Automatic matching method for post transfer in enterprise

Abstract

The invention belongs to the technical field of artificial intelligence, and particularly relates to an automatic matching method for post transfer in an enterprise. In order to solve the problem that in the prior art, enterprise employee information integration is not comprehensive, and due to the influence of subjective factors of human resource managers, the matching degree of post-transfer posts and original posts is low, the method comprises the following steps of: according to a key element data set of a current post of an employee to be subjected to post transfer and a key element data set of a corresponding gap post; and calculating a matching degree result of each key element of the current post of each to-be-transferred employee, integrating the calculation results to obtain a total matching degree result of the current post of the to-be-transferred employee and the notched post, and determining the to-be-transferred employee meeting the notched post demand according to the total matching degree result. The effects of reducing the subjective factor influence of the human resource management personnel, shortening the time for the employees to adapt to new posts after the employees are transferred and breaking through the scene limitation of a traditional post recommendation method are achieved.

Description

Technical Field

[0001] The present invention belongs to the technical field of artificial intelligence, and particularly relates to an automatic matching method for internal job transfers in enterprises. Background Art

[0002] In the current human resource management of enterprises, due to the influence of various factors such as the existing system and enterprise management system, there are problems of unreasonable human resource allocation in enterprises. With the adjustment of enterprise business and the improvement of informatization level, the manpower demand for certain positions has been greatly reduced and new position demands have emerged, seriously restricting the development of enterprises.

[0003] When matching positions in the prior art, it mainly focuses on the external recruitment scenario of enterprises, that is, recommending relevant positions according to the personal information and job hunting intentions of job seekers. There are obvious differences between the internal job transfer scenario of enterprises and the external recruitment scenario of enterprises: Difference one, job seekers in external recruitment will actively provide personal information and job hunting intentions, while internal employees of enterprises are passively transferred due to enterprise development, and the suitable positions are unclear; Difference two, in the external recruitment scenario of enterprises, the goal is to match the most suitable candidates with specific positions, while in the internal job transfer scenario of enterprises, the goal is to meet the business changes within the organization and match the existing employees with the gap positions suitable for them; Difference three, in the external recruitment scenario of enterprises, it is difficult to obtain accurate information of job seekers, especially the work content of the original position, so the matching degree between the intended job hunting position and the recommended position is low, and accurate employee information and the work content of the employee's current position can be obtained within the enterprise.

[0004] For example, the Chinese invention patent application with the application publication number CN112199602A discloses a position recommendation method for the external recruitment scenario. This patent is to obtain the comparison between the user's job hunting intention information and candidate position information, and push the candidate positions corresponding to the total similarity score that meets the preset conditions to the user according to the total similarity score between the candidate position and the intended position. However, this method is limited to the external recruitment scenario, and the prerequisite for application is that the job hunting intention information of the user must be obtained, which is not applicable to the internal job transfer scenario of enterprises.

[0005] The existing human resource information system is not comprehensive enough in integrating enterprise employee information. In the face of a large number of enterprise employees and the lack of quantifiable data to support decision-making, human resource managers will consider whether the basic information of employees meets the basic requirements of the gap positions during the process of adjusting employees' positions. Judging whether employees are competent for the gap positions depends more on the experience of managers and is easily affected by subjective factors, resulting in a low matching degree between the positions of employees after job transfer and their original positions. Summary of the Invention

[0006] The object of the present invention is to provide an automatic matching method for internal job transfers in enterprises, so as to solve the problems in the prior art that the integration of enterprise employee information is not comprehensive and affected by the subjective factors of human resource managers, resulting in a low matching degree between the post after the job transfer and the original post.

