Post competency matching method and device, equipment and storage medium

Abstract

The invention relates to a post competency matching method and device, equipment and a storage medium. The method comprises the following steps: acquiring original data, wherein the original data comprises talent features and post features; calculating a distance vector between the talent feature and the post feature; calculating mutual information and conditional entropy according to talent features, post features and distance vectors; optimizing the distance vector based on the mutual information amount and the conditional entropy; constructing a post competency model based on the optimized distance vector; the post competency model is used for matching talents and posts. According to the method, the AI algorithm big database and the post competency model are combined, an intelligent conditional entropy clustering calculation method is utilized, accurate classification matching of post competency is achieved, and a new and more efficient solution is provided for the field of human resource management.

Description

Technical Field

[0001] The present application relates to the field of artificial intelligence technology, and in particular to a job competency matching method, apparatus, device, and storage medium. Background Art

[0002] With the rapid development of information technology and the continuous advancement of human resource management, the matching methods for talent and positions are also evolving. Traditional human resource management methods rely primarily on manual screening and interviews. This method is not only time-consuming and labor-intensive, but also susceptible to subjective factors, resulting in poor matching between talent and positions.

[0003] With the rise of technologies like big data and artificial intelligence, human resources service platforms have begun to leverage technology to achieve preliminary matches between talent and positions. These platforms build preliminary matching models by collecting and analyzing talent data, such as educational background, work experience, and skills certifications, as well as job information, such as job responsibilities and skill requirements. However, these models often lack depth and professionalism, and are unable to accurately measure the degree of fit between talent characteristics and job characteristics.

[0004] To further improve the accuracy and efficiency of matching, professional human resources talent identification tools and methods are beginning to be introduced into the matching process. These methods combine knowledge from multiple disciplines, such as psychology and organizational behavior, to construct more sophisticated competency models. However, current human resources service platforms rely primarily on simple technical means to match talent and positions and recruit, lacking solutions for building competency models that integrate with professional human resources talent identification tools and methods. As a result, in practical applications, the classification and matching of job competencies is often inaccurate, making it difficult to meet the needs of both businesses and individuals. Specifically, these methods fail to fully utilize job competency models and job attributes, and are unable to effectively measure the degree of match between talent characteristics and job characteristics, thereby affecting the efficiency and accuracy of talent selection and job assignment. Summary of the Invention

[0005] In order to solve or partially solve the problems existing in the relevant technologies, the present application provides a job competency matching method, device, equipment and storage medium, which can achieve accurate classification and matching of job competencies by combining AI algorithm big database and job competency model, and using intelligent conditional entropy clustering calculation method, providing a new and more efficient solution for the field of human resources management.

[0006] The first aspect of the present application provides a job competency matching method, comprising: Obtaining original data, wherein the original data includes talent characteristics and job characteristics; Calculating the distance vector between the talent characteristics and the job characteristics; Calculating mutual information and conditional entropy based on the talent characteristics, the job characteristics, and the distance vector; Optimizing the distance vector based on the mutual information and the conditional entropy; A job competency model is constructed based on the optimized distance vector; the job competency model is used to match talents and jobs.

[0007] In some implementations, calculating the distance vector between the talent characteristics and the position characteristics includes: Calculating a first distance vector from the talent feature to the position feature, and a second distance vector from the position feature to the talent feature; The calculating of mutual information and conditional entropy based on the talent characteristics, the job characteristics, and the distance vector includes: Calculating mutual information using a mutual information function according to the talent characteristics, the job characteristics, the first distance vector, and the second distance vector; Conditional entropy is calculated based on the talent characteristics, the position characteristics, and the first distance vector.

[0008] In some embodiments, optimizing the distance vector based on the mutual information and the conditional entropy includes: Optimizing the first distance vector and the second distance vector based on the mutual information and the conditional entropy; The step of constructing a job competency model based on the optimized distance vector includes: The job competency model is constructed based on the optimized first distance vector and the second distance vector.

[0009] In some embodiments, calculating the mutual information using a mutual information function based on the talent characteristics, the position characteristics, the first distance vector, and the second distance vector includes: Calculating recovery variables based on the talent characteristics, the job characteristics, the first distance vector, and the second distance vector; the recovery variables include the recovery variables of the talent characteristics and the recovery variables of the job characteristics; The mutual information is calculated using a mutual information function according to the talent characteristics, the job characteristics, the first distance vector, the second distance vector, the recovery variable of the talent characteristics, and the recovery variable of the job characteristics.

[0010] In some embodiments, after calculating the restoration variable based on the talent characteristics, the position characteristics, the first distance vector, and the second distance vector, the method further includes: Calculating a distortion metric based on the talent characteristic, the job characteristic, the recovery variable of the talent characteristic, and the recovery variable of the job characteristic; the distortion metric is used to evaluate the difference between the recovery variable of the talent characteristic and the recovery variable of the job characteristic and the talent characteristic and the job characteristic; A rate distortion function is used to calculate rate distortion based on the talent characteristics, the job characteristics, and the distortion metric; the rate distortion is used to represent the information rate restored between the talent characteristics and the job characteristics.

[0011] In some embodiments, after calculating the distortion metric based on the talent characteristics, the job characteristics, the recovery variables of the talent characteristics, and the recovery variables of the job characteristics, the method further includes: When the distortion metric is within a preset range, calculating a conditional rate using a conditional rate-distortion function based on the talent characteristics, the job characteristics, and the distortion metric; Calculating conditional mutual information using a conditional mutual information function according to the talent characteristics, the job characteristics, the first distance vector, the second distance vector, the recovery variable of the talent characteristics, and the recovery variable of the job characteristics; determining an original range of a compression distance based on the conditional rate distortion and the conditional mutual information; Calculating the compression distance between the talent characteristics and the job characteristics; When the compression distance is outside the original range, the parameters of the conditional rate-distortion function and the conditional mutual information function are adjusted, and the compression distance is recalculated until the compression distance is within the original range.

