People and post matching method and device, storage medium and computer program product

Through a large language model, analyzing the client input text, building an impact matrix and capability matrix, and automatically recommending positions, solving the problems of low efficiency and poor accuracy in person-position matching, and achieving efficient and accurate automated matching.

CN118982184APending Publication Date: 2024-11-19CHINA TELECOM CORP LTD
View PDF 0 Cites 2 Cited by

Patent Information

Application Number
CN202411046811.5
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-07-31
Publication Date
2024-11-19

AI Technical Summary

Technical Problem

The existing technology has low matching efficiency and poor accuracy in human-job matching, and reliance on manual analysis leads to waste of resources and inefficiency.

Method used

A large language model is used to analyze the text input by the client, build an impact matrix and capability matrix, predict employee ability scores through hierarchical analysis method and deep learning model, and automatically recommend the most suitable positions.

Benefits of technology

It improves the accuracy and efficiency of human-position matching, reduces human resources investment, and achieves automated and efficient matching.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118982184A_ABST
    Figure CN118982184A_ABST
Patent Text Reader

Abstract

The invention discloses a person and post matching method and device, a storage medium and a computer program product. Comprising the steps of obtaining a target input text, and analyzing the target input text by using a large language model to obtain a text tag; when the tag is the man-post matching text, analyzing the man-post matching text by using a large language model to obtain target post demand information related to a target post and the influence degree of a related target capability item on the target post demand information so as to construct an influence matrix; obtaining to-be-recommended employees with each target ability item and target ability item scores of the to-be-recommended employees from the employee ability item set by using a large language model to construct an ability matrix; and analyzing the two matrixes by using a large language model, determining a recommendation score of each to-be-recommended employee, and taking the to-be-recommended employee with the recommendation score higher than a preset threshold as a target recommended employee of the target post. According to the invention, the technical problems of low matching efficiency and poor accuracy of man-post matching in a manual mode are solved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present application relates to the field of intelligent network operation technology, and more specifically, to a person-job matching method, device, storage medium and computer program product. Background Art

[0002] With the development of business, many companies have seen a surge in business volume and an increasing number of employees. In order to effectively allocate and rationally use human resources, it is necessary to ensure a two-way match between employees and positions, so as to match employees to the most suitable positions. However, when matching people with positions, the relevant technologies manually analyze the degree of match between position information and employee information, resulting in a low matching rate and poor matching efficiency.

[0003] To address the above-mentioned problems, no effective solution has been proposed yet. Summary of the invention

[0004] The embodiments of the present application provide a person-job matching method, device, storage medium and computer program product to at least solve the technical problems of low matching efficiency and poor accuracy in manual person-job matching.

[0005] According to one aspect of an embodiment of the present application, a person-job matching method is provided, including: obtaining a target input text of a client, and analyzing the target input text using a large language model to obtain a text label of the target input text, wherein the text label is used to mark whether the input text is a person-job matching text; in the case where the text label is a person-job matching text, analyzing the person-job matching text using a large language model to obtain at least one target position requirement information related to the target position and the degree of influence of at least one target capability item related to the target position requirement information on the target position requirement information, so as to construct an influence matrix; obtaining employees to be recommended who have each target capability item and the target capability item scores of the employees to be recommended from a preset employee capability item set using the large language model to construct a capability matrix; analyzing the influence matrix and the capability matrix using the large language model to determine the recommendation score of each employee to be recommended, and selecting employees to be recommended whose recommendation score is higher than a preset threshold as target recommended employees for the target position.

[0006] Optionally, the target input text is analyzed using a large language model to obtain a text label of the target input text, including: obtaining a sample data set for iteratively training the large language model, wherein the sample data set includes multiple groups of sample data consisting of sample texts and text labels corresponding to the sample texts; using the large language model to analyze the text similarity between the target input text and each sample text, and using the text label of the target sample text whose text similarity is higher than a preset similarity threshold as the text label of the target input text.

[0007] Optionally, a large language model is used to analyze the person-job matching text to obtain at least one target job requirement information related to the target job and the degree of influence of at least one target capability item related to the target job requirement information on the target job requirement information, so as to construct an influence matrix, including: using a large language model to analyze the person-job matching text to extract at least one target job requirement information related to the target job; obtaining a job information database, wherein the job information database includes at least one capability item corresponding to multiple job requirement information related to the job and the degree of influence of each capability item on the job requirement information; matching each target job requirement information with the job information database to obtain at least one target capability item corresponding to each target job requirement information and the target degree of influence of each target capability item on the corresponding target job requirement information; constructing an influence matrix with the target job requirement information as rows, the target capability items as columns, and the target degree of influence of the target capability items on the target job requirement information as matrix elements.

[0008] Optionally, a large language model is used to obtain employees to be recommended who have each target competency item and the target competency item scores of the employees to be recommended from a preset set of employee competency items to construct a competency matrix, including: obtaining an employee competency item set, wherein the employee competency item set includes competency items possessed by multiple employees and corresponding competency item scores, and the competency item scores are determined by the first performance data of the employees in a historical time period and the second performance data predicted in a future time period; fuzzy matching each target competency item with the employee competency item set to obtain employees to be recommended who have the target competency item and the target competency item scores of the employees to be recommended; constructing a competency matrix with the target competency items as rows, the employees to be recommended as columns, and the target competency item scores of the employees to be recommended as matrix elements.