[0007] To solve the above technical problems, the present invention provides an automatic matching method for internal job transfers in enterprises, including the following steps:

[0008] 1) Preprocess the constructed human resource data set to obtain the key element feature data set of the current post of the employee to be transferred, and the corresponding key element feature data set of the vacant post;

[0009] 2) Use the corresponding matching mechanism to calculate the matching degree of each key element feature corresponding to the current post and the vacant post of the employee to be transferred, obtain the matching degree results of each key element feature corresponding to the current post and the vacant post of the employee to be transferred, and process the matching degree results of each key element feature to obtain the total matching degree result of the current post and the vacant post of the employee to be transferred;

[0010] 3) Determine whether the employee to be transferred meets the requirements of the vacant post according to the total matching degree result of the current post and the vacant post of the employee to be transferred.

[0011] Further, the key element features include: post name, salary, professional skill certificate, age, gender, education background, and job content.

[0012] Further, the matching mechanism includes a direct text matching mechanism, a calculation matching mechanism, and a text similarity matching mechanism; among them, the direct text matching mechanism refers to directly comparing whether a certain key element feature corresponding to the current post and the vacant post is the same; the calculation matching mechanism refers to comparing whether a certain key element feature corresponding to the current post and the vacant post satisfies a magnitude relationship; the text similarity matching mechanism refers to calculating the cosine similarity between a certain key element feature corresponding to the current post and the vacant post.

[0013] Further, the matching degree result of each key element feature is a matching degree score. The key element features using the direct text matching mechanism for matching include gender and professional skill certificate, and its calculation formula is:

[0014]

[0015]

[0016] where s 1 is the gender matching degree score of the employee's current post and the vacant post; L 1 is the gender of the employee; N 1Gender requirements for the vacant position; s 2 The matching score of the employee's current position and the vacant position in terms of professional skill certificates; L 2 The professional skill certificates the employee already has; N 2 The requirements for professional skill certificates of the vacant position.

[0017] Furthermore, the matching results of each key element feature are matching scores. The key elements matched using the calculation matching mechanism include salary, age, and education level, and their calculation formula is:

[0018]

[0019]

[0020]

[0021] Among them, s 31 The age matching score of the employee's current position and the vacant position; L 3 The age of the employee; N 3 The age requirement of the vacant position; s 32 The salary matching score of the employee's current position and the vacant position; L 4 The salary of the employee's current position; N 4 The salary of the vacant position; s 33 The education level matching score of the employee's current position and the vacant position; L 5 The education level of the employee; N 5 The education level requirement of the vacant position.

[0022] Furthermore, the matching results of each key element feature are matching scores. The key elements matched using the text similarity matching mechanism include the position name and job content, and their calculation formula is:

[0023] s 4 (N 6 , L 6 ) = cos_sim(N 6 h jt , L 6 h it )

[0024] s 5 (N 7 , L 7 ) = cos_sim(N 7 h jt , L 7 h it )

[0025] Among them, L 6It is the job title of the employee's current position; N 6 It is the job title of the vacant position; L 6 h it It is the node after t - layer propagation of the name of the i - th position in the employee's current position; N 6 h jt It is the node after t - layer propagation of the name of the j - th position in the vacant position; s 4 It is the matching score of the job titles between the employee's current position and the vacant position; cos_sim is the cosine similarity calculation function; L 7 It is the job content of the employee's current position; L 7 h it It is the node after t - layer propagation of the i - th sentence in the job content of the employee's current position; N 7 It is the job content of the vacant position; N 7 h jt It is the node after t - layer propagation of the j - th sentence in the job content of the vacant position; s 5 It is the matching score of the job contents between the employee's current position and the vacant position.

[0026] Furthermore, the specific steps for processing the matching results of each key element feature are as follows: First, integrate the matching results of salary, age, and education level to obtain the first comprehensive matching result. Then, integrate the first comprehensive matching result with the matching results of gender and professional skill certificates to obtain the second comprehensive matching result. Finally, perform a weighted sum of the second comprehensive matching result and the matching results of job title and job content to obtain the total matching result of the current position of the employee to be transferred. The calculation formula is as follows:

[0027]

[0028]