[0012] In some embodiments, the method further comprises: The job competency model clusters jobs based on the conditional entropy; Talents are classified based on the clustering results, and matching of talents and positions is achieved based on the clustering results and the classification results.

[0013] A second aspect of the present application provides a job competency matching device, comprising: an acquisition module, configured to acquire raw data, wherein the raw data includes talent characteristics and job characteristics; A distance vector module, used to calculate the distance vector between the talent characteristics and the job characteristics; A calculation module, configured to calculate mutual information and conditional entropy based on the original data and the distance vector; an optimization module, configured to optimize the distance vector based on the mutual information and the conditional entropy; A construction module is used to construct a job competency model based on the optimized distance vector; the job competency model is used to match talents and jobs.

[0014] A third aspect of the present application provides an electronic device, including: processor; and The memory stores executable codes thereon, and when the executable codes are executed by the processor, the processor is caused to execute the method described above.

[0015] A fourth aspect of the present application provides a computer-readable storage medium having executable code stored thereon. When the executable code is executed by a processor of an electronic device, the processor is caused to execute the method described above.

[0016] The technical solution provided by this application may have the following beneficial effects: The present application provides a method for matching job competencies, including obtaining raw data, including talent characteristics and job characteristics; calculating the distance vector between the talent characteristics and job characteristics; calculating mutual information and conditional entropy based on the talent characteristics, job characteristics, and distance vector; optimizing the distance vector based on the mutual information and conditional entropy; constructing a job competency model based on the optimized distance vector; and using the job competency model to match talent and jobs. The above method combines a large database of AI algorithms with a job competency model and utilizes intelligent conditional entropy clustering calculations to achieve accurate classification and matching of job competencies, providing a new and more efficient solution for the field of human resources management.

[0017] It should be understood that the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the present application. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] The above and other objects, features and advantages of the present application will become more apparent by describing in more detail the exemplary embodiments of the present application in conjunction with the accompanying drawings, wherein the same reference numerals generally represent the same components in the exemplary embodiments of the present application.

[0019] Figure 1 1 is a flow chart of a method for matching job competencies according to an embodiment of the present application; Figure 2 1 is another flow chart of the job competency matching method shown in an embodiment of the present application; Figure 3 Schematic diagram of the structure of the job competency matching device shown in an embodiment of the present application; Figure 4 It is a structural diagram of an electronic device shown in an embodiment of the present application. DETAILED DESCRIPTION

[0020] The following describes embodiments of the present application in more detail with reference to the accompanying drawings. Although the accompanying drawings illustrate embodiments of the present application, it should be understood that the present application can be implemented in various forms and should not be limited by the embodiments described herein. Rather, these embodiments are provided to make the present application more thorough and complete, and to fully convey the scope of the present application to those skilled in the art.

[0021] The terms used in this application are for the purpose of describing specific embodiments only and are not intended to limit this application. As used in this application and the appended claims, the singular forms "a," "an," "the," and "the" are intended to include the plural forms, unless the context clearly indicates otherwise. It should also be understood that the term "and / or" as used herein refers to and encompasses any and all possible combinations of one or more of the associated listed items.

[0022] It should be understood that although the terms "first", "second", "third", etc. may be used in this application to describe various information, this information should not be limited to these terms. These terms are only used to distinguish information of the same type from each other. For example, without departing from the scope of this application, the first information may also be referred to as the second information, and similarly, the second information may also be referred to as the first information. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of the features. In the description of this application, the meaning of "plurality" is two or more, unless otherwise clearly and specifically defined.

[0023] The relevant methods do not make full use of job competency models and job attributes, and cannot effectively measure the degree of match between talent characteristics and job characteristics, thus affecting the efficiency and accuracy of talent selection and job allocation.

[0024] This application is proposed in this context. It aims to achieve accurate classification and matching of job competencies by combining the AI ​​algorithm big database and job competency model, and using the intelligent conditional entropy clustering calculation method, so as to provide a new and more efficient solution for the field of human resource management.

[0025] The technical solutions of the embodiments of the present application are described in detail below with reference to the accompanying drawings.

[0026] Figure 1 It is a flow chart of the job competency matching method shown in an embodiment of the present application.

[0027] See also Figure 1 , a job competency matching method, the method comprising: Step 110: Obtain original data, which includes talent characteristics and job characteristics.

[0028] Collect talent-related data, such as education, work experience, skill certificates, and personal traits, and organize it into a structured format. Collect job-related data, such as job responsibilities, skill requirements, and work experience, and organize it into a structured format, such as JSON (JavaScript Object Notation) or CSV (Comma-Separated Values) files. Preprocess the collected data, such as cleaning and removing missing and outliers. De-noise the data, such as removing irrelevant characters and whitespace from text data. Normalize the data so that all feature values ​​are on the same scale, such as using z-score normalization to convert the data to a distribution with a mean of 0 and a standard deviation of 1. Convert the preprocessed data into talent characteristics X and job characteristics Y. Both characteristics X and Y are data vectors, where each element in X represents a characteristic, such as education or work experience, and each element in Y represents a job characteristic, such as job responsibilities or skill requirements.

[0029] Set Y = X + N, where N is a Gaussian process with a variance of σ_N^2, representing the random error or difference between talent characteristic X and job characteristic Y. Furthermore, we can also set a correlation coefficient ρ between talent characteristic X and job characteristic Y to quantify the correlation between talent characteristic X and job characteristic Y. The correlation coefficient ρ can be calculated by calculating the Pearson correlation coefficient between talent characteristic X and job characteristic Y.

[0030] Step 120: Calculate the distance vector between the talent characteristics and the job characteristics.

