A31: Determine the second loss value based on the difference between the obtained predicted value and the answer results of the specified personnel type for the target test questions;

[0096] The second loss value is determined based on the predicted value and the actual responses of the specified personnel type to the target test questions;

[0097] Step A32: Estimate the second gradient change value of the specified parameter when performing network parameter tuning on the second prediction model based on the second loss value; wherein, the specified parameter is a parameter related to the ability representation information corresponding to the specified personnel type;

[0098] In one implementation, a second gradient change value for the parameters representing the capability representation information therein can be obtained when the second prediction model backpropagates based on the second loss value.

[0099] Step A33: Based on the second gradient change value, adjust the capability representation information corresponding to the specified personnel type to obtain the target capability representation information.

[0100] For example, the second gradient value described above can be used as an adjustment value to adjust the parameters of the capability representation information in the second prediction model to obtain the target capability representation information.

[0101] For example, in one implementation, the second loss value can be determined using the following loss function:

[0102]

[0103] Loss cdm This is the loss value of the aforementioned CDM, s i The text sentence containing the target skills obtained from the initialization of data representing target description information; e j Recommended exercises j; r ij It is the predicted probability that person i of a specified personnel type will correctly answer question j, that is, the predicted value y for question j. ij It is the result of the actual answer to exercise j by personnel of the specified personnel type i, that is, the answer result of personnel of the specified personnel type for exercise j.

[0104] The second loss value is calculated using the loss function described above. The capability representation information θ can be adjusted based on the second gradient change value of the second loss value during backpropagation of the second prediction model to obtain the target capability representation information. The capability representation information θ is the specified parameter mentioned above.

[0105] The update methods described in steps A1-A3 above combine the skills reflected in the target description information with the characteristics of the target test questions, and utilize model prediction methods to ensure both the accuracy and efficiency of the update.

[0106] Step S105: Based on the target capability representation information and the demand representation information, determine the matching result of the specified personnel type for the specified position; wherein, the demand representation information represents the degree of demand for each target skill in the specified position;

[0107] The demand representation information can be expressed in vector form. Furthermore, the demand representation information can be obtained through content analysis of the recruitment requirements for a specific position. The method for analyzing the recruitment requirements for a specific position is similar to the method described above for analyzing the target description information of a specific personnel type. The only difference is that it involves target description information for a specific personnel type and recruitment requirements for a specific position. Therefore, the implementation method for content analysis of recruitment requirements will not be elaborated upon here.

[0108] After obtaining the target competency representation information and demand representation information corresponding to a specified personnel type, a matching analysis can be performed on the target competency representation information and demand representation information to obtain the matching result of the specified personnel type for the specified position. The specified position is the position to be tested for person-job matching; the matching result can be either a result indicating whether a match has occurred, or a matching probability indicating the degree of matching, both of which are reasonable.

[0109] It should be noted that there are multiple ways to determine the matching result of the specified personnel type for the specified job position based on the target capability representation information and the demand representation information.

[0110] Optionally, in one implementation, considering that both the target capability representation information and the demand representation information can be in vector form, the matching result of the specified personnel type for the specified job can be determined by calculating the vector distance.

[0111] Optionally, in another implementation, determining the matching result of the specified personnel type for the specified position based on the target capability representation information and the demand representation information includes:

[0112] The target capability representation information and demand representation information are input into a pre-trained job matching model to obtain the matching result of the specified personnel type for the specified job.

[0113] The job matching model is a model trained using the ability representation information corresponding to the sample personnel type, the demand representation information of the sample job, and the ground truth of the matching results of the sample personnel type with the sample job.

[0114] The job matching model is a model that applies an activation function. For example, It is an activation function in a job matching model, where N represents the number of target skills, and θ′ is the target ability representation information. It represents job requirement information, where s represents the job and i represents the target skill.

[0115] This approach, based on a job matching model, allows for more accurate matching results and ensures a higher analysis speed.

[0116] For example, the target competency representation information θ′(0.6,0,0.7,0,0.8,0.5) and the job requirement representation information are obtained by having personnel of a specified personnel type answer recommended exercises and updating them. The components of the two row vectors mentioned above represent the skill mastery scores and skill requirement levels for skills a, b, c, d, e, and f for a specified personnel type. Each component position corresponds to the same skill. For generality, the capability representation information θ′ and the job requirement representation information are also included. Set any non-existent skills to 0, and combine the aforementioned ability representation information θ′ with the job requirement representation information. Input to activation function This retrieves the matching results for the specified personnel type and the job position. The above interactive formula is understood to be correct. Loss function cdm and activation function This is just one implementation method in the embodiments of this disclosure. The embodiments of this disclosure do not specifically limit the interaction information, the method of determining the second loss value, or the form of the activation function.

[0117] Optionally, the method provided in this disclosure may further include:

[0118] Based on the target capability representation information, a capability profile of the specified personnel type is generated and output; wherein, the capability profile is used to represent the business capabilities of the specified personnel type for each target skill.

[0119] The aforementioned competency profile allows individuals of a specific skill type to understand their business capabilities for each target skill, i.e., their level of mastery of each target skill. Therefore, individuals of a specific skill type can use the competency profile to understand their weak skills so that they can subsequently address these weaknesses and improve their overall capabilities.

[0120] In the solution provided in this disclosure, when performing person-job matching detection, the content analysis of the target description information can be performed to obtain the ability representation information corresponding to the specified personnel type; target test questions can be output; the answer results of the personnel of the specified personnel type on the target test questions can be obtained; based on the answer results, the ability representation information corresponding to the specified personnel type can be updated to obtain target ability representation information; based on the target ability representation information and the requirement representation information, the matching result of the specified personnel type for the specified position can be determined, thus completing the person-job matching detection for the specified personnel type.

[0121] This solution intelligently assesses the capabilities of designated personnel types based on target description information and outputs test questions to these personnel. The solution updates the capability representation information corresponding to each personnel type based on their responses to the test questions, thus obtaining target capability representation information. Then, a person-job matching test is performed based on the target capability representation information and the requirement representation information to obtain the person-job matching result. Therefore, this solution can intelligently perform person-job matching detection without relying on manual detection, thereby improving the efficiency of person-job matching detection.

[0122] Furthermore, since there is no need for manual detection of whether the person-job fit is correct, but rather based on the target description information and the answers to the target test questions, the precise target ability representation information corresponding to the specified personnel type is obtained for the person-job fit analysis, thereby conducting job matching analysis. Therefore, this solution can avoid detection bias caused by subjectivity and improve the detection accuracy of person-job fit.

[0123] Alternatively, in another embodiment of this disclosure, such as Figure 3 As shown, a person-job matching detection method may include the following steps:

[0124] Step S301: Perform content analysis on the target description information to obtain the ability representation information corresponding to the specified personnel type; wherein, the target description information is a description of the skills mastered by the personnel of the specified personnel type, and the ability representation information is used to represent the degree of mastery of the personnel of the specified personnel type for each target skill;

[0125] Step S302: Based on the ability representation information corresponding to the specified personnel type, determine the information content corresponding to the test questions in the question bank; wherein, the information content corresponding to each test question is used to characterize the degree of influence of incorrect and / or correct answers on the ability representation information corresponding to the specified personnel type;

[0126] To personalize exercise recommendations to specific personnel types and improve the accuracy of ability analysis, after obtaining the ability representation information of personnel of a specific personnel type based on the target description information, the information content of the test exercises in the question bank can be determined based on the ability representation information corresponding to the specified personnel type. Then, based on the information content of the test exercises, personalized exercise recommendations can be made for personnel of the specified personnel type.

[0127] Provided that the amount of information corresponding to each test question is sufficient to reflect the impact of incorrect and / or correct answers on the ability representation information corresponding to the specified personnel type, the method for determining the amount of information corresponding to the test questions in the question bank is not limited.