[0009] Optionally, the process of determining an employee's competency item score includes: obtaining the employee's first position data in a historical time period, and preprocessing the first position data, wherein the first position data includes at least one of the following: attendance data, salary data, and first performance data for characterizing the competency item; using a non-time-series linear non-Gaussian model to extract features from the preprocessed first position data to obtain at least one feature vector, wherein the type of feature vector includes at least: a feature vector for characterizing the competency item; using a pre-trained prediction model to analyze each feature vector to predict the employee's second performance data in a future time period, wherein the prediction model is a deep learning model that combines a convolutional neural network, a long short-term memory network, and a self-attention mechanism; determining the employee's competency item score based on the employee's first position data in the historical time period and the employee's second performance data in the future time period.

[0010] Optionally, a large language model is used to analyze the influence matrix and the capability matrix to determine the recommendation score of each employee to be recommended, including: converting the capability matrix into a comparison matrix, wherein the comparison matrix has each employee to be recommended as a row and a column, and the relative capability item scores of the employees to be recommended on the target capability items are used as matrix elements of a second comparison matrix; calculating the maximum eigenvalues ​​and corresponding eigenvectors of the influence matrix and the comparison matrix respectively, and determining the consistency verification results of the influence matrix and the comparison matrix respectively based on the maximum eigenvalues; when the consistency verification results of the influence matrix and the comparison matrix are both verified to be passed, the recommendation score of each employee to be recommended is obtained based on the sum of the products of the eigenvectors of the influence matrix and the column vectors in the capability matrix.

[0011] Optionally, the consistency check results of the influence matrix and the comparison matrix are respectively determined based on the maximum eigenvalue, including: for any one of the influence matrix and the comparison matrix, determining a consistency index based on the maximum eigenvalue of the matrix; determining a random consistency index corresponding to the matrix dimension of the matrix, and determining a consistency ratio based on the consistency index and the random consistency index; determining the consistency check result of the matrix based on the relationship between the consistency ratio and a preset rate value threshold.

[0012] According to another aspect of an embodiment of the present application, a person-job matching device is also provided, including: an acquisition module, used to acquire a target input text of a client, and analyze the target input text using a large language model to obtain a text label of the target input text, wherein the text label is used to mark whether the input text is a person-job matching text; a first construction module, used to analyze the person-job matching text using a large language model when the text label is a person-job matching text, to obtain at least one target position requirement information related to the target position and the degree of influence of at least one target capability item related to the target position requirement information on the target position requirement information, so as to construct an influence matrix; a second construction module, used to obtain employees to be recommended who have each target capability item and the target capability item scores of the employees to be recommended from a preset employee capability item set using a large language model, so as to construct a capability matrix; a matching module, used to analyze the influence matrix and the capability matrix using a large language model, determine the recommendation score of each employee to be recommended, and use the employees to be recommended whose recommendation scores are higher than a preset threshold as target recommended employees for the target position.

[0013] According to another aspect of an embodiment of the present application, a non-volatile storage medium is further provided, which includes a stored computer program, wherein the device where the non-volatile storage medium is located executes the above-mentioned person-job matching method by running the computer program.

[0014] According to another aspect of an embodiment of the present application, a computer program product is further provided, the computer program product comprising a stored computer program, wherein the computer program implements the above-mentioned person-job matching method when executed by a processor.

[0015] In the embodiment of the present application, the person-job matching system integrates an advanced artificial intelligence dialogue system, so that in the person-job matching process, the target input text entered by the user through the client can be processed by a large language model for natural language processing, solving the traditional person-job matching solution that requires a lot of human resources; at the same time, in the entire processing process, the person-job matching process is completed by a large language model to achieve automated and efficient matching. In addition, in the person-job matching stage, the employee's historical performance data and predicted future performance data are comprehensively considered to determine the ability score of the employee's various ability items, which can make the fit between the position and the employee higher, greatly improve the accuracy of person-job matching, and thus solve the technical problems of low matching efficiency and poor accuracy in manual person-job matching. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] The drawings described herein are used to provide a further understanding of the present application and constitute a part of the present application. The illustrative embodiments of the present application and their descriptions are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:

[0017] Figure 1 This is a hardware structure block diagram of an optional computer terminal (or mobile device) for implementing a person-job matching method according to an embodiment of the present application;

[0018] Figure 2 It is a flowchart of an optional person-job matching method according to an embodiment of the present application;

[0019] Figure 3 is a schematic structural diagram of an optional large model guide according to an embodiment of the present application;

[0020] Figure 4 is a schematic structural diagram of an optional person-job matching device according to an embodiment of the present application;

[0021] Figure 5 It is a schematic diagram of the structure of an optional electronic device according to an embodiment of the present application. DETAILED DESCRIPTION

[0022] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present application.

[0023] It should be noted that the terms "first", "second", etc. in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments of the present application described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device comprising a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0024] In addition, the relevant information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data for display, data for analysis, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties. For example, an interface is set up between this system and the relevant user or organization. Before obtaining relevant information, it is necessary to send an acquisition request to the aforementioned user or organization through the interface, and obtain relevant information after receiving the consent information fed back by the aforementioned user or organization.