[0031] The distance vector between talent characteristic X and position characteristic Y can be calculated using the Euclidean distance formula, the Manhattan distance formula, or other appropriate distance metric. The distance vector should retain sufficient information about the original data vector and maximize the mutual information between the data vector and the distance vector. Those skilled in the art will appreciate that this application does not restrict the method for calculating the distance vector. In practical applications, the most appropriate calculation method can be selected based on the application scenario.

[0032] Step 130 , calculating mutual information and conditional entropy based on talent characteristics, job characteristics, and distance vectors.

[0033] Based on talent characteristics, job characteristics, and distance vectors, we derive the relative entropy (KL divergence) between the joint probability distribution and the marginal probability distribution to calculate the mutual information I. This measures the correlation between these data. Furthermore, based on the entropy difference between the joint probability distribution and the conditional probability distribution, we can calculate the conditional entropy H. This conditional entropy measures the uncertainty of the data under given conditions.

[0034] Step 140: Optimize the distance vector based on mutual information and conditional entropy.

[0035] The distance vector is optimized according to the calculation results of mutual information I and conditional entropy H, so that the distance vector can maximize the retention of information about talent characteristics X and job characteristics Y and minimize the recovery error.

[0036] Step 150: construct a job competency model based on the optimized distance vector; the job competency model is used to match talents and jobs.

[0037] Based on the optimized distance vector, a job competency model is constructed using job attributes and a large database of AI (Artificial Intelligence) algorithms. This model can be based on a large language model. This model can be used to match talent with jobs. Matching results can include a match score, a list of recommended jobs, and recommended resumes.

[0038] The present application provides a method for matching job competencies, comprising obtaining raw data, including talent characteristics and job characteristics; calculating the distance vector between the talent characteristics and the job characteristics; calculating mutual information and conditional entropy based on the talent characteristics, job characteristics, and the distance vector; optimizing the distance vector based on the mutual information and conditional entropy; constructing a job competency model based on the optimized distance vector; and using the job competency model to match talent and jobs. The above method combines a large database of AI algorithms with a job competency model and utilizes intelligent conditional entropy clustering calculations to achieve precise classification and matching of job competencies, providing a new and more efficient solution for the field of human resources management.

[0039] Figure 2 This is another flow chart of the job competency matching method shown in the embodiment of the present application. Figure 2 , the job competency matching method includes the following steps: Step 210: Obtain original data, which includes talent characteristics and job characteristics.

[0040] Collect talent-related data, such as education, work experience, skill certificates, and personal traits, and organize this data into a structured format. Collect job-related data, such as job responsibilities, skill requirements, and work experience, and organize this data into a structured format, such as a JSON or CSV file. Preprocess the collected data, such as cleaning and removing missing and outliers. De-noise the data, such as removing irrelevant characters and whitespace from text data. Normalize the data so that all feature values ​​are on the same scale, for example, using z-score normalization to convert the data to a distribution with a mean of 0 and a standard deviation of 1. Convert the preprocessed data into talent characteristics X and job characteristics Y. Both talent characteristics X and job characteristics Y are data vectors, where each element in talent characteristics X represents a characteristic, such as education background or work experience, and each element in job characteristics Y represents a characteristic, such as job responsibilities or skill requirements.

[0041] Set Y = X + N, where N is a Gaussian process with a variance of σ_N^2, representing the random error or difference between talent characteristic X and job characteristic Y. Furthermore, we can also set a correlation coefficient ρ between talent characteristic X and job characteristic Y to quantify the correlation between talent characteristic X and job characteristic Y. The correlation coefficient ρ can be calculated by calculating the Pearson correlation coefficient between talent characteristic X and job characteristic Y.

[0042] Step 220 , calculating a first distance vector from the talent characteristics to the job characteristics, and a second distance vector from the job characteristics to the talent characteristics.

[0043] The first distance vector S_Y from the talent feature X to the position feature Y may be calculated using the Euclidean distance formula, the Manhattan distance formula, or other suitable distance measurement methods.

[0044] The second distance vector S_X from the position feature Y to the talent feature X may be calculated using the Euclidean distance formula, the Manhattan distance formula, or other suitable distance measurement methods.

[0045] Step 230 : Calculate mutual information using a mutual information function based on the talent characteristics, the job characteristics, the first distance vector S_Y, and the second distance vector.

[0046] Based on the talent feature X, job feature Y, the first distance vector S_Y, and the second distance vector S_X, the mutual information function is used to calculate the mutual information I based on the relative entropy (KL divergence) between the joint probability distribution and the marginal probability distribution. The mutual information I can measure the correlation between the talent features, job features, and distance vectors.

[0047] In an optional embodiment of the present application, step 230 includes: Sub-step 231 , calculating recovery variables based on the talent characteristics, job characteristics, the first distance vector, and the second distance vector; the recovery variables include recovery variables of the talent characteristics and recovery variables of the job characteristics.

[0048] By encoding the talent feature X, the job feature Y, the first distance vector S_Y and the second distance vector S_X and performing distance calculation, the restored variable X1 of the talent feature X and the restored variable Y1 of the job feature Y can be restored.

[0049] Sub-step 232 , calculating the mutual information using a mutual information function based on the talent characteristics, the job characteristics, the first distance vector, the second distance vector, the recovery variables of the talent characteristics, and the recovery variables of the job characteristics.

[0050] According to the talent feature X, the position feature Y, the first distance vector S_Y, the second distance vector S_X, the talent feature recovery variable X1, and the position feature recovery variable Y1, the mutual information I is calculated based on the relevant function or self-written code through the relative entropy (KL divergence) between the joint probability distribution and the marginal probability distribution. Specifically, these mutual information amounts can be calculated by using the mutual information function using the Bayesian formula and the definition of entropy.