[0128] To ensure clarity and a well-defined layout, subsequent examples, in conjunction with other embodiments, will exemplarily describe how to determine the amount of information corresponding to test questions in the question bank based on the ability representation information corresponding to the specified personnel type.

[0129] Step S303: Using the amount of information corresponding to the determined test questions, output the target test questions in the question bank;

[0130] In this embodiment, the amount of information corresponding to the test questions in the question bank can be used as the basis for recommending test questions, and target test questions can be output to personnel of a specified personnel type.

[0131] Optionally, in one implementation, the target test questions in the question bank are output based on the amount of information corresponding to the determined test questions, including steps B1-B3:

[0132] Step B1: Based on the amount of information corresponding to the determined test questions, determine the amount of information corresponding to each question set in the question bank; wherein, the question set is obtained by grouping the test questions in the question bank, and the amount of information corresponding to each question set is determined based on the amount of information corresponding to the test questions contained in the question set;

[0133] Step B2: Obtain the test questions from the set of questions with the largest amount of information to obtain the target test questions;

[0134] Step B3: Output the obtained target test questions.

[0135] Steps B1-B3 above can select a set of exercises with the largest amount of information based on the amount of information corresponding to the test exercises and output it to the personnel of the specified personnel type. The amount of information corresponding to the test exercises is calculated based on the ability representation information corresponding to the specified personnel type. The amount of information corresponding to the test exercises is used to represent the degree of influence of incorrect and / or correct answers on the ability representation information corresponding to the specified personnel type. Therefore, it is possible to recommend a set of exercises in a personalized and fast manner for personnel of the specified personnel type.

[0136] Alternatively, in another implementation, the importance and / or the amount of skills covered by the test questions can be used as the selection criteria for target test questions, thereby ensuring the diversity and popularity of the skills covered by the output target test questions, and further ensuring the accuracy of the ability analysis. Then, using the amount of information corresponding to the determined test questions, the target test questions in the question bank are output, including steps C1-C4.

[0137] Step C1: Based on the amount of information corresponding to the determined test questions, select multiple candidate questions from the question bank; wherein the amount of information corresponding to any candidate question is greater than the amount of information corresponding to any other test questions in the question bank excluding the candidate questions.

[0138] For example, when selecting multiple candidate exercises, the top K test exercises can be selected after sorting the test exercises according to their information content, or test exercises with information content greater than a predetermined threshold can be selected to obtain multiple candidate exercises.

[0139] Step C2: Obtain a specified evaluation value for each candidate exercise; wherein, the specified evaluation value for each candidate exercise is used to characterize the skill coverage of the candidate exercise and / or the importance of the skills covered by the candidate exercise;

[0140] The skill coverage of candidate exercises can be represented by the number of target skills associated with each candidate exercise. If a specified evaluation value is used to characterize skill coverage, a representation value representing the number of associated target skills can be generated as the specified evaluation value. Additionally, the importance of the skills covered by candidate exercises can be determined based on the word frequency of the skills in historical recruitment data; higher word frequency indicates higher importance. If a specified evaluation value is used to characterize the importance of the covered skills, a representation value representing the importance of the covered skills can be generated as the specified evaluation value. If the specified evaluation value is used to characterize two types of information, the representation values ​​of the two types of information can be weighted to obtain the specified evaluation value.

[0141] Step C3: Select the target test question from the plurality of candidate questions based on the magnitude of the specified evaluation value of each candidate question;

[0142] For example, when selecting target test questions, the top K candidate questions can be selected after sorting the candidate questions according to a specified evaluation value, or candidate questions with a specified evaluation value greater than a predetermined threshold can be selected to obtain the target test questions; this solution does not limit the selection method of target test questions.

[0143] Step C4: Output the selected target test questions.

[0144] After selecting the target test questions, the selected target test questions can be output to personnel of a specified personnel type.

[0145] Optionally, in one implementation, step C3 may include step C31:

[0146] Step C31: Select at least one target test question from the specified sequence that is ranked first among the multiple candidate questions;

[0147] The specified sequence is a sequence obtained by sorting the multiple candidate exercises in descending order of their specified evaluation values.

[0148] The above steps select multiple candidate exercises based on the amount of information corresponding to the test exercises, and then select a target number of target test exercises from the candidate exercises according to the specified evaluation value and output them to the specified personnel type. This allows for a personalized and rapid output of a set of exercises for the specified personnel type, and can more comprehensively assess the abilities of the specified personnel type.

[0149] Alternatively, in another implementation,

[0150] Step C5 may also be included after step C4 and before step S306.

[0151] Step C5: If it is detected that the total number of currently output target test questions has not reached the target number, then based on the specified ability representation information, the information content corresponding to the test questions in the question bank is re-determined, and the step of selecting multiple candidate questions from the question bank based on the determined information content of the test questions is returned, until the target number of target test questions is output.

[0152] The specified ability representation information is the target ability representation information updated based on the answers given by the specified personnel type to the most recently output target test questions.

[0153] Accordingly, updating the ability representation information corresponding to the specified personnel type based on the response result to obtain the target ability representation information may include:

[0154] Each time a target test question is output and the responses from personnel of a specified personnel type are obtained, the ability representation information corresponding to that personnel type is updated based on the obtained responses, thus obtaining the target ability representation information. Step C5 above involves updating the ability representation information corresponding to the specified personnel type based on the responses of personnel of the specified personnel type to obtain the target ability representation information, updating the information content of the test questions based on the target ability representation information, selecting candidate questions based on the information content, and selecting the target question based on the specified evaluation value of the candidate questions. This process is repeated until the number of answered questions reaches the target number. Therefore, this scheme can personalized recommend the target number of questions based on changes in the ability status of personnel of a specified personnel type, and the target questions selected by this scheme can more comprehensively assess the abilities of personnel of the specified personnel type.

[0155] The method of redetermining the amount of information corresponding to test questions in the question bank based on specified ability representation information is the same as the method of determining the amount of information corresponding to test questions in the question bank based on the ability representation information corresponding to specified personnel types, and will not be elaborated here.

[0156] Step S304: Obtain the answer results of the specified personnel type for the target test questions;

[0157] Step S305: Based on the response result, update the ability representation information corresponding to the specified personnel type to obtain the target ability representation information;

[0158] Step S306: Based on the target capability representation information and the demand representation information, determine the matching result of the specified personnel type for the specified position; wherein, the demand representation information represents the degree of demand for each target skill in the specified position.

[0159] The content of step S301 is the same as that of step S101 in the above embodiment, and the content of steps S304-S306 is the same as that of steps S103-S105 in the above embodiment, which will not be repeated here.

[0160] This solution enables intelligent person-job matching detection without the need for manual checks, thus improving detection efficiency. Furthermore, since it eliminates the need for manual checks on job matching, instead relying on target description information and responses to target test questions to obtain precise target competency representation information for specific personnel types for job matching analysis, this solution avoids detection biases caused by subjectivity, thereby improving the accuracy of person-job matching detection.

[0161] Furthermore, when outputting target recommended exercises, the information content of the test exercises in the question bank is first determined based on the ability representation information corresponding to the specified personnel type. Then, the target test exercises in the question bank are output using the determined information content of the test exercises. Therefore, target test exercises can be output to personnel of a specified personnel type in a personalized manner, so that the output target test exercises can better reflect the ability of personnel of the specified personnel type, thereby further improving the accuracy of person-job matching detection.

[0162] Optionally, in another embodiment of this disclosure, such as Figure 4 As shown, a person-job matching detection method may include the following steps:

[0163] Step S401: Obtain each specified text sentence from the target description information; wherein each specified text sentence is a text sentence that records at least one target skill;

[0164] In this embodiment, the target description information is in text form. Therefore, when performing capability analysis on personnel of a specified personnel type, we can first obtain each specified text sentence from the target description information, and then further analyze the specified text sentences.