[0025] In order to better understand the embodiments of the present application, the technical terms involved in the embodiments of the present application are explained as follows:

[0026] The Analytic Hierarchy Process (AHP) method is a multi-criteria decision-making method that decomposes complex problems into multiple components by establishing a hierarchical model, and performs pairwise comparison and consistency check between the factors to determine the relative importance of each factor, thereby solving the problem.

[0027] Example 1

[0028] According to an embodiment of the present application, a method embodiment of a person-job matching method is provided. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and although a logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0029] The method embodiments provided in the embodiments of the present application can be executed in a mobile terminal, a computer terminal or a similar computing device. Figure 1 FIG. 1 shows a hardware structure block diagram of a computer terminal (or mobile device) for implementing a person-job matching method. Figure 1 As shown, the computer terminal 10 (or mobile device 10) may include one or more (102a, 102b, ..., 102n are used to illustrate) processors 102 (the processor 102 may include but is not limited to a processing device such as a microprocessor MCU or a programmable logic device FPGA), a memory 104 for storing data, and a transmission device 106 for communication functions. In addition, it may also include: a display, an input / output interface (I / O interface), a universal serial bus (USB) port (which may be included as one of the ports of the BUS bus), a network interface, a power supply and / or a camera. It can be understood by those skilled in the art that Figure 1 The structure shown is only for illustration and does not limit the structure of the above electronic device. Figure 1 More or fewer components as shown, or with Figure 1 Different configurations are shown.

[0030] It should be noted that the one or more processors 102 and / or other data processing circuits described above may generally be referred to herein as "data processing circuits". The data processing circuits may be embodied in whole or in part as software, hardware, firmware, or any other combination thereof. In addition, the data processing circuit may be a single independent processing module, or may be incorporated in whole or in part into any of the other components in the computer terminal 10 (or mobile device). As described in the embodiments of the present application, the data processing circuit acts as a processor control (e.g., selection of a variable resistor terminal path connected to an interface).

[0031] The memory 104 can be used to store software programs and modules of application software, such as the program instructions / data storage device corresponding to the person-job matching method in the embodiment of the present application. The processor 102 executes various functional applications and data processing by running the software programs and modules stored in the memory 104, that is, the person-job matching method of the above-mentioned application is realized. The memory 104 may include a high-speed random access memory, and may also include a non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some examples, the memory 104 may further include a memory remotely arranged relative to the processor 102, and these remote memories may be connected to the computer terminal 10 via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0032] The transmission device 106 is used to receive or send data via a network. The specific example of the above network may include a wireless network provided by a communication provider of the computer terminal 10. In one example, the transmission device 106 includes a network adapter (Network Interface Controller, NIC), which can be connected to other network devices through a base station so as to communicate with the Internet. In one example, the transmission device 106 can be a radio frequency (RF) module, which is used to communicate with the Internet wirelessly.

[0033] The display may be, for example, a touch screen liquid crystal display (LCD) that enables a user to interact with a user interface of the computer terminal 10 (or mobile device).

[0034] In the above operating environment, Figure 2 is a flow chart of an optional person-job matching method according to an embodiment of the present application, such as Figure 2 As shown, the method at least includes steps S202-S208, wherein:

[0035] Step S202: acquiring the target input text of the client, and analyzing the target input text using the large language model to obtain a text label of the target input text.

[0036] In the technical solution provided in step S202, the person-job matching system first obtains the target input text input by the relevant personnel through the client, and then uses the large language model to classify the target input text to obtain the text label of the target input text, that is, to determine whether the target input text is a person-job matching text. If the target input text is a person-job matching text, such as the target input text is "matching an employee who is proficient in Windows operating system, Linux operating system, and familiar with VMware installation and debugging for the IT department", continue with the following steps; otherwise, the large language model will autonomously answer relevant content.

[0037] Step S204, when the text label is a person-job matching text, the person-job matching text is analyzed using a large language model to obtain at least one target job requirement information related to the target job and the degree of influence of at least one target capability item related to the target job requirement information on the target job requirement information, so as to construct an influence matrix.

[0038] In the technical solution provided in step S204, when the target input text is a person-job matching text, the person-job matching system continues to use the large language model to analyze the person-job matching text, obtains at least one target job requirement information related to the target job (which can be understood as the job requirement) and the degree of influence of at least one target capability item related to the target job requirement information on the target job requirement information (which can be understood as the importance of the employee's capability items to the job requirement), and constructs an influence matrix based on the target job requirement information and the degree of influence of the target capability item on the target job requirement information.

[0039] Step S206, using the large language model to obtain the recommended employees with each target competency item and the target competency item scores of the recommended employees from the preset employee competency item set, so as to construct a competency matrix.

[0040] In the technical solution provided in step S206, the person-job matching system uses a large language model to match the requirement information of each target position with a preset set of employee competency items, obtains the target competency item score with the highest matching degree and the corresponding employee to be recommended, and constructs a competency matrix based on this, wherein the employee competency item set includes competency items possessed by multiple employees and corresponding competency item scores, and the competency item scores are determined by the employee's first performance data in a historical time period and the second performance data in a predicted future time period.

[0041] Step S208, using the large language model to analyze the influence matrix and the capability matrix, determine the recommendation score of each employee to be recommended, and select the employees to be recommended whose recommendation score is higher than a preset threshold as target recommended employees for the target position.