[0051] Calculate I(Y;S_Y|X), that is, the mutual information between job feature Y and the first distance vector S_Y when talent feature X is given, to measure the correlation between job feature Y and the first distance vector S_Y given talent feature X.

[0052] Calculate I(X;S_X|Y), that is, the mutual information between talent feature X and the second distance vector S_X when the job feature Y is given, to measure the correlation between talent feature X and the second distance vector S_X given the job feature Y.

[0053] Calculate I(Y; Y1|XS_Y), that is, the mutual information between job characteristic Y and recovery variable Y1 when talent characteristic X and the first distance vector S_Y are given, to measure the accuracy of the recovery variable Y1's estimation of job characteristic Y given talent characteristic X and the first distance vector S_Y.

[0054] Calculate I(X; X1|YS_X), that is, the mutual information between talent characteristic X and recovery variable X1 when the job characteristic Y and the second distance vector S_X are given, to measure the estimation accuracy of talent characteristic X by recovery variable X1 given the job characteristic Y and the second distance vector S_X.

[0055] In an optional embodiment of the present application, after calculating the recovery variable according to the talent characteristics, the position characteristics, the first distance vector, and the second distance vector, the method further includes: The distortion measure is calculated based on the talent characteristics, job characteristics, the recovery variables of the talent characteristics, and the recovery variables of the job characteristics; the distortion measure is used to evaluate the difference between the recovery variables of the talent characteristics and the recovery variables of the job characteristics and the talent characteristics and the job characteristics; The rate distortion is calculated using a rate-distortion function based on talent characteristics, job characteristics, and distortion metrics; the rate distortion is used to represent the information rate restored between talent characteristics and job characteristics.

[0056] Based on talent characteristics X and job characteristics Y, and the talent characteristic recovery variables X1 and Y1, a distortion metric D is calculated by calculating the squared error. Distortion metric D assesses the difference between the talent characteristic recovery variables and the job characteristic recovery variables and talent characteristics X and Y. Furthermore, it can be determined whether distortion metric D is greater than σ_N^2. If so, talent characteristics X and job characteristics Y are considered to be mutually estimable, indicating a high degree of match between the talent characteristics and job characteristics.

[0057] Based on talent characteristics X, job characteristics Y, and distortion metric D, a rate-distortion function is used to calculate rate distortion. Rate distortion is used to represent the minimum information rate required to recover between talent characteristics X and job characteristics Y. The specific form of the rate-distortion function can be determined through experiments or theoretical derivation, or it can be approximated using an existing rate-distortion function model. In one example, the rate distortion R_(X|Y)(D) is set as the rate-distortion function of talent characteristics X when job characteristics Y are given. It represents the minimum information rate required to recover talent characteristics X from job characteristics Y under a given distortion metric D. Similarly, R_(Y|X)(D) is set as the rate-distortion function of job characteristics Y when talent characteristics X are given. Similarly, it represents the minimum information rate required to recover job characteristics Y from talent characteristics X.

[0058] In an optional embodiment of the present application, after calculating the distortion metric based on the talent characteristics, the job characteristics, the recovery variables of the talent characteristics, and the recovery variables of the job characteristics, the following steps are further included: When the distortion metric is within a preset range, a conditional rate distortion function is used to calculate the conditional rate distortion based on the talent characteristics, job characteristics, and the distortion metric; The conditional mutual information function is used to calculate the conditional mutual information according to the talent characteristics, job characteristics, the first distance vector, the second distance vector, the recovery variable of the talent characteristics, and the recovery variable of the job characteristics; Determine the original range of the compression distance based on the conditional rate distortion and conditional mutual information; Calculate the compression distance between talent characteristics and job characteristics; When the compression distance is outside the original range, the parameters of the conditional rate-distortion function and the conditional mutual information function are adjusted, and the compression distance is recalculated until the compression distance is within the original range.

[0059] This embodiment provides a method for accurately measuring the compression distance between two original data within a given distortion range, and performing information processing or data matching based on this. By introducing the concepts of conditional rate-distortion function and conditional mutual information, this method realizes an accurate description and quantification of the complex relationship between two original data.

[0060] If the distortion measure D is within the preset range, that is, 0 < D < σ_N^2, the conditional rate-distortion functions R_(X|Y)(D) and R_(Y|X)(D) are calculated based on the talent characteristics X, the job characteristics Y, and the distortion measure D. The calculation of the conditional rate-distortion can be based on the rate-distortion theory in information theory and solved using existing numerical methods or algorithms. For example, iterative methods such as the Blahut-Arimoto algorithm can be used for numerical calculation. During the calculation process, the joint probability distribution and conditional probability distribution of the talent characteristics X and the job characteristics Y, as well as the constraint conditions of the distortion measure D, need to be considered.

[0061] According to the talent characteristics X, the job characteristics Y, the first distance vector S_Y, the second distance vector S_X, the recovery variable X1 of the talent characteristics, and the recovery variable Y1 of the job characteristics, the conditional mutual information I is calculated based on the relative entropy (KL divergence) between the joint probability distribution and the marginal probability distribution through a correlation function or self-written code. Specifically, these conditional mutual informations can be calculated through the conditional mutual information function using Bayes' formula and the definition of entropy.

[0062] Calculate I(Y; S_Y|X), that is, the mutual information between the job characteristics Y and the first distance vector S_Y given the talent characteristics X, which measures the correlation between the job characteristics Y and the first distance vector S_Y given the talent characteristics X.

[0063] Calculate I(X; S_X|Y), that is, the mutual information between the talent characteristics X and the second distance vector S_X given the job characteristics Y, which measures the correlation between the talent characteristics X and the second distance vector S_X given the job characteristics Y.