[0165] Optionally, in one implementation, obtaining each specified text sentence from the target description information includes steps D1-D2:

[0166] Step D1: Extract each text sentence from the target description information;

[0167] Step D2: For each text sentence, if the text sentence contains a skill from the skill dictionary, then the text sentence is determined to be the specified text sentence.

[0168] The skill dictionary contains relevant skills for the job position to be tested. Matching using the skill dictionary ensures efficient selection of specific text sentences.

[0169] The above implementation is merely an example and should not be construed as limiting the embodiments of this disclosure.

[0170] Step S402: For each target skill recorded in each specified text sentence, determine the evaluation value corresponding to the target skill based on the text content of the specified text sentence that records the target skill; wherein, the evaluation value corresponding to the target skill is used to characterize the degree of mastery of the target skill by the personnel of the specified personnel type;

[0171] Optionally, in one implementation, considering that keywords representing the degree of mastery usually exist in the specified text sentence, such as "proficient mastery", the evaluation value corresponding to the target skill is determined based on the keywords representing the degree of mastery in the text content of the specified text sentence that records the target skill.

[0172] Optionally, in one implementation, determining the evaluation value corresponding to the target skill based on the text content of a specified text sentence containing the target skill includes steps E1-E2:

[0173] Step E1: Vectorize the text content of the specified text sentence that records the target skill to obtain the text vector to be used.

[0174] Each skill can be represented by one or more specified text sentences. When multiple text sentences contain the same skill, the average of the vectors of the text sentences containing the same skill is calculated to obtain the text vector to be used. Alternatively, the text sentences containing the same skill can be concatenated and vectorized to obtain the text vector to be used.

[0175] Step E2: Input the text vector to be used into the pre-trained first prediction model to obtain the evaluation value corresponding to the target skill;

[0176] The first prediction model is trained using the text vector corresponding to the sample sentence and the ground truth value corresponding to the sample sentence. The sample sentence is a text sentence that records the sample skills possessed by people of the sample personnel type. The ground truth value represents the degree of mastery of the sample skills by people of the sample personnel type.

[0177] Optionally, in one implementation, if the text vector to be utilized contains multiple skills, then the evaluation value of all skills in this text vector is the evaluation value obtained by inputting the text vector to be utilized into a pre-trained first prediction model. For example, if a text vector contains two skills, A and B, and the evaluation value obtained by inputting it into the pre-trained first prediction model is 0.6, then the evaluation value of both skills A and B is 0.6.

[0178] The model structure and training process of the first prediction model are not limited in the embodiments disclosed herein.

[0179] Using the first prediction model to analyze the evaluation values ​​can ensure accuracy and improve the analysis speed.

[0180] Step S403: Using the evaluation value corresponding to each target skill that has been determined, generate the ability representation information corresponding to the specified personnel type.

[0181] It is understandable that when generating competency representation information corresponding to a specified personnel type, the evaluation value corresponding to each target skill is determined as the representation information of that target skill in the competency representation information.

[0182] For example, in one implementation, the ability representation information corresponding to a specified personnel type can be represented in the form of a vector. Each component in the vector corresponds to a target skill, and the value of each component can represent the degree of mastery of the target skill corresponding to that component by the personnel of the specified personnel type. A higher value indicates a higher degree of mastery. The ability representation information corresponding to a personnel of a specified personnel type is a three-dimensional row vector (0.6, 0.5, 0.8), where each component corresponds to a target skill, and the value of each component represents the score of the corresponding target skill. For example, the target skills corresponding to the components in the above three-dimensional row vector from left to right are target skill A, target skill B, and target skill C, respectively. Through this three-dimensional row vector, it can be seen that the score of target skill A is 0.6, the score of target skill B is 0.5, and the score of target skill C is 0.8. The score range of the skill is 0 to 1. The higher the score of the skill, the higher the degree of mastery of the skill by the personnel of the specified personnel type. Of course, the form of ability representation information is not limited to this, and the representation of ability representation information in this embodiment does not specifically limit the form of ability representation information.

[0183] In one implementation, the number of target skills involved in the target description information may be less than the number of pre-set target skills. In this case, for target skills that do not exist, the corresponding evaluation value can be 0. That is, the number of target skills whose evaluation values ​​are determined through step S402 may be less than the required number of each target skill. In this case, when generating the ability representation information, the evaluation value of the target skills whose evaluation values ​​are not determined can be 0. For example, for position X, four target skills A, B, C, and D are pre-set. If the target skills involved in a certain target description information are A and B, then the evaluation values ​​of target skills C and D for the specified personnel type are 0. Finally, the ability representation information of the specified personnel type is (0.6, 0.4, 0, 0). The target skills corresponding to the components in the above four-dimensional row vector from left to right are target skill A, target skill B, target skill C, and target skill D, respectively. Through this four-dimensional row vector, it can be seen that the score of target skill A for this specified personnel type is 0.6, the score of target skill B is 0.4, and the scores of target skills C and D are 0.

[0184] Step S404: Output target test questions; wherein, the target test questions are test questions from the question bank used for job matching analysis.

[0185] Step S405: Obtain the answer results of the specified personnel type for the target test questions.

[0186] Step S406: Based on the response result, update the capability representation information corresponding to the specified personnel type to obtain the target capability representation information.

[0187] Step S407: Based on the target capability representation information and the demand representation information, determine the matching result of the specified personnel type for the specified position; wherein, the demand representation information represents the degree of demand for each target skill in the specified position.

[0188] Steps S404-S407 are the same as steps S102-S105 in the above embodiment, and will not be repeated here.

[0189] This solution enables intelligent person-job matching detection without the need for manual checks, thus improving detection efficiency. Furthermore, since it eliminates the need for manual checks on job matching, instead relying on target description information and responses to target test questions to obtain precise target competency representation information for specific personnel types for job matching analysis, this solution avoids detection biases caused by subjectivity, thereby improving the accuracy of person-job matching detection.

[0190] In addition, in this embodiment, for target description information in text form, the mastery level of the target skills recorded in the target description information is identified by analyzing specified text sentences, thereby generating ability representation information corresponding to the specified personnel type. In this way, the analysis speed and accuracy of the content analysis of the target description information can be guaranteed.

[0191] Optionally, in another embodiment of this disclosure, such as Figure 5 As shown, a person-job matching detection method may include the following steps:

[0192] Step S501: Perform content analysis on the target description information to obtain the ability representation information corresponding to the specified personnel type; wherein, the target description information is a description of the skills mastered by the personnel of the specified personnel type, and the ability representation information is used to represent the degree of mastery of the personnel of the specified personnel type for each target skill;

[0193] Step S502: For each test question in the question bank, using the specified description parameters of the test question and the ability representation information corresponding to the specified personnel type, construct the interaction information corresponding to the test question; wherein, the interaction information corresponding to each test question is used to characterize the ability advantages of the personnel of the specified personnel type in relation to the test question;

[0194] The step of constructing interactive information corresponding to the test question using the specified descriptive parameters of the test question and the ability representation information corresponding to the specified personnel type includes:

[0195] When the specified description parameters include the first description parameter, the second description parameter, and the third description parameter, the ability representation information corresponding to the specified personnel type is subtracted from the first description parameter, the result of the subtraction is multiplied digit-wise with the third description parameter, and the result of the multiplication is multiplied with the second description parameter to obtain the information content corresponding to the test question.

[0196] By constructing interactive information in the above manner, the ability advantages of a specified personnel type in relation to the target test questions can be comprehensively and accurately reflected. The specific content of the specified descriptive parameters is the same as in the above embodiments and will not be repeated here.

[0197] For example, there exists a problem difficulty parameter. Exercise Discrimination Parameter Knowledge vector Q associated with exercises j And a capability representation information θ, which is in vector form and can be expressed using an interactive formula. To determine the ability advantages of a specific personnel type for the target test questions, the first layer of interaction formula is the aforementioned interaction information.

[0198] In this embodiment, the method for constructing the interactive information corresponding to each test question in the question bank can refer to the method for constructing interactive information for the target test question in the above embodiment, and will not be repeated here.