[0042] In the technical solution provided in step S208, the person-job matching system inputs the influence matrix and the capability matrix into the large language model and analyzes them using the hierarchical analysis method to determine the recommendation score of each recommended employee in the capability matrix, and selects the recommended employees whose recommendation scores are higher than the preset threshold as the target recommended employees for the target position.

[0043] It should be noted that the above preset threshold can be determined according to the number of recommended employees or can be set according to the actual application scenario. For example, if two employees need to be recommended, the recommendation scores of the employees to be recommended can be arranged in descending order, and then the recommendation score greater than the third employee to be recommended and less than the second employee to be recommended is used as the preset threshold.

[0044] The above method of this embodiment is further introduced below.

[0045] As an optional implementation, in the technical solution provided in the above step S202, the method may include:

[0046] Step S2021, obtaining a sample data set for iterative training of a large language model, wherein the sample data set includes a plurality of groups of sample data consisting of sample texts and text labels corresponding to the sample texts;

[0047] Step S2022: Analyze the text similarity between the target input text and each sample text using the large language model, and use the text label of the target sample text whose text similarity is higher than a preset similarity threshold as the text label of the target input text.

[0048] Specifically, the above-mentioned large language model can be ChatGLM2 with natural language processing capabilities. Since ChatGLM2 pre-stores multiple conversation records (i.e., a sample data set composed of multiple sample data), and each conversation record includes a sample text and a text label corresponding to the sample text. Therefore, the embodiment of the present application can calculate the text similarity between the target input text and the sample text in each sample data, and use the text label of the target sample text whose text similarity with the target input text is higher than a preset threshold as the text label of the target input text. If the text label of the target input text cannot be determined by text similarity, the semantic similarity between the target input text and the sample text in each sample data can also be calculated, and the text label of the target sample text whose text similarity with the target input text is higher than a preset threshold is used as the text label of the target input text. Wherein, the above-mentioned similarity threshold is preferably set to 0.6.

[0049] It should be noted that, before guiding ChatGLM2 to perform natural language processing, the embodiment of the present application can be performed according to Figure 3The flowchart shown encapsulates the large model guide, and the large model guide mainly guides it by operating the history array (i.e., sample data set) of ChatGLM2. The history array stores the conversation records of the large language model (ChatGLM2) (i.e., multiple groups of sample data consisting of sample texts and text labels corresponding to the sample texts). ChatGLM2 completes the task in the context by reading the conversation records in the history array. In addition, the large model guide mainly includes: isolation layer, identity layer, sample layer, instruction layer, input layer, and output layer, among which:

[0050] The isolation layer is used to start before starting the task to clear the history array in ChatGLM2, thereby preventing the historical records generated by the previous task from disrupting the task execution of ChatGLM2.

[0051] The identity layer is used to write identity recognition guidance sentences into the history array. By reading the identity recognition guidance sentences in the history array, identity recognition can be generated, thereby improving the accuracy of task processing. The formula of the identity guidance sentence is as follows:

[0052]

[0053] in, represents the fixed part of the identity-introducing sentence, X i Indicates the new identity assigned to ChatGLM2.

[0054] The instruction layer is used to write task description statements in the history array, which mainly include the operations that ChatGLM2 needs to perform and the form of the answer output. Generally, ChatGLM2 will read the instructions in the history array and perform related operations according to the instructions. Among them, the task description statement generation formula is as follows:

[0055]

[0056] Where θ represents the fixed part of the task description sentence, X i Indicates the task description given to ChatGLM2. Indicates the processing result output.

[0057] The sample layer is used to write the answer sample of the task description statement in the history array. During the execution process, ChatGLM2 will imitate the sample to get the answer required by the user.

[0058]

[0059] Among them, Q irepresents the i-th input text, represents the fixed part of the identity-introducing sentence, X i represents the job proficiency requirement, Y i Indicates job requirements, A i Represents the output answer corresponding to the i-th input text.

[0060] The input layer is used to convert the original data written in the history array that the user needs to operate into a JSON string so that ChatGLM2 can read the original data and complete the operation.

[0061] The output layer is used to store the results after ChatGLM2 processing. Users can read the final desired results from it.

[0062] Furthermore, the JSON format result output by the above output layer is checked for non-emptiness, and only when all items in the JSON format result are not empty can the subsequent steps be continued. Otherwise, the client is prompted to continue inputting text.

[0063] As an optional implementation, in the technical solution provided in the above step S204, the method may include:

[0064] Step S2041 uses a large language model to analyze the person-job matching text and extract at least one target job requirement information related to the target job;

[0065] Step S2042, obtaining a job information database, wherein the job information database includes at least one capability item corresponding to a plurality of job requirement information related to the job and the influence degree of each capability item on the job requirement information;

[0066] Step S2043, matching each target job requirement information with the job information database to obtain at least one target capability item corresponding to each target job requirement information and the target impact degree of each target capability item on the corresponding target job requirement information;

[0067] Step S2044, constructing an influence matrix with target job requirement information as rows, target capability items as columns, and target impact degrees of target capability items on target job requirement information as matrix elements.