[0064] Calculate I(Y; Y1|XS_Y), that is, the mutual information between the job characteristics Y and the recovery variable Y1 given the talent characteristics X and the first distance vector S_Y, which measures the estimation accuracy of the recovery variable Y1 for the job characteristics Y given the talent characteristics X and the first distance vector S_Y.

[0065] Calculate I(X; X1|YS_X), that is, the mutual information between the talent characteristics X and the recovery variable X1 given the job characteristics Y and the second distance vector S_X, which measures the estimation accuracy of the recovery variable X1 for the talent characteristics X given the job characteristics Y and the second distance vector S_X.

[0066] Based on the calculation results of the conditional rate-distortion function, a lower bound for the compression distance CD(X,Y) between talent characteristic X and job characteristic Y is derived. Specifically, the compression distance CD(X,Y) should be greater than or equal to the larger of R_(X|Y)(D) and R_(Y|X)(D), that is, CD(X,Y) ≥ max{R_(X|Y)(D),R_(Y|X)(D)}. This step ensures that the lower bound of the compression distance reflects the minimum information recovery required between talent characteristic X and job characteristic Y under a given distortion metric D. Based on the calculation results of the conditional mutual information, an upper bound for the compression distance CD(X,Y) between two data vectors X and Y is derived. Specifically, the compression distance CD(X,Y) should be less than or equal to the infimum of the following expression: CD(X,Y)≤inf{max{I(Y;S_Y|X),I(X;S_X|Y)}+I(Y;Y1|XS_Y)+I(X;X1|YS_X)}. This step ensures that the upper bound of the compression distance reflects the maximum information loss of talent characteristics X and job characteristics Y under given distortion D and mutual information constraints. This process determines the original range of the compression distance.

[0067] The above method can be applied to actual information processing or data matching tasks for job competency models. Based on the specific application scenario and needs of the enterprise, the talent characteristics X and job characteristics Y, the distortion metric D, the first distance vector S_Y, and the second distance vector S_X can be set and adjusted. Specifically, the compressed distance CD(X,Y) between the talent characteristics and job characteristics can be calculated to determine whether CD(X,Y) is within the original range. If the compressed distance CD(X,Y) is outside the original range, the parameters of the conditional rate-distortion function and the conditional mutual information function are adjusted, and the compressed distance CD(X,Y) is recalculated until it is within the original range. Experimental verification and cross-validation can be used to evaluate and optimize the parameter settings. Furthermore, the matching results of the actual processing process can be evaluated, including the accuracy, robustness, and computational efficiency of the information processing or data matching. Based on the evaluation results, the model and function parameters can be further improved and optimized.

[0068] Step 240 , calculating conditional entropy based on the talent characteristics, job characteristics, and the first distance vector.

[0069] To find the conditional entropy H(Y|X) of position feature Y given talent feature X, we can calculate the joint probability distribution of position feature Y and the marginal probability distribution of position feature Y given talent feature X, then calculate the entropy difference to obtain the conditional entropy H(Y|X). Similarly, to find the conditional entropy H(Y|XS_Y) of position feature Y given talent feature X and the first distance vector S_Y, we can calculate the joint probability distribution of position feature Y and the marginal probability distribution of position feature Y given talent feature X and the first distance vector S_Y, then calculate the entropy difference to obtain the conditional entropy H(Y|XS_Y). Specifically, you can use relevant functions or write your own code to perform the calculation.

[0070] Step 250: Optimize the first distance vector and the second distance vector based on mutual information and conditional entropy.

[0071] First, you can set the specific method used by the optimization algorithm, such as gradient descent or stochastic gradient descent. Second, you can set the optimization algorithm parameters, such as the learning rate and number of iterations. After setting the optimization algorithm, construct an objective function based on the calculated results of mutual information I and conditional entropy H, aiming to maximize mutual information I and minimize conditional entropy H. Use the selected optimization algorithm to optimize the objective function, iteratively updating the values ​​of the first distance vector S_Y and the second distance vector S_X. Monitor the optimization process and stop when the objective function value converges or the preset number of iterations is reached, resulting in the optimized first and second distance vectors.

[0072] Using optimization algorithms such as gradient descent, the distance vectors S_Y and S_X are optimized based on the calculation results of mutual information and conditional entropy, so that they can maximize the retention of information about talent characteristics X and job characteristics Y and minimize the recovery error, thereby improving the accuracy and efficiency of the matching process.

[0073] Step 260: construct a job competency model based on the optimized first distance vector and the second distance vector; the job competency model is used to match talents and jobs.

[0074] Using the optimized first and second distance vectors S_Y and S_X, combined with job attributes and a large database of AI algorithms, a job competency model is constructed. This model is used to match talent with jobs. Machine learning algorithms, such as support vector machines and random forests, can be used to predict job competencies based on distance vectors and job attributes.

[0075] In an optional embodiment of the present application, the method further includes: The job competency model clusters jobs based on conditional entropy; Talents are classified based on clustering results, and talents and positions are matched based on clustering and classification results.

[0076] Leveraging the concept of conditional entropy, we perform cluster analysis on job competencies, identifying groups of competencies with similar characteristics. Talents are then categorized based on the clustering results, achieving precise job competency matching. Based on the clustering results and conditional entropy analysis, we output talent-job matching results, including a match score and a list of recommended positions. These matching results can be presented visually to users, such as through tables, charts, or interactive interfaces. Download and export functions are provided, allowing users to easily save matching results locally or conduct further analysis.

[0077] Clustering algorithms can be selected based on specific scenarios and requirements, such as K-means, hierarchical clustering, or DBSCAN (Density-Based Spatial Clustering of Applications with Noise). Parameters for clustering algorithms, such as the number of clusters and distance metric, can be set. The quality of clustering results can also be evaluated using metrics such as the silhouette coefficient and the Calinski-Harabasz index (CH index). Similarity metrics, such as cosine similarity and Euclidean distance, can be used to assess the degree of fit between talent data and job competency groups. The degree of fit can be comprehensively assessed using methods such as weighted average and the maximum-minimum method.