[0199] Step S503: Input the interaction information corresponding to the test question into the pre-trained second prediction model to obtain the predicted value of the answer result of the specified personnel type for the test question; wherein, the second prediction model is a model trained with sample interaction information and the ground truth value corresponding to the sample interaction information, the sample interaction information is used to characterize the ability advantage of the sample personnel type for the sample question, and the ground truth value corresponding to the sample interaction information is the actual answer result of the sample personnel type for the sample question.

[0200] The detailed description of the second prediction model, and the detailed description of inputting the interaction information corresponding to the test question into the pre-trained second prediction model to obtain the predicted value of the answer result of the specified personnel type for the test question, can be found in the corresponding description of determining the predicted value of the answer result of the specified personnel type for the target test question in the above embodiments, and will not be repeated in this embodiment.

[0201] Step S504: Calculate the information content corresponding to the test question using the difference between the obtained predicted value and the specified true value; wherein, the specified true value is the answer result of the person of the specified person type regarding the test question, whether the answer is correct or incorrect.

[0202] For any test question, since the predicted value corresponding to the test question may differ from the specified true value corresponding to the test question, and the difference can reflect the difference between the skill ability recorded in the target description information and the actual skill ability, the information content corresponding to the test question can be calculated by using the difference between the obtained predicted value and the specified true value corresponding to the test question.

[0203] Optionally, considering that the model loss caused by the difference between the predicted value and the specified true value of any test question can reflect the ability differences of a specified personnel type to a certain extent, the model loss can be applied to the calculation process of the information content of the test question, thereby ensuring the validity of the calculated information content. The step of calculating the information content corresponding to the test question using the difference between the obtained predicted value and the specified true value includes steps F1-F3:

[0204] Step F1: Determine the first loss value based on the difference between the obtained predicted value and the specified true value;

[0205] Step F2: Estimate the first gradient change value of a specified parameter when tuning the network parameters of the second prediction model based on the first loss value; wherein the specified parameter is a parameter related to the ability representation information corresponding to the specified personnel type;

[0206] Step F3: Based on the first gradient change value, determine the information content corresponding to the test question.

[0207] For example, in one implementation, after obtaining the predicted and true values, the first loss value can be determined using the following loss function:

[0208]

[0209] Loss cdm This is the loss value of the aforementioned CDM, s i The text sentence containing the target skills obtained from the initialization of data representing target description information; e j Recommended exercises j; r ij It is the predicted probability that person i of a specified personnel type will correctly answer question j, that is, the predicted value y for question j. ij It is the result of the actual answer to exercise j by personnel of the specified personnel type i, that is, the answer result of personnel of the specified personnel type for exercise j.

[0210] The first loss value is calculated using the loss function described above. The first loss value includes two loss values: one for a person of the specified personnel type who answers the exercise correctly and the other for a person of the specified personnel type who answers the exercise incorrectly. The information content of the exercise can be determined based on the first gradient change of the ability representation information θ during backpropagation of the second prediction model using the two first loss values. The ability representation information θ is the specified parameter mentioned above.

[0211] Step S505: Using the amount of information corresponding to the determined test questions, output the target test questions in the question bank;

[0212] Step S506: Obtain the answer results of the specified personnel type for the target test questions;

[0213] Step S507: Based on the response result, update the ability representation information corresponding to the specified personnel type to obtain the target ability representation information;

[0214] Step S508: Based on the target capability representation information and the demand representation information, determine the matching result of the specified personnel type for the specified position; wherein, the demand representation information represents the degree of demand for each target skill in the specified position.

[0215] Step S501 is the same as step S301 in the above embodiment, and steps S505-S508 are the same as steps S303-S306 in the above embodiment, so they will not be repeated here.

[0216] This solution enables intelligent person-job matching detection without the need for manual checks, thus improving detection efficiency. Furthermore, since it eliminates the need for manual checks on job matching, instead relying on target description information and responses to target test questions to obtain precise target competency representation information for specific personnel types for job matching analysis, this solution avoids detection biases caused by subjectivity, thereby improving the accuracy of person-job matching detection.

[0217] Furthermore, this solution calculates interaction information based on the target description information and the specified description parameters of the exercises. It then calculates the information content of each exercise based on the interaction information, and selects questions for specific personnel types based on the information content. Based on the answers of personnel of specific personnel types, it obtains accurate ability representation information for job matching analysis corresponding to the specific personnel type, thereby conducting job matching analysis. Therefore, this solution can fully detect the abilities of personnel of specific personnel types, thereby avoiding detection bias caused by subjectivity and further improving the detection accuracy of job matching.

[0218] To better understand this solution, the following specific example illustrates the person-job matching detection method:

[0219] Step 1: Use a skill dictionary to match the text content of the target description information to obtain text sentences containing specific skills;

[0220] The text sentence containing a specific skill is the designated text sentence that records the skills in the skill dictionary.

[0221] The set of text sentences containing specific skills obtained from the analysis of target description information can be: in It is a sentence that contains a specific skill mentioned above.

[0222] Step 2: Initialize the text sentence representation using a pre-trained BERT (Bidirectional Encoder Representation from Transformers) model and model the text sentence using a BiLSTM (long short-term memory) model.

[0223] Step two is the vectorization process for the specified text sentence in the above embodiments.

[0224] Step 3: For each text sentence The final representation predicts the mastery of a given personnel type in a skill through a neural network layer, thereby obtaining the ability state θ of the given personnel type.

[0225] Each text sentence The final representation refers to each text sentence The vector; the neural network layer is the first prediction model in the above embodiment; the ability state θ of the specified personnel type is the ability representation information in the above embodiment.

[0226] Step 4: For each test question in the question bank, construct an interaction formula based on the item response theory in psychology. It also uses the strong fitting ability of neural networks to learn the interaction between a specified type of person and the exercise, and finally predicts the probability that a specified type of person will answer the exercise correctly.

[0227] The interaction formula is used to calculate the interaction information in the above embodiments; the neural network is the second prediction model in the above embodiments, wherein... This refers to the exercise difficulty parameter in the above embodiments. Q is the exercise discrimination parameter in the above embodiments. jθ is the exercise-related knowledge vector in the above embodiments, and θ is the ability representation information corresponding to the specified personnel type in the above embodiments.

[0228] Step 5: For each test question in the question bank, using the probability that a person of a specified personnel type will answer the question correctly, the information content corresponding to the question is determined by calculating the sum of the gradient changes of the ability state θ corresponding to the specified personnel type in both correct and incorrect cases.

[0229] This step is the step in the above embodiment to calculate the information content corresponding to the test exercise by using the difference between the obtained predicted value and the specified true value; the following is a loss function of a second prediction model CDM;

[0230]

[0231] Loss cdm This is the loss value of the aforementioned CDM, s i This refers to the text sentence containing the target skills obtained from the initialization of the target description information data; e j Recommended exercises j; r ij It is the predicted probability that person i of a specified personnel type will correctly answer question j, that is, the predicted value y for question j. ij It is the result of the actual answer to exercise j by personnel of the specified personnel type i, that is, the answer result of personnel of the specified personnel type for exercise j.

[0232] The second loss value is calculated using the loss function described above. The capability representation information θ can be adjusted based on the second gradient change value of the second loss value during backpropagation of the second prediction model. The capability representation information θ is the specified parameter mentioned above.

[0233] The first loss value is calculated using the loss function described above. Based on the first loss value, the gradient changes of correct and incorrect answers with respect to ability θ during backpropagation of the neural model can be calculated. The sum of gradient changes obtained from the correct and incorrect answers is then used to determine the information content corresponding to the test question. The larger the sum of gradient changes, the greater the information content corresponding to the test question.