[0068] As an optional implementation, in the technical solution provided in the above step S206, the method may include:

[0069] Step S2061, obtaining an employee capability item set, wherein the employee capability item set includes a plurality of employee capability items and corresponding capability item scores, and the capability item scores are determined by the employee's first performance data in a historical time period and second performance data in a predicted future time period;

[0070] Step S2062, fuzzy matching each target capability item with the employee capability item set to obtain the recommended employees who have the target capability item and the target capability item scores of the recommended employees;

[0071] Step S2063, constructing a capability matrix with target capability items as rows, employees to be recommended as columns, and target capability item scores of employees to be recommended as matrix elements.

[0072] Specifically, since the related technology only evaluates the employee's ability score based on the employee's historical performance, but does not consider the employee's future performance trend, the employee's ability score is inaccurate. Therefore, in the embodiment of the present application, the employee's ability score is determined by the employee's first performance data in the historical time period and the second performance data in the predicted future time period.

[0073] The competency item scores of each competency item of each employee in the above employee competency item set can be obtained by the following steps, including:

[0074] Step S1, obtaining the first position data of the employee in a historical time period, and preprocessing the first position data.

[0075] The first position data includes at least one of the following: attendance data, salary data, and first performance data for characterizing competence items. The preprocessing operations include but are not limited to unit normalization, data cleaning, and the like.

[0076] Step S2, using a non-time-series linear non-Gaussian model to perform feature extraction on the preprocessed first position data to obtain at least one feature vector.

[0077] Specifically, in the above step S2, a non-time-series linear non-Gaussian model (such as a LiNGAM model) is used to extract features from the processed first position data to at least obtain a feature vector (that is, a feature vector representing a capability item) that has a causal relationship with the employee's business data.

[0078] Step S3, using the pre-trained prediction model to analyze each feature vector and predict the second performance data of the employee in the future time period.

[0079] The above prediction model is a deep learning model that combines convolutional neural network (CNN), long short-term memory network (LSTM) and self-attention mechanism (Attention). It should be noted that when using the prediction model to analyze the feature vector, it is necessary to increase the weight of the feature vector that has a causal relationship with the employee's business data to improve the accuracy of the prediction results.

[0080] Step S4, determining the employee's ability score based on the employee's first position data in a historical time period and the employee's second performance data in a future time period.

[0081] Specifically, in the technical solution provided in the above step S4, the weighted sum of each sub-item performance data (i.e., performance data on different projects) of the employee in the first performance data in the historical time period is performed using the following formula to obtain the employee's ability item score in the historical time dimension:

[0082]

[0083] in represents the employee's performance data of the i-th item in the historical period, ω i Indicates the weight of the sub-item performance data. The higher the weight, the more important the sub-item performance data is. max(α ago Indicates the maximum value of the employee's performance among n sub-items, min(α ago Indicates the minimum value of the employee's performance among n sub-items. It represents the normalized result of the performance of the i-th sub-item, thereby avoiding abnormal results caused by different performance measurement units.

[0084] Similarly, the weighted sum of each sub-item performance data (i.e., performance data on different projects) of the employee's second performance data in the future time period is performed using the following formula to obtain the employee's ability item score in the future time dimension:

[0085]

[0086] Then, the ability score of the above employees in the historical time dimension ε ago and the employee's ability score in the future time dimension ε pre Perform weighted summation to obtain the employee's competency score for this competency item:

[0087] ε=β1ε ago +β2ε pre

[0088] Among them, the weight β1 of the past time dimension and the weight β2 of the future time dimension are preferably set to 0.5 and 0.5. Of course, the embodiment of the present application does not impose specific restrictions on the setting of the weights, and they can be set according to the actual application scenario.

[0089] As an optional implementation, in the technical solution provided in the above step S208, the large language model can analyze the influence matrix and the capability matrix through a hierarchical comparative analysis method, and obtain the recommendation score of the employee through the following steps, including:

[0090] Step S2081, converting the capability matrix into a comparison matrix.

[0091] The comparison matrix uses the employees to be recommended as rows and columns, and the relative ability item scores of the employees to be recommended on the target ability items are used as matrix elements of the second comparison matrix.

[0092] Step S2082, respectively calculating the maximum eigenvalue and the corresponding eigenvector of the influence matrix and the contrast matrix, and determining the consistency check results of the influence matrix and the contrast matrix based on the maximum eigenvalue.

[0093] Specifically, in the technical solution provided in the above step S208, the embodiment of the present application can use the eigenvalue decomposition method to calculate the maximum eigenvalue and corresponding eigenvector of the influence matrix and the comparison matrix. Then, based on the maximum eigenvalue, the consistency check results of the influence matrix and the comparison matrix are determined by the following method:

[0094] First, according to the maximum eigenvalue of the matrix, the consistency index is determined by the following formula:

[0095]

[0096] Where n represents the matrix dimension of the contrast matrix, λ max Represents the largest eigenvalue of the contrast matrix.

[0097] Then, the random consistency index (RI) corresponding to the matrix dimension of the matrix is ​​determined, and based on the consistency index CI and the random consistency index RI, the consistency ratio is determined by the following formula:

[0098]

[0099] Finally, based on the relationship between the consistency ratio CR and the preset rate value threshold, the consistency check result of the matrix is ​​determined. Among them, the above rate value threshold is preferably set to 0.1. Of course, the threshold can also be set according to the actual application scenario. Therefore, when CR is less than the rate value threshold, it means that the matrix has satisfactory consistency, otherwise it is necessary to readjust the pairwise comparison values ​​in the comparison matrix.