[0078] In the following example of this application, a method for calculating the mutual information based on conditional information and conditional entropy under an ideal compression distance is given. If the mutual information between the position characteristic Y and the recovery variable Y1 is 0 under the conditions of setting the talent characteristic X and the first distance vector S_Y, that is, I(Y; Y1|XS_Y)=0, it means that under the conditions of knowing the talent characteristic X and the first distance vector S_Y, there is no additional shared information between the position characteristic Y and the recovery variable Y1, which usually means a perfect or sufficiently accurate estimate of the position characteristic Y and the recovery variable Y1 under the given conditions. Similarly, it is also set that under the conditions of giving the position characteristic Y and the second distance vector S_X, the mutual information between the talent characteristic X and the recovery variable X1 is 0, that is, I(X; X1|YS_X)=0, which means that under the conditions of knowing the position characteristic Y and the second distance vector S_X, there is no additional shared information between the talent characteristic X and the recovery variable X1.

[0079] Based on the basic relationship between conditional mutual information (I) and conditional entropy (H), we can derive the expression I(Y;S_Y|X): I(Y;S_Y|X) = H(Y|X) - H(Y|XS_Y). This formula states that, given talent characteristic X, the mutual information between job characteristic Y and S_Y is equal to the conditional entropy of job characteristic Y given X minus the conditional entropy of job characteristic Y given talent characteristic X and the first distance vector S_Y.

[0080] When the job characteristic Y follows a Gaussian distribution given the talent characteristic X, the entropy formula of the Gaussian distribution can be used to calculate H(Y|X): H(Y|X) = 1 / 2∙log (2πeσ_N^2), where σ_N^2 represents the noise variance.

[0081] To calculate H(Y|XS_Y), we need to determine the specific form or representation of the first distance vector S_Y. This often depends on the specific application scenario and data characteristics. For example, the first distance vector S_Y can be a linear transformation, nonlinear transformation, feature extraction result, or compressed representation of the position feature Y. Once the form of the first distance vector S_Y is determined, H(Y|XS_Y) can be solved according to the definition and calculation method of conditional entropy.

[0082] Substituting the calculated results of H(Y|X) and H(Y|XS_Y) into the formula I(Y;S_Y|X), we obtain: I(Y;S_Y|X)=1 / 2∙log (2πeσ_N^2)-H(Y|XS_Y). This formula reveals the relationship between the mutual information and noise variance between the position characteristic Y and the first distance vector S_Y given the talent characteristic X, as well as the conditional entropy of the position characteristic Y given the talent characteristic X and the first distance vector S_Y.

[0083] During data compression in job competency models, I(Y;S_Y|X) can be used to measure the amount of information lost when the first distance vector S_Y is used as a compressed representation of job feature Y. Specifically, the I(Y;S_Y|X) values ​​can be calculated for different compressed representations, and the one with the least information loss is selected as the optimal solution. This method can also be used to uncover potential relationships and patterns between data during job competency data matching analysis. The conditional mutual information (CMI) between two datasets can be calculated, and its magnitude can be used to determine their correlation or similarity.

[0084] The above methods can be appropriately adjusted and optimized based on specific application scenarios and requirements. For example, more auxiliary variables can be introduced to improve the accuracy of information extraction; the optimal compression representation or feature extraction method can be automatically learned; and other information theory methods or data analysis techniques can be combined to further improve performance. During the optimization process, it is necessary to set appropriate thresholds or objective functions based on the specific application scenario and requirements. For example, in data compression, an upper limit on the amount of information loss can be set to ensure that the compressed data is still sufficiently usable; in data matching analysis, a correlation or similarity threshold can be set to filter out qualified data pairs. When calculating the conditional entropy H(Y|XS_Y), it is important to ensure that the selection and representation of the first distance vector S_Y accurately reflects the information characteristics of the job characteristic Y given the talent characteristic X. Improper selection or inaccurate representation of the first distance vector S_Y can lead to biased or inaccurate calculation results.

[0085] When using I(Y;S_Y|X) for data processing or optimization, it is important to fully consider the characteristics of the data and the limitations of the application scenario. For example, in data compression, it is necessary to balance the relationship between information loss and compression rate; in data matching analysis, it is necessary to consider the impact of factors such as data dimension, distribution, and noise on the results.

[0086] In practical applications, other information theory methods or data analysis techniques can be combined to further improve performance. For example, feature extraction methods such as principal component analysis (PCA) and linear discriminant analysis (LDA) can be used to reduce data dimensionality and improve information utilization. Machine learning algorithms such as support vector machines (SVM) and neural networks can also be used for tasks such as classification, regression, and clustering.

[0087] This example proposes a method for calculating mutual information under an ideal compression distance. By setting the mutual information under specific conditions and leveraging the relationship between conditional mutual information and conditional entropy, a formula for calculating the mutual information between talent characteristic Y and its restored variable Y1 (or position characteristic X and its restored variable X1) under given conditions is derived. This method provides a new approach and solution for accurately measuring the compression distance between two data vectors within a given distortion metric.

[0088] The job competency matching method proposed in the embodiments of the present application is a job competency matching method based on intelligent conditional entropy clustering calculation. This method combines an AI algorithm big database and a job competency model to achieve accurate classification and matching of job competencies. By constructing a job competency and talent data model based on distance measurement, the distance from talent feature X to job feature Y, as well as the reverse distance, is calculated, and the concepts of mutual information and conditional entropy are introduced to measure the amount of information and uncertainty carried by these distances, which can more accurately reflect the degree of matching between talent features and job features. By calculating the conditional entropy between different job competencies, job competency groups with similar characteristics can be identified, and talent data can be classified based on these clustering results to achieve accurate matching of job competencies, which not only improves the accuracy of matching, but also enhances the robustness and stability of matching.