[0234] Step 6: Select the top K1 questions with the most information from the question bank;

[0235] The top K1 exercises are the candidate exercises in the above embodiments;

[0236] Step 7: Select the top K2 exercises from the top K1 exercises according to the importance and diversity of the skills, and recommend them to the designated personnel type;

[0237] The importance of a skill is determined based on the word frequency of skills in historical job posting data; a higher word frequency indicates a higher skill importance. Skill diversity represents the coverage of skills in historical job posting data; the more skills covered in historical job posting data, the stronger the skill diversity. Candidate exercises can be scored by weighting skill importance and skill diversity, and the top K2 exercises with the highest scores are selected and recommended to designated personnel. The scores obtained using skill importance and skill diversity are the designated evaluation values ​​for candidate exercises in the above embodiment; the top K2 exercises are the target test exercises in the above embodiment.

[0238] Figure 6 This is a schematic diagram illustrating the principles of the topic selection strategies in steps six and seven above. For example... Figure 6 As shown, the candidate question bank is the question bank mentioned in the above embodiments. Information maximization means maximizing the amount of information mentioned above. Figure 6 θ 1 and θ 2 , respectively, represent the degree of influence of correct and incorrect answers on the ability representation information corresponding to the specified personnel type, using θ 1 and θ 2 It can represent the amount of information corresponding to the test questions.

[0239] Step 8: Update the ability status θ′ corresponding to the specified personnel type based on the answers given by the personnel of the specified personnel type;

[0240] The capability status θ′ corresponding to the specified personnel type is the updated target capability representation information in the above embodiment; in this step, it can be achieved through the following Loss cdm The loss value of the function is determined;

[0241]

[0242] The loss value is calculated using the loss function described above. The capability representation information θ can be adjusted based on the gradient change of the loss value with respect to the capability θ during backpropagation of the model, thus obtaining θ′.

[0243] Step 9: Based on the job posting text information, obtain the text sentences containing skills and use the BiLSTM model to model the text sentences;

[0244] Among them, the vector of the text sentence can be obtained by using the BiLSTM model to model the text sentence;

[0245] Step 10: Predict the vector of job requirements for specific skills.

[0246] Where 's' represents the target description information, and 'i' represents the target skill in the target description information. In this step, a skill demand prediction model can be used to predict the degree of demand for a specific skill in a given job position. This skill demand prediction model is trained using the text vector corresponding to a sample sentence and the ground truth value corresponding to that sample sentence. The sample sentence is a text sentence recording the sample skills possessed by the sample job position, and the ground truth value represents the degree of demand for that sample skill in the sample job position. Based on the degree of demand for a specific skill in a job position, a demand vector for that specific skill can be generated.

[0247] Step 11: Obtain the ability state θ′ of personnel of the specified personnel type after answering the recommended exercises, and the skill requirement vector for the specified position. Then, the formula is used. Predict the match score between a specified personnel type and a specified job position;

[0248] This step uses a job matching model to predict the match score between a specified personnel type and a specified job, wherein the formula is as follows: This refers to the activation function of the job matching model in the above embodiments; the loss function of the job matching model is:

[0249]

[0250] Loss pjf The loss value of the above job matching model is s. i This refers to a text sentence containing the target skill, obtained through data initialization using target description information; J j Indicates the position, f ij t is the predicted probability that a person of a specified personnel type i matches a job position j. ij The result of matching personnel of the specified personnel type i with job position j, i.e. whether the personnel of the specified personnel type have been hired.

[0251] Figure 7 This is a schematic diagram illustrating the principle of jointly training the first prediction model, the second prediction model, and the job matching model in steps three, eight, and eleven above to calculate the matching results of person-job matching. During joint training, the model's loss function is:

[0252] Loss = α * LosS cdm +β*Loss pjf

[0253] α and β are weights used to balance the two types of loss values; α and β are set based on experience; the loss function of the joint training model can replace the loss function of the second prediction model in the above steps. During the backpropagation of the joint training model, the gradient change of the loss value with respect to the ability representation information θ is calculated, thereby completing the calculation of the information content corresponding to the test questions and the update of the ability status based on the gradient change of the ability representation information θ.

[0254] In addition, such as Figure 7 As shown, the job skill requirement vector This refers to the demand characterization information described in the above embodiments. The elements in the capability representation information θ, obtained by content analysis of the target description information, are used to characterize the various dimensions of the capability representation information. Used to represent the elements in the demand representation information, L and M can have the same value for ease of calculation; Figure 7 Sum+Sigmoid is used to characterize the activation function mentioned above; by using CDM, the capability representation information θ can be updated, i.e., diagnosed, to obtain θ', which is the target capability representation information.

[0255] Based on the above embodiments of the person-job matching detection method, this disclosure also provides a person-job matching detection device. Figure 8 This is a schematic diagram of a person-job matching detection device provided in an embodiment of the present disclosure, as shown below. Figure 8 As shown, the person-job matching detection device may include:

[0256] Analysis module 801 is used to perform content analysis on target description information to obtain capability representation information corresponding to a specified personnel type; wherein, the target description information is a description of the skills mastered by personnel of the specified personnel type, and the capability representation information is used to represent the degree of mastery of personnel of the specified personnel type for each target skill;

[0257] Output module 802 is used to output target test questions; wherein, the target test questions are test questions from a question bank used for job matching analysis;

[0258] The acquisition module 803 is used to acquire the answer results of the specified personnel type for the target test questions;

[0259] The update module 804 is used to update the ability representation information corresponding to the specified personnel type according to the answer result, so as to obtain the target ability representation information;

[0260] The determining module 805 is used to determine the matching result of the specified personnel type for the specified position based on the target capability representation information and the demand representation information; wherein, the demand representation information represents the degree of demand for each target skill in the specified position.

[0261] In the solution provided in this disclosure, when performing person-job matching detection, the content analysis of the target description information can be performed to obtain the ability representation information corresponding to the specified personnel type; target test questions can be output; the answer results of the personnel of the specified personnel type on the target test questions can be obtained; based on the answer results, the ability representation information corresponding to the specified personnel type can be updated to obtain target ability representation information; based on the target ability representation information and the requirement representation information, the matching result of the specified personnel type for the specified position can be determined, thus completing the person-job matching detection for the specified personnel type.

[0262] As can be seen, this solution intelligently assesses the capabilities of personnel of a specified personnel type based on the target description information, outputs test questions to these personnel, updates the capability representation information corresponding to the specified personnel type based on their responses to the test questions, thus obtaining target capability representation information. Then, based on the target capability representation information and the requirement representation information, a person-job matching test is performed to obtain the person-job matching result. Therefore, this solution can intelligently perform person-job matching testing without relying on manual testing, thereby improving the efficiency of person-job matching testing.

[0263] Furthermore, since it eliminates the need for manual checks on whether there is a person-job fit, and instead obtains precise target ability representation information for the specified personnel type for person-job fit analysis based on the target description information and the answers to the target test questions, this solution can avoid detection bias caused by subjectivity and improve the accuracy of person-job fit detection.

[0264] Optionally, the output module includes:

[0265] The first determining submodule is used to determine the amount of information corresponding to the test questions in the question bank based on the ability representation information corresponding to the specified personnel type; wherein, the amount of information corresponding to each test question is used to characterize the degree of influence of incorrect and / or correct answers on the ability representation information corresponding to the specified personnel type;

[0266] The output submodule is used to output the target test questions in the question bank based on the amount of information corresponding to the determined test questions.

[0267] Optionally, the analysis module includes:

[0268] The acquisition submodule is used to acquire each specified text sentence from the target description information; wherein each specified text sentence is a text sentence that records at least one target skill;

[0269] The second determining submodule is used to determine the evaluation value corresponding to each target skill recorded in each specified text sentence based on the text content of the specified text sentence containing the target skill; wherein, the evaluation value corresponding to the target skill is used to characterize the degree of mastery of the target skill by the personnel of the specified personnel type;

[0270] The generation submodule is used to generate capability representation information corresponding to the specified personnel type using the evaluation value corresponding to each target skill.