[0100] Step S2083, when the consistency check results of the influence matrix and the comparison matrix are both verified to be passed, the recommendation score of each employee to be recommended is obtained based on the sum of the product of the eigenvector of the influence matrix and the column vector in the capability matrix.

[0101] Specifically, let the influence matrix be A(m*n), the competence matrix be B(n*l), the target position requirement information be m, the number of employees to be recommended be l, and the number of competence items be n. Then, the dimension of the eigenvector w of the influence matrix A (which can be expressed as the weight of each target competence item in the influence matrix after normalization) is n*1. Therefore, the competence item scores of each employee to be recommended in the competence matrix can be written as:

[0102]

[0103] Among them, w i represents the weight of the i-th target capability item in the impact matrix, b ij It represents the competency score of the j-th employee on the i-th target competency item.

[0104] It should be noted that the large language model used in the embodiment of the present application can be loaded into the memory. For example, the raw data of the large language model can be loaded from the non-volatile memory to the volatile memory so that the processor can run the large language model. The raw data of the large language model refers to unprocessed data, which usually includes the parameters and structural data of the large language model. The structural data can be a calculation relationship based on the parameters, such as the forward propagation calculation relationship between the intermediate layers and the neurons. Specifically, the structural data can include the structure-related code of the large language model, such as the code for performing related calculations between the intermediate layers and the neurons.

[0105] In one embodiment, an area for loading a large language model can be divided in the memory, which may include a structure data storage area and a parameter storage area. The structure data storage area is used to store structure-related codes, and the parameters referenced by it can point to the address of a specific parameter in the parameter storage area through a pointer. During the training process of the large language model, it may be necessary to frequently update the parameters, and the parameter values ​​in the parameter storage area can be updated.

[0106] Based on the scheme defined in the above steps S202 to S208, it can be learned that in the embodiment, the person-job matching system integrates an advanced artificial intelligence dialogue system, so that in the person-job matching process, the target input text entered by the user through the client can be processed by a large language model for natural language processing, solving the problem that the traditional person-job matching solution requires a lot of human resources; at the same time, in the entire processing process, the person-job matching process is completed by a large language model to achieve automated and efficient matching. In addition, in the person-job matching stage, the employee's historical performance data and predicted future performance data are comprehensively considered to determine the ability score of the employee's various ability items, which can make the fit between the position and the employee higher, greatly improve the accuracy of person-job matching, and thus solve the technical problem of low matching efficiency and poor accuracy in manual person-job matching.

[0107] Example 2

[0108] Based on Embodiment 1 of the present application, an embodiment of a person-job matching device is also provided, and the person-job matching method of the above embodiment is executed when the device is running. Figure 4 is a schematic structural diagram of an optional person-job matching device according to an embodiment of the present application, such as Figure 4 As shown, the person-job matching device at least includes an acquisition module 41, a first construction module 42, a second construction module 43 and a matching module 44, wherein:

[0109] The acquisition module 41 is used to acquire the target input text of the client, and analyze the target input text using the large language model to obtain a text label of the target input text, wherein the text label is used to mark whether the input text is a human-position matching text;

[0110] The first construction module 42 is used to analyze the person-job matching text using a large language model when the text label is a person-job matching text, obtain at least one target job requirement information related to the target job and the influence degree of at least one target capability item related to the target job requirement information on the target job requirement information, so as to construct an influence matrix;

[0111] The second construction module 43 is used to obtain the recommended employees with each target competency item and the target competency item scores of the recommended employees from the preset employee competency item set by using the large language model to construct a competency matrix;

[0112] The matching module 44 is used to analyze the influence matrix and the capability matrix using a large language model, determine the recommendation score of each recommended employee, and select the recommended employees whose recommendation scores are higher than a preset threshold as target recommended employees for target positions.

[0113] It should be noted that the various modules in the above-mentioned person-job matching device can be program modules (for example, a set of program instructions that implement a certain specific function) or hardware modules. For the latter, it can be expressed in the following forms, but is not limited to this: the expression form of each of the above-mentioned modules is a processor, or the functions of each of the above-mentioned modules are implemented by a processor.

[0114] Example 3

[0115] According to an embodiment of the present application, a non-volatile storage medium is further provided, in which a program is stored, wherein when the program is running, the device where the non-volatile storage medium is located is controlled to execute the person-job matching method in Example 1.

[0116] Optionally, the device where the non-volatile storage medium is located implements the following steps by running the program: obtaining the target input text of the client, and analyzing the target input text using a large language model to obtain a text label of the target input text, wherein the text label is used to mark whether the input text is a person-job matching text; in the case where the text label is a person-job matching text, analyzing the person-job matching text using a large language model to obtain at least one target job requirement information related to the target job and the degree of influence of at least one target capability item related to the target job requirement information on the target job requirement information, so as to construct an influence matrix; obtaining the recommended employees with each target capability item and the target capability item scores of the recommended employees from a preset employee capability item set using a large language model to construct a capability matrix; analyzing the influence matrix and the capability matrix using a large language model to determine the recommendation score of each employee to be recommended, and selecting the recommended employees with a recommendation score higher than a preset threshold as the target recommended employees for the target job.