[0089] Figure 3 It is a structural diagram of a job competency matching device shown in an embodiment of the present application.

[0090] See also Figure 3 , the job competency matching device 300 of this embodiment: An acquisition module 310 is used to acquire raw data, including talent characteristics and job characteristics; A distance vector module 320 is used to calculate the distance vector between talent characteristics and job characteristics; A calculation module 330 is used to calculate mutual information and conditional entropy based on the original data and the distance vector; An optimization module 340 is configured to optimize the distance vector based on mutual information and conditional entropy; The construction module 350 is used to construct a job competency model based on the optimized distance vector; the job competency model is used to match talents and jobs.

[0091] In an optional embodiment of the present application, the distance vector module 320 includes: A distance vector submodule, used to calculate a first distance vector from talent characteristics to job characteristics, and a second distance vector from job characteristics to talent characteristics; The calculation module 330 includes: A first calculation submodule is configured to calculate mutual information using a mutual information function based on the talent characteristics, the job characteristics, the first distance vector, and the second distance vector; The second calculation submodule is used to calculate the conditional entropy according to the talent characteristics, the position characteristics and the first distance vector.

[0092] The optimization module 340 includes: An optimization submodule, configured to optimize the first distance vector and the second distance vector based on mutual information and conditional entropy; The job competency model is constructed based on the optimized distance vector, including: A submodule is constructed to construct a job competency model based on the optimized first distance vector and the second distance vector.

[0093] The first calculation submodule includes: A recovery variable unit, configured to calculate a recovery variable based on the talent characteristics, the job characteristics, the first distance vector, and the second distance vector; the recovery variable includes a recovery variable of the talent characteristics and a recovery variable of the job characteristics; The mutual information unit is used to calculate the mutual information using a mutual information function according to the talent characteristics, the job characteristics, the first distance vector, the second distance vector, the recovery variable of the talent characteristics, and the recovery variable of the job characteristics.

[0094] The device also includes: a distortion metric module, configured to calculate distortion metrics based on talent characteristics, job characteristics, recovery variables of talent characteristics, and recovery variables of job characteristics; the distortion metrics are configured to evaluate the differences between the recovery variables of talent characteristics and the recovery variables of job characteristics and the talent characteristics and job characteristics; The rate distortion module is used to calculate the rate distortion based on the talent characteristics, job characteristics and distortion measurement using the rate distortion function; the rate distortion is used to represent the information rate restored between the talent characteristics and the job characteristics.

[0095] The device also includes: A conditional rate module, configured to calculate a conditional rate using a conditional rate distortion function based on talent characteristics, job characteristics, and distortion metrics when the distortion metric is within a preset range; A conditional mutual information module is used to calculate the conditional mutual information using a conditional mutual information function according to the talent characteristics, the job characteristics, the first distance vector, the second distance vector, the recovery variable of the talent characteristics, and the recovery variable of the job characteristics; An original range module, used to determine the original range of the compression distance based on conditional rate distortion and conditional mutual information; Compression distance module, used to calculate the compression distance between talent characteristics and job characteristics; The adjustment module is used to adjust the parameters of the conditional rate-distortion function and the conditional mutual information function when the compression distance is outside the original range, and recalculate the compression distance until the compression distance is within the original range.

[0096] The device also includes: Clustering module, used for job competency model to cluster jobs based on conditional entropy; The classification module is used to classify talents based on clustering results and match talents with positions based on clustering results and classification results.

[0097] The embodiments of the present application provide a job competency matching device, which can be based on a job competency matching method based on intelligent conditional entropy clustering calculation. The method combines an AI algorithm large database and a job competency model to achieve accurate classification and matching of job competencies. By constructing a job competency and talent data model based on distance measurement, the distance from talent feature X to job feature Y, as well as the reverse distance, is calculated, and the concepts of mutual information and conditional entropy are introduced to measure the amount of information and uncertainty carried by these distances, which can more accurately reflect the degree of matching between talent features and job features. By calculating the conditional entropy between different job competencies, job competency groups with similar characteristics can be identified, and talent data can be classified based on these clustering results to achieve accurate matching of job competencies, which not only improves the accuracy of matching, but also enhances the robustness and stability of matching.

[0098] Regarding the apparatus in the above embodiment, the specific manner in which each module performs operations has been described in detail in the embodiment of the method, and will not be elaborated again here.

[0099] Figure 4 It is a structural diagram of an electronic device shown in an embodiment of the present application.

[0100] See also Figure 4 , the electronic device 400 includes a memory 410 and a processor 420.

[0101] The processor 420 may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or any conventional processor.

[0102] Memory 410 may include various types of storage units, such as system memory, read-only memory (ROM), and permanent storage. ROM may store static data or instructions required by processor 420 or other computer modules. Permanent storage may be a readable and writable storage device. Permanent storage may be a non-volatile storage device that retains stored instructions and data even when the computer is powered off. In some embodiments, the permanent storage device utilizes a mass storage device (e.g., a magnetic or optical disk, flash memory). In other embodiments, the permanent storage device may be a removable storage device (e.g., a floppy disk, optical drive). System memory may be a readable and writable storage device or a volatile readable and writable storage device, such as dynamic random access memory (DRAM). System memory may store some or all instructions and data required by the processor during operation. Furthermore, memory 410 may include any combination of computer-readable storage media, including various types of semiconductor memory chips (e.g., DRAM, SRAM, SDRAM, flash memory, programmable read-only memory), as well as magnetic disks and / or optical disks. In some embodiments, the memory 410 may include a readable and / or writable removable storage device, such as a compact disc (CD), a read-only digital versatile disc (e.g., DVD-ROM, double-layer DVD-ROM), a read-only Blu-ray disc, an ultra-density optical disc, a flash memory card (e.g., SD card, mini SD card, Micro-SD card, etc.), a magnetic floppy disk, etc. Computer-readable storage media do not contain carrier waves and transient electronic signals transmitted wirelessly or wired.