[0271] Optionally, the second determining submodule includes:

[0272] The processing unit is used to vectorize the text content of a specified text sentence containing the target skill to obtain the text vector to be used.

[0273] The first input unit is used to input the text vector to be used into the pre-trained first prediction model to obtain the evaluation value corresponding to the target skill.

[0274] The first prediction model is trained using the text vector corresponding to the sample sentence and the ground truth value corresponding to the sample sentence. The sample sentence is a text sentence that records the sample skills possessed by people of the sample personnel type. The ground truth value represents the degree of mastery of the sample skills by people of the sample personnel type.

[0275] Optionally, the acquisition submodule includes:

[0276] The extraction unit is used to extract individual text sentences from the target description information;

[0277] The first determining unit is used to determine that for each text sentence, if the text sentence contains a skill from the skill dictionary, the text sentence is designated as the specified text sentence.

[0278] Optionally, the first determining submodule includes:

[0279] A construction unit is configured to construct interactive information corresponding to each test question in the question bank, using specified descriptive parameters of the test question and ability representation information of the specified personnel type; wherein, the interactive information corresponding to each test question is used to characterize the ability advantages of the specified personnel type regarding the test question;

[0280] The second input unit is used to input the interaction information corresponding to the test question into a pre-trained second prediction model to obtain the predicted value of the answer result of the specified personnel type for the test question; wherein, the second prediction model is a model trained with sample interaction information and the ground truth value corresponding to the sample interaction information, the sample interaction information is used to characterize the ability advantage of the sample personnel type for the sample question, and the ground truth value corresponding to the sample interaction information is the actual answer result of the sample personnel type for the sample question.

[0281] The calculation unit is used to calculate the information content corresponding to the test question by using the difference between the obtained predicted value and the specified true value; wherein the specified true value is the answer result of the person of the specified person type regarding the test question regarding correct and / or incorrect answers.

[0282] Optionally, the computing unit includes:

[0283] The first determining subunit is used to determine the first loss value based on the difference between the obtained predicted value and the specified true value;

[0284] The prediction subunit is used to predict the first gradient change value of a specified parameter when the second prediction model is tuned based on the first loss value; wherein the specified parameter is a parameter related to the ability representation information corresponding to the specified personnel type.

[0285] The second determining subunit is used to determine the amount of information corresponding to the test exercise based on the first gradient change value.

[0286] Optionally, the specified description parameters include one or more of the following: a first description parameter, a second description parameter, and a third description parameter;

[0287] The first description parameter is used to characterize the difficulty of the exercise;

[0288] The second description parameter is used to characterize the discrimination of the exercises;

[0289] The third description parameter is used to characterize the associated knowledge vector, which is a vector used to characterize whether each target skill is included.

[0290] Optionally, the building unit includes:

[0291] The calculation subunit is configured to, when the specified description parameters include the first description parameter, the second description parameter, and the third description parameter, subtract the ability representation information corresponding to the specified personnel type from the first description parameter, multiply the result of the subtraction with the third description parameter, and multiply the result of the multiplication with the second description parameter to obtain the interaction information corresponding to the test exercise.

[0292] Optionally, the output submodule includes:

[0293] The first selection unit is used to select multiple candidate exercises from the question bank based on the amount of information corresponding to the determined test exercises; wherein the amount of information corresponding to any candidate exercise is greater than the amount of information corresponding to other test exercises in the question bank excluding the candidate exercises.

[0294] The first acquisition unit is used to acquire a specified evaluation value for each candidate exercise; wherein, the specified evaluation value for each candidate exercise is used to characterize the skill coverage of the candidate exercise and / or the importance of the skills covered by the candidate exercise;

[0295] The second selection unit is used to select a target test question from the plurality of candidate questions based on the magnitude of a specified evaluation value for each candidate question.

[0296] The first output unit outputs the selected target test question.

[0297] Optionally, the second selection unit includes:

[0298] Selecting a sub-unit is used to select at least one target test question from a specified sequence that ranks first among the multiple candidate questions;

[0299] The specified sequence is a sequence obtained by sorting the multiple candidate exercises in descending order of their specified evaluation values.

[0300] Optionally, the output submodule includes:

[0301] The second output unit is used to, if it is detected that the total number of currently output target test questions has not reached the target number, redetermine the information content corresponding to the test questions in the question bank based on the specified ability representation information, and return the step of selecting multiple candidate questions from the question bank based on the determined information content of the test questions, until the target number of target test questions is output;

[0302] The specified ability representation information is the target ability representation information updated based on the answers given by personnel of a specified personnel type to the target test questions in the most recent output.

[0303] Optionally, the update module includes:

[0304] The update submodule is used to update the ability representation information corresponding to the specified personnel type based on the obtained answer results of the target test questions each time they are output, in order to obtain the target ability representation information.

[0305] Optionally, the output submodule includes:

[0306] The second determining unit is used to determine the information content corresponding to each set of exercises in the question bank based on the amount of information corresponding to the determined test exercises; wherein, the set of exercises is obtained by grouping and dividing the test exercises in the question bank, and the information content corresponding to each set of exercises is determined based on the information content corresponding to the test exercises contained in the set of exercises;

[0307] The second acquisition unit is used to acquire test questions from the set of questions with the largest amount of information to obtain the target test questions;

[0308] The third output unit is used to output the obtained target test questions to the personnel of the specified personnel type.

[0309] Optionally, the update module includes:

[0310] A construction submodule is used to construct interactive information corresponding to the target test question based on the specified description parameters of the target test question and the ability representation information corresponding to the specified personnel type; wherein, the interactive information corresponding to the target test question is used to characterize the ability advantages of the specified personnel type in relation to the target test question;

[0311] The first input submodule is used to input the interaction information corresponding to the target test question into a pre-trained second prediction model to obtain the predicted value of the answer result of the specified personnel type for the target test question; wherein, the second prediction model is a model trained with sample interaction information and the ground truth value corresponding to the sample interaction information, the sample interaction information is used to characterize the ability advantage of the sample personnel type for the sample question, and the ground truth value corresponding to the sample interaction information is the actual answer result of the sample personnel type for the sample question;

[0312] The second update submodule is used to update the ability representation information corresponding to the specified personnel type based on the difference between the obtained predicted value and the answer results of the specified personnel type on the target test questions, so as to obtain the target ability representation information.

[0313] Optionally, the second update submodule includes:

[0314] The third determining unit is used to determine a second loss value based on the difference between the obtained predicted value and the answer results of the specified personnel type for the target test questions;

[0315] The prediction unit is used to predict the second gradient change value of a specified parameter when the network parameter tuning of the second prediction model is performed based on the second loss value; wherein the specified parameter is a parameter relating to the ability representation information of the specified personnel type.

[0316] The adjustment unit is used to adjust the capability representation information corresponding to the specified personnel type based on the second gradient change value to obtain the target capability representation information.

[0317] Optionally, the determining module includes:

[0318] The second input submodule is used to input the target capability representation information and demand representation information into a pre-trained job matching model to obtain the matching result of the specified personnel type for the specified job.

[0319] The job matching model is a model trained using the ability representation information of sample users, the demand representation information of sample jobs, and the ground truth of the matching results of sample personnel types with the sample jobs.

[0320] Optionally, the job matching model is a model with an activation function.

[0321] Optionally, a capability profile of the specified personnel type is generated and output based on the target capability representation information; wherein the capability profile is used to represent the business capabilities of the specified personnel type for each target skill.

[0322] According to embodiments of this disclosure, this disclosure also provides an electronic device, a readable storage medium, and a computer program product.

[0323] An electronic device provided in this disclosure may include:

[0324] At least one processor; and

[0325] A memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to enable the at least one processor to perform the steps of the above-described person-job matching detection method.

[0326] This disclosure provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps of any of the above-described person-job matching detection methods.

[0327] In yet another embodiment provided in this disclosure, a computer program product containing instructions is also provided, which, when run on a computer, causes the computer to perform the steps of any of the person-job matching detection methods described above.