[0117] According to an embodiment of the present application, a computer program product is also provided, which includes a stored computer program, wherein when the computer program is executed by a processor, the person-job matching method in Example 1 is implemented.

[0118] Optionally, the computer program executes the following steps: obtaining a target input text from the client, and analyzing the target input text using a large language model to obtain a text label of the target input text, wherein the text label is used to mark whether the input text is a person-job matching text; in the case where the text label is a person-job matching text, analyzing the person-job matching text using a large language model to obtain at least one target job requirement information related to the target job and the degree of influence of at least one target capability item related to the target job requirement information on the target job requirement information, so as to construct an influence matrix; obtaining employees to be recommended who have each target capability item and the target capability item scores of the employees to be recommended from a preset employee capability item set using a large language model, so as to construct a capability matrix; analyzing the influence matrix and the capability matrix using a large language model to determine the recommendation score of each employee to be recommended, and selecting employees to be recommended whose recommendation score is higher than a preset threshold as target recommended employees for the target job.

[0119] According to an embodiment of the present application, a processor is also provided, which is used to run a program, wherein the person-job matching method in Example 1 is executed when the program is running.

[0120] Optionally, when the program is running, the following steps are executed: obtaining a target input text from the client, and analyzing the target input text using a large language model to obtain a text label of the target input text, wherein the text label is used to mark whether the input text is a person-job matching text; in the case where the text label is a person-job matching text, analyzing the person-job matching text using a large language model to obtain at least one target position requirement information related to the target position and the degree of influence of at least one target capability item related to the target position requirement information on the target position requirement information, so as to construct an influence matrix; obtaining recommended employees with each target capability item and the target capability item scores of the recommended employees from a preset employee capability item set using a large language model to construct a capability matrix; analyzing the influence matrix and the capability matrix using a large language model to determine the recommendation score of each employee to be recommended, and selecting the recommended employees with a recommendation score higher than a preset threshold as target recommended employees for the target position.

[0121] According to an embodiment of the present application, an electronic device is also provided, wherein: Figure 5 is a schematic diagram of the structure of an optional electronic device according to an embodiment of the present application, such as Figure 5 As shown, the electronic device includes one or more processors; a memory for storing one or more programs, which, when the one or more programs are executed by the one or more processors, enables the one or more processors to run the programs, wherein the programs are configured to execute the person-job matching method in the above-mentioned embodiment 1 when running.

[0122] Optionally, the processor is configured to execute the following steps through a computer program: obtaining a target input text of the client, and analyzing the target input text using a large language model to obtain a text label of the target input text, wherein the text label is used to mark whether the input text is a person-job matching text; in the case where the text label is a person-job matching text, analyzing the person-job matching text using a large language model to obtain at least one target job requirement information related to the target job and the degree of influence of at least one target capability item related to the target job requirement information on the target job requirement information, so as to construct an influence matrix; obtaining employees to be recommended who have each target capability item and the target capability item scores of the employees to be recommended from a preset employee capability item set using a large language model, so as to construct a capability matrix; analyzing the influence matrix and the capability matrix using a large language model to determine the recommendation score of each employee to be recommended, and selecting employees to be recommended whose recommendation score is higher than a preset threshold as target recommended employees for the target job.

[0123] The serial numbers of the above-mentioned embodiments of the present application are for description only and do not represent the advantages or disadvantages of the embodiments.

[0124] In the above embodiments of the present application, the description of each embodiment has its own emphasis. For parts that are not described in detail in a certain embodiment, please refer to the relevant description of other embodiments.

[0125] In the several embodiments provided in this application, it should be understood that the disclosed technical content can be implemented in other ways. Among them, the device embodiments described above are only schematic. For example, the division of units can be a logical function division. There may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of units or modules, which can be electrical or other forms.

[0126] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed over multiple units. Some or all of the units may be selected according to actual needs to achieve the purpose of the present embodiment.

[0127] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0128] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application, or the part that contributes to the relevant technology or all or part of the technical solution, can be embodied in the form of a software product, which is stored in a storage medium and includes several instructions for a computer device (which can be a personal computer, server or network device, etc.) to perform all or part of the steps of each embodiment method of the present application. The aforementioned storage medium includes: U disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), mobile hard disk, disk or optical disk, etc. Various media that can store program codes.

[0129] The above are only preferred implementations of the present application. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present application. These improvements and modifications should also be regarded as the scope of protection of the present application.

Claims

1. A person-job matching method, characterized in that: include: Obtain the target input text of the client, and analyze the target input text using the large language model to obtain a text label of the target input text, wherein the text label is used to mark whether the input text is a human-position matching text; In the case where the text label is a person-job matching text, the person-job matching text is analyzed using the large language model to obtain at least one target job requirement information related to the target job and the influence degree of at least one target capability item related to the target job requirement information on the target job requirement information, so as to construct an influence matrix; Using the large language model, obtaining the recommended employees with each of the target competence items and the target competence item scores of the recommended employees from a preset set of employee competence items, so as to construct a competence matrix; The influence matrix and the capability matrix are analyzed by using the large language model to determine the recommendation score of each employee to be recommended, and the employees to be recommended whose recommendation score is higher than a preset threshold are used as target recommended employees for the target position.