[0103] The memory 410 stores executable codes. When the executable codes are processed by the processor 420 , the processor 420 may execute part or all of the above-mentioned methods.

[0104] In addition, the method according to the present application may also be implemented as a computer program or a computer program product, which includes computer program code instructions for executing some or all of the steps in the above method of the present application.

[0105] Alternatively, the present application can also be implemented as a computer-readable storage medium (or non-transitory machine-readable storage medium or machine-readable storage medium) on which executable code (or computer program or computer instruction code) is stored. When the executable code (or computer program or computer instruction code) is executed by a processor of an electronic device (or server, etc.), the processor executes part or all of the steps of the above-mentioned method according to the present application.

[0106] The embodiments of the present application have been described above. The above description is exemplary, not exhaustive, and is not limited to the disclosed embodiments. Many modifications and variations will be apparent to those skilled in the art without departing from the scope and spirit of the described embodiments. The terminology used herein is selected to best explain the principles of the embodiments, their practical applications, or improvements to the technology in the market, or to enable other persons skilled in the art to understand the embodiments disclosed herein.

Claims

1. A job competency matching method, characterized in that: include: Obtaining original data, wherein the original data includes talent characteristics and job characteristics; Calculating the distance vector between the talent characteristics and the job characteristics; Calculating mutual information and conditional entropy based on the talent characteristics, the job characteristics, and the distance vector; Optimizing the distance vector based on the mutual information and the conditional entropy; A job competency model is constructed based on the optimized distance vector; the job competency model is used to match talents and jobs.

2. The method according to claim 1, characterized in that The calculating of the distance vector between the talent characteristics and the position characteristics includes: Calculating a first distance vector from the talent feature to the position feature, and a second distance vector from the position feature to the talent feature; The calculating of mutual information and conditional entropy based on the talent characteristics, the job characteristics, and the distance vector includes: Calculating mutual information using a mutual information function according to the talent characteristics, the job characteristics, the first distance vector, and the second distance vector; Conditional entropy is calculated based on the talent characteristics, the position characteristics, and the first distance vector.

3. The method according to claim 2, characterized in that The optimizing the distance vector based on the mutual information and the conditional entropy includes: Optimizing the first distance vector and the second distance vector based on the mutual information and the conditional entropy; The step of constructing a job competency model based on the optimized distance vector includes: The job competency model is constructed based on the optimized first distance vector and the second distance vector.

4. The method according to claim 2, characterized in that The calculating the mutual information using a mutual information function according to the talent characteristics, the position characteristics, the first distance vector, and the second distance vector includes: Calculating recovery variables based on the talent characteristics, the job characteristics, the first distance vector, and the second distance vector; the recovery variables include the recovery variables of the talent characteristics and the recovery variables of the job characteristics; The mutual information is calculated using a mutual information function according to the talent characteristics, the job characteristics, the first distance vector, the second distance vector, the recovery variable of the talent characteristics, and the recovery variable of the job characteristics.

5. The method according to claim 4, characterized in that After calculating the recovery variable according to the talent characteristics, the position characteristics, the first distance vector, and the second distance vector, the method further includes: Calculating a distortion metric based on the talent characteristic, the job characteristic, the recovery variable of the talent characteristic, and the recovery variable of the job characteristic; the distortion metric is used to evaluate the difference between the recovery variable of the talent characteristic and the recovery variable of the job characteristic and the talent characteristic and the job characteristic; A rate distortion function is used to calculate rate distortion based on the talent characteristics, the job characteristics, and the distortion metric; the rate distortion is used to represent the information rate restored between the talent characteristics and the job characteristics.

6. The method according to claim 5, characterized in that After calculating the distortion measure according to the talent characteristics, the job characteristics, the recovery variables of the talent characteristics, and the recovery variables of the job characteristics, the method further includes: When the distortion metric is within a preset range, calculating a conditional rate using a conditional rate-distortion function based on the talent characteristics, the job characteristics, and the distortion metric; Calculating conditional mutual information using a conditional mutual information function according to the talent characteristics, the job characteristics, the first distance vector, the second distance vector, the recovery variable of the talent characteristics, and the recovery variable of the job characteristics; determining an original range of a compression distance based on the conditional rate distortion and the conditional mutual information; Calculating the compression distance between the talent characteristics and the job characteristics; When the compression distance is outside the original range, the parameters of the conditional rate-distortion function and the conditional mutual information function are adjusted, and the compression distance is recalculated until the compression distance is within the original range.

7. The method according to claim 1, characterized in that The method further comprises: The job competency model clusters jobs based on the conditional entropy; Talents are classified based on the clustering results, and matching of talents and positions is achieved based on the clustering results and the classification results.

8. A job competency matching device, characterized in that: include: An acquisition module, configured to acquire raw data, wherein the raw data includes talent characteristics and job characteristics; A distance vector module, used to calculate the distance vector between the talent characteristics and the job characteristics; A calculation module, configured to calculate mutual information and conditional entropy using a mutual information function according to the original data and the distance vector; an optimization module, configured to optimize the distance vector based on the mutual information and the conditional entropy; A construction module is used to construct a job competency model based on the optimized distance vector; the job competency model is used to match talents and jobs.

9. An electronic device, characterized in that: include: processor; as well as A memory having executable codes stored thereon, which, when executed by the processor, causes the processor to execute the method according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that An executable code is stored thereon, and when the executable code is executed by a processor of an electronic device, the processor is caused to execute the method according to any one of claims 1 to 7.

Images