[0328] Figure 9 A schematic block diagram of an example electronic device 900 that can be used to implement embodiments of the present disclosure is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device may also represent various forms of mobile devices, such as personal digital processors, cellular phones, smartphones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely illustrative and are not intended to limit the implementation of the present disclosure described and / or claimed herein.

[0329] like Figure 9 As shown, device 900 includes a computing unit 901, which can perform various appropriate actions and processes based on a computer program stored in read-only memory (ROM) 902 or a computer program loaded from storage unit 908 into random access memory (RAM) 903. RAM 903 may also store various programs and data required for the operation of device 900. The computing unit 901, ROM 902, and RAM 903 are interconnected via bus 904. Input / output (I / O) interface 905 is also connected to bus 904.

[0330] Multiple components in device 900 are connected to I / O interface 905, including: input unit 906, such as keyboard, mouse, etc.; output unit 907, such as various types of monitors, speakers, etc.; storage unit 908, such as disk, optical disk, etc.; and communication unit 909, such as network card, modem, wireless transceiver, etc. Communication unit 909 allows device 900 to exchange information / data with other devices through computer networks such as the Internet and / or various telecommunications networks.

[0331] The computing unit 901 can be a variety of general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of the computing unit 901 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special-purpose artificial intelligence (AI) computing chips, various computing units running machine learning model algorithms, a digital signal processor (DSP), and any suitable processor, controller, microcontroller, etc. The computing unit 901 performs the various methods and processes described above, such as the job matching detection method. For example, in some embodiments, the job matching detection method can be implemented as a computer software program tangibly contained in a machine-readable medium, such as storage unit 908. In some embodiments, part or all of the computer program can be loaded and / or installed on device 900 via ROM 902 and / or communication unit 909. When the computer program is loaded into RAM 903 and executed by the computing unit 901, one or more steps of the job matching detection method described above can be performed. Alternatively, in other embodiments, the computing unit 901 can be configured to perform the job matching detection method by any other suitable means (e.g., by means of firmware).

[0332] Various embodiments of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field-programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), application-specific standard products (ASSPs), systems-on-a-chip (SoCs), complex programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments may include implementations in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which may be a dedicated or general-purpose programmable processor, capable of receiving data and instructions from a storage system, at least one input device, and at least one output device, and transmitting data and instructions to the storage system, the at least one input device, and the at least one output device.

[0333] The program code used to implement the methods of this disclosure may be written in any combination of one or more programming languages. This program code may be provided to a processor or controller of a general-purpose computer, special-purpose computer, or other programmable data processing apparatus, such that when executed by the processor or controller, the program code causes the functions / operations specified in the flowcharts and / or block diagrams to be implemented. The program code may be executed entirely on a machine, partially on a machine, as a standalone software package partially on a machine and partially on a remote machine, or entirely on a remote machine or server.

[0334] In the context of this disclosure, a machine-readable medium can be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, apparatus, or device. A machine-readable medium can be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium can be, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination of the foregoing. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the foregoing.

[0335] To provide interaction with a user, the systems and techniques described herein can be implemented on a computer having: a display device for displaying information to the user (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor); and a keyboard and pointing device (e.g., a mouse or trackball) through which the user provides input to the computer. Other types of devices can also be used to provide interaction with the user; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including sound input, voice input, or tactile input).

[0336] The systems and technologies described herein can be implemented in computing systems that include backend components (e.g., as a data server), or computing systems that include middleware components (e.g., an application server), or computing systems that include frontend components (e.g., a user computer with a graphical user interface or web browser through which a user can interact with embodiments of the systems and technologies described herein), or any combination of such backend, middleware, or frontend components. The components of the system can be interconnected via digital data communication of any form or medium (e.g., a communication network). Examples of communication networks include local area networks (LANs), wide area networks (WANs), and the Internet.

[0337] Computer systems can include clients and servers. Clients and servers are generally located far apart and typically interact via communication networks. Client-server relationships are created by computer programs running on the respective computers and having a client-server relationship with each other. Servers can be cloud servers, servers in distributed systems, or servers incorporating blockchain technology.

[0338] It should be understood that the various forms of processes shown above can be used to rearrange, add, or delete steps. For example, the steps described in this disclosure can be executed in parallel, sequentially, or in different orders, as long as the desired result of the technical solution disclosed in this disclosure can be achieved, and this is not limited herein.

[0339] The specific embodiments described above do not constitute a limitation on the scope of protection of this disclosure. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this disclosure should be included within the scope of protection of this disclosure.

Claims

1. A method for detecting person-job fit, the method comprising: Content analysis is performed on the target description information to obtain the ability representation information corresponding to the specified personnel type; wherein, the target description information is a description of the skills mastered by the personnel of the specified personnel type, and the ability representation information is used to represent the degree of mastery of the personnel of the specified personnel type for each target skill, and is represented in vector form; Output target test questions; wherein, the target test questions are test questions from a question bank used for job matching analysis; Obtain the answers of the specified personnel type to the target test questions; Based on the specified descriptive parameters of the target test question and the ability representation information corresponding to the specified personnel type, interactive information corresponding to the target test question is constructed; wherein, the interactive information corresponding to the target test question is used to characterize the ability advantages of the specified personnel type in relation to the target test question; The interaction information corresponding to the target test question is input into a pre-trained second prediction model to obtain the predicted value of the answer result of the specified personnel type for the target test question; the second prediction model is a model trained with sample interaction information and the ground truth value corresponding to the sample interaction information, wherein the sample interaction information is used to characterize the ability advantage of the sample personnel type for the sample question, and the ground truth value corresponding to the sample interaction information is the actual answer result of the sample personnel type for the sample question. A second loss value is determined based on the difference between the obtained predicted value and the answers given by the specified personnel type to the target test questions; The second gradient change value of a specified parameter is estimated when the network parameters of the second prediction model are tuned based on the second loss value; wherein the specified parameter is a parameter relating to the ability representation information corresponding to the specified personnel type; Based on the second gradient change value, the ability representation information corresponding to the specified personnel type is adjusted to obtain the target ability representation information; Based on the target capability representation information and demand representation information, the matching result of the specified personnel type for the specified position is determined; wherein, the demand representation information represents the degree of demand for each target skill in the specified position.

2. The method according to claim 1, wherein, The output target test questions include: Based on the ability representation information corresponding to the specified personnel type, the information content corresponding to the test questions in the question bank is determined; wherein, the information content corresponding to each test question is used to characterize the degree of influence of incorrect and / or correct answers on the ability representation information corresponding to the specified personnel type; Based on the amount of information corresponding to the determined test questions, the target test questions in the question bank are output.

3. The method according to claim 1 or 2, wherein, The content analysis of the target description information to obtain the capability representation information corresponding to the specified personnel type includes: Extract each specified text sentence from the target description information; wherein each specified text sentence is a text sentence that records at least one target skill; For each target skill recorded in each specified text sentence, an evaluation value corresponding to the target skill is determined based on the text content of the specified text sentence containing the target skill; wherein, the evaluation value corresponding to the target skill is used to characterize the degree of mastery of the target skill by the personnel of the specified personnel type; Using the evaluation value corresponding to each target skill, the ability representation information corresponding to the specified personnel type is generated.

4. The method according to claim 3, wherein, The process of determining the evaluation value corresponding to the target skill based on the text content of a specified text sentence containing the target skill includes: The text content of the specified text sentence containing the target skill is vectorized to obtain the text vector to be used; The text vector to be used is input into a pre-trained first prediction model to obtain the evaluation value corresponding to the target skill; The first prediction model is trained using the text vector corresponding to the sample sentence and the ground truth value corresponding to the sample sentence. The sample sentence is a text sentence that records the sample skills possessed by people of the sample personnel type. The ground truth value represents the degree of mastery of the sample skills by people of the sample personnel type.