2. The method according to claim 1, characterized in that: Analyzing the target input text using a large language model to obtain a text label of the target input text includes: Acquire a sample data set for iteratively training the large language model, wherein the sample data set includes multiple groups of sample data consisting of sample texts and text labels corresponding to the sample texts; The large language model is used to analyze the text similarity between the target input text and each sample text, and the text label of the target sample text whose text similarity is higher than a preset similarity threshold is used as the text label of the target input text.

3. The method according to claim 1, characterized in that The person-job matching text is analyzed by using the large language model to obtain at least one target job requirement information related to the target job and the influence degree of at least one target capability item related to the target job requirement information on the target job requirement information, so as to construct an influence matrix, including: Analyzing the person-job matching text using the large language model to extract at least one target job requirement information related to the target job; Acquire the position information database, wherein the position information database includes at least one capability item corresponding to a plurality of position requirement information related to the position and the influence degree of each capability item on the position requirement information; Matching each of the target job requirement information with the job information database to obtain at least one of the target capability items corresponding to each of the target job requirement information and the target impact degree of each of the target capability items on the corresponding target job requirement information; The influence matrix is ​​constructed by taking target job requirement information as rows, target capability items as columns, and the target influence degree of the target capability items on the target job requirement information as matrix elements.

4. The method according to claim 1, characterized in that: The large language model is used to obtain the recommended employees with each target competency item and the target competency item scores of the recommended employees from the preset employee competency item set to construct a competency matrix, including: Acquire the employee capability item set, wherein the employee capability item set includes a plurality of capability items possessed by the employees and corresponding capability item scores, and the capability item scores are determined by first performance data of the employees in a historical time period and second performance data in a predicted future time period; Perform fuzzy matching on each of the target competency items and the employee competency item set to obtain the recommended employees who have the target competency items and the target competency item scores of the recommended employees; The capability matrix is ​​constructed by taking target capability items as rows, employees to be recommended as columns, and target capability item scores of the employees to be recommended as matrix elements.

5. The method according to claim 4, characterized in that The process of determining the employee's competency score includes: Acquire the first position data of the employee in a historical time period, and pre-process the first position data, wherein the first position data includes at least one of the following: attendance data, salary data, and first performance data for characterizing a capability item; Using a non-time-series linear non-Gaussian model to extract features from the preprocessed first job data to obtain at least one feature vector, wherein the type of the feature vector at least includes: a feature vector for characterizing a capability item; Analyzing each of the feature vectors using a pre-trained prediction model to predict the second performance data of the employee in a future time period, wherein the prediction model is a deep learning model that combines a convolutional neural network, a long short-term memory network, and a self-attention mechanism; The ability item score of the employee is determined based on the first position data of the employee in a historical time period and the second performance data of the employee in a future time period.

6. The method according to claim 1, characterized in that The influence matrix and the capability matrix are analyzed by using the large language model to determine the recommendation score of each employee to be recommended, including: Converting the competence matrix into a comparison matrix, wherein the comparison matrix has the employees to be recommended as rows and columns, and the relative competence item scores of the employees to be recommended on the target competence items are used as matrix elements of the second comparison matrix; Calculating the maximum eigenvalue and the corresponding eigenvector of the influence matrix and the comparison matrix respectively, and determining the consistency check results of the influence matrix and the comparison matrix respectively based on the maximum eigenvalue; When the consistency check results of the influence matrix and the comparison matrix are both verified to be passed, the recommendation score of each of the employees to be recommended is obtained based on the sum of the products of the eigenvector of the influence matrix and the column vectors in the capability matrix.

7. The method according to claim 6, characterized in that Determining consistency check results of the influence matrix and the comparison matrix based on the maximum eigenvalue, respectively, includes: For any one of the influence matrix and the comparison matrix, determining a consistency index based on the maximum eigenvalue of the matrix; determining a random consistency index corresponding to a matrix dimension of the matrix, and determining a consistency ratio based on the consistency index and the random consistency index; Based on the relationship between the consistency ratio and a preset rate value threshold, a consistency check result of the matrix is ​​determined.

8. A person-job matching device, characterized in that: include: An acquisition module is used to acquire a target input text from a client, and analyze the target input text using a large language model to obtain a text label of the target input text, wherein the text label is used to mark whether the input text is a human-position matching text; A first construction module is used for analyzing the person-job matching text using the large language model when the text label is a person-job matching text, obtaining at least one target job requirement information related to the target job and the influence degree of at least one target capability item related to the target job requirement information on the target job requirement information, so as to construct an influence matrix; A second construction module is used to obtain the recommended employees with each of the target competence items and the target competence item scores of the recommended employees from a preset employee competence item set by using the large language model to construct a competence matrix; A matching module is used to analyze the influence matrix and the capability matrix using the large language model to determine the recommendation score of each recommended employee, and to select the recommended employees whose recommendation scores are higher than a preset threshold as target recommended employees for the target position.

9. A non-volatile storage medium, characterized in that: The non-volatile storage medium stores a computer program, wherein the device where the non-volatile storage medium is located executes the person-job matching method described in any one of claims 1 to 7 by running the computer program.

10. A computer program product, characterized in that include: A computer program, wherein when the computer program is executed by a processor, the person-job matching method described in any one of claims 1 to 7 is implemented.

Citation Information

Cited By

  • Human resource management method based on generative adversarial network and large language model

    CN119963144A

  • Matching recommendation method and system based on large model digital talent inventory

    CN120561373A