5. The method according to claim 3, wherein, The step of obtaining each specified text sentence from the target description information includes: Extract each text sentence from the target description information; For each text sentence, if the text sentence contains a skill from the skill dictionary, then the text sentence is determined to be the specified text sentence.

6. The method according to claim 2, wherein, The step of determining the amount of information corresponding to the test questions in the question bank based on the ability representation information corresponding to the specified personnel type includes: For each test question in the question bank, interactive information corresponding to the test question is constructed using the specified description parameters of the test question and the ability representation information corresponding to the specified personnel type; wherein, the interactive information corresponding to each test question is used to characterize the ability advantages of the personnel of the specified personnel type in relation to the test question; The interaction information corresponding to the test question is input into a pre-trained second prediction model to obtain the predicted value of the answer result of the specified personnel type for the test question; wherein, the second prediction model is a model trained with sample interaction information and the ground truth value corresponding to the sample interaction information, the sample interaction information is used to characterize the ability advantage of the sample personnel type for the sample question, and the ground truth value corresponding to the sample interaction information is the actual answer result of the sample personnel type for the sample question. The information content corresponding to the test question is calculated by using the difference between the obtained predicted value and the specified true value; wherein, the specified true value is the answer result of the person of the specified person type regarding the test question regarding correct and / or incorrect answers.

7. The method according to claim 6, wherein, The step of calculating the information content corresponding to the test question by using the difference between the obtained predicted value and the specified true value includes: The first loss value is determined based on the difference between the obtained predicted value and the specified true value; The first gradient change value of a specified parameter is estimated when the second prediction model is tuned based on the first loss value; wherein the specified parameter is a parameter relating to the ability representation information corresponding to the specified personnel type; Based on the first gradient change value, the information content corresponding to the test question is determined.

8. The method according to claim 6 or 7, wherein, The specified description parameters include one or more of the following: a first description parameter, a second description parameter, and a third description parameter; The first description parameter is used to characterize the difficulty of the exercise; The second description parameter is used to characterize the discrimination of the exercises; The third description parameter is used to characterize the associated knowledge vector, which is a vector used to characterize whether each target skill is included.

9. The method according to claim 8, wherein, The step of constructing interactive information corresponding to the test question using the specified descriptive parameters of the test question and the ability representation information corresponding to the specified personnel type includes: When the specified description parameters include the first description parameter, the second description parameter, and the third description parameter, the ability representation information corresponding to the specified personnel type is subtracted from the first description parameter, the result of the subtraction is multiplied digit-wise with the third description parameter, and the result of the multiplication is multiplied with the second description parameter to obtain the interaction information corresponding to the test question.

10. The method according to claim 2, wherein, The step of outputting the target test questions in the question bank based on the amount of information corresponding to the determined test questions includes: Based on the amount of information corresponding to the determined test questions, multiple candidate questions are selected from the question bank; wherein the amount of information corresponding to any candidate question is greater than the amount of information corresponding to any other test question in the question bank excluding the candidate questions. Obtain a specified evaluation value for each candidate exercise; wherein, the specified evaluation value for each candidate exercise is used to characterize the skill coverage of the candidate exercise and / or the importance of the skills covered by the candidate exercise; Based on the magnitude of the specified evaluation value of each candidate exercise, a target test exercise is selected from the plurality of candidate exercises; Output the selected target test questions.

11. The method according to claim 10, wherein, The selection of the target test problem from the plurality of candidate problems based on the magnitude of a specified evaluation value for each candidate problem includes: From the plurality of candidate exercises, select at least one target test exercise that is ranked first in a specified sequence; The specified sequence is a sequence obtained by sorting the multiple candidate exercises in descending order of their specified evaluation values.

12. The method according to claim 10, wherein, Before determining the matching result of the specified personnel type for the specified position based on the target capability representation information and the demand representation information, and after outputting the selected target test questions, the method further includes: If it is detected that the total number of currently output target test questions has not reached the target number, then based on the specified ability representation information, the information content corresponding to the test questions in the question bank is re-determined, and the step of selecting multiple candidate questions from the question bank based on the determined information content of the test questions is returned, until the target number of target test questions is output; The specified ability representation information is the target ability representation information updated based on the answers of personnel of a specified personnel type to the target test questions in the most recent output. The step of updating the ability representation information corresponding to the specified personnel type based on the response result to obtain the target ability representation information includes: Each time a target test question is output and the answers of a specified personnel type to the target test question are obtained, the ability representation information corresponding to the specified personnel type is updated based on the obtained answers to obtain the target ability representation information.

13. The method according to claim 2, wherein, The step of outputting the target test questions in the question bank based on the amount of information corresponding to the determined test questions includes: Based on the amount of information corresponding to the determined test questions, the amount of information corresponding to each set of questions in the question bank is determined; wherein, the set of questions is obtained by grouping the test questions in the question bank, and the amount of information corresponding to each set of questions is determined based on the amount of information corresponding to the test questions contained in that set of questions; Obtain the test questions from the set of questions with the largest amount of information to obtain the target test questions; Output the obtained target test questions.

14. The method according to claim 1 or 2, wherein, The step of determining the matching result of the specified personnel type for the specified position based on the target capability representation information and the demand representation information includes: The target capability representation information and demand representation information are input into a pre-trained job matching model to obtain the matching result of the specified personnel type for the specified job. The job matching model is a model trained using the ability representation information corresponding to the sample personnel type, the demand representation information of the sample job, and the ground truth of the matching results of the sample personnel type with the sample job.

15. The method according to claim 14, wherein, The job matching model is a model that uses an activation function.

16. The method according to claim 1 or 2, further comprising: Based on the target capability representation information, a capability profile of the specified personnel type is generated and output; wherein, the capability profile is used to represent the business capabilities of the specified personnel type for each target skill.

17. A person-job matching detection device, comprising: The analysis module is used to perform content analysis on the target description information to obtain the ability representation information corresponding to the specified personnel type; wherein, the target description information is a description of the skills mastered by the personnel of the specified personnel type, and the ability representation information is used to represent the degree of mastery of the personnel of the specified personnel type for each target skill, and is represented in vector form; The output module is used to output target test questions; wherein, the target test questions are test questions from a question bank used for job matching analysis; The acquisition module is used to acquire the answer results of the specified personnel type for the target test questions; The update module is used to construct interaction information corresponding to the target test question based on specified description parameters of the target test question and ability representation information corresponding to the specified personnel type. The interaction information corresponding to the target test question is used to characterize the ability advantages of the specified personnel type regarding the target test question. The interaction information corresponding to the target test question is input into a pre-trained second prediction model to obtain a predicted value of the answer result of the specified personnel type for the target test question. A second loss value is determined based on the difference between the obtained predicted value and the answer result of the specified personnel type for the target test question. The module also estimates the... When tuning the network parameters of the second prediction model based on the second loss value, the second gradient change value of the specified parameter is used, wherein the specified parameter is a parameter related to the ability representation information corresponding to the specified personnel type; based on the second gradient change value, the ability representation information corresponding to the specified personnel type is adjusted to obtain the target ability representation information; the second prediction model is a model trained with sample interaction information and the ground truth value corresponding to the sample interaction information, wherein the sample interaction information is used to characterize the ability advantages of personnel of the sample personnel type for sample exercises, and the ground truth value corresponding to the sample interaction information is the actual answer result of personnel of the sample personnel type for sample exercises. The determination module is used to determine the matching result of the specified personnel type for the specified position based on the target capability representation information and the demand representation information; wherein, the demand representation information represents the degree of demand of the specified position for each target skill.

18. An electronic device, comprising: At least one processor; as well as A memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor to enable the at least one processor to perform the method of any one of claims 1-16.

19. A non-transitory computer-readable storage medium storing computer instructions, wherein, The computer instructions are used to cause the computer to perform the method according to any one of claims 1-16.

20. A computer program product comprising a computer program that, when executed by a processor, implements the method according to any one of claims 1-16.

Images