[0029] s total =k 1 s p +k 2 s 4 +k 3 s 5

[0030] Among them, s 31 is the age matching result between the employee's current position and the vacant position; s 32 is the salary matching result between the employee's current position and the vacant position; s 33 is the education level matching result between the employee's current position and the vacant position; s 3 is the first comprehensive matching result; s 1 is the gender matching result between the employee's current position and the vacant position; s2 is the matching degree result of the professional skill certificates for the employee's current position and the vacant position; s p is the second comprehensive matching degree result; s 4 is the matching degree result of the position names for the employee's current position and the vacant position; s 5 is the matching degree result of the job contents for the employee's current position and the vacant position; s total is the total matching degree result of the employee's current position and the vacant position for the employee to be transferred; k 1 、k 2 and k 3 are respectively the weight of the second comprehensive matching degree result, the weight of the matching degree result of the position names for the employee's current position and the vacant position, and the weight of the matching degree result of the job contents for the employee's current position and the vacant position.

[0031] Furthermore, in the process of calculating the matching degree score of the education background for the employee's current position and the vacant position, each education background needs to be first converted into corresponding scores and then calculated using the scores.

[0032] Its beneficial effects are as follows: The present invention is an exploratory invention creation. To solve the problems in the prior art that the integration of enterprise employee information is not comprehensive and is affected by the subjective factors of human resource managers, resulting in a low matching degree between the transferred position and the original position, the present invention automatically determines whether the employee to be transferred meets the requirements of the vacant position according to the matching degree of the key element features corresponding to the employee to be transferred and the vacant position. The specific process is as follows: According to the key element data sets of the employee's current position and the corresponding vacant position, calculate the matching degree results of each key element of the employee's current position, and integrate the calculation results to obtain the total matching degree result of the employee's current position and the vacant position, and determine the employees to be transferred who meet the requirements of the vacant position according to the total matching degree result. The present invention improves the matching degree of position recommendation, achieves the effect of reducing the influence of the subjective factors of human resource managers, shortening the time for employees to adapt to the new position after transfer, and breaking the scenario limitations of traditional position recommendation methods. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] Figure 1 is the flowchart of an embodiment of the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0034] The basic concept of the present invention is as follows: By preprocessing the obtained data set, the present invention obtains the key element data set of the current position of the employee to be transferred and the key element data set of the corresponding vacant position, calculates the matching degree results of each key element of the current position of each employee to be transferred, and integrates the calculation results to obtain the total matching degree result of the current position of the employee to be transferred and the vacant position. According to the total matching degree result, it is judged whether the employee to be transferred meets the requirements of the vacant position. The principle of the present invention is: Preprocess the constructed human resource data set to obtain the key element feature data set of the current position of the employee and the key element feature data set of the corresponding vacant position, match various key element features to obtain the matching degree results of each key element feature of the current position of the employee to be transferred, and calculate the total matching degree result of the current position of the employee to be transferred. According to the total matching degree result, it is judged whether the employee to be transferred meets the requirements of the vacant position. Based on this concept, an automatic matching method for internal job transfer in an enterprise of the present invention can be realized.

[0035] The present invention will be described in detail below in conjunction with the accompanying drawings and method embodiments.

[0036] Method embodiment:

[0037] An automatic matching method for internal job transfer in an enterprise of the present invention, the flowchart is as Figure 1 shown. Taking 2,000 employees of an enterprise as the employees to be transferred (excluding backend software development engineers) and the vacant position as the backend software development engineer as an example, the specific steps include:

[0038] Step 1, obtain the data of the human resource database within the enterprise and construct a human resource data set.

[0039] The data set includes the "name", "department", "position name", "salary", "existing professional skill certificates", "age", "gender", "education background" and "work content" of the employees. In this embodiment, the above 9 types of data are selected, but it is not limited to these data and can also be adjusted according to actual needs.

[0040] Step 2, preprocess the constructed human resource data set, and screen the corresponding data from the human resource data set to construct the key element feature data set of the current position of the employee and the key element feature data set of the vacant position.

[0041] 1) After preprocessing the human resource data set, select the "position name", "salary", "existing professional skill certificates", "age", "gender", "education background" and "work content" of the current positions of all employees within the enterprise as the 7 key element features of the current position of the employee, and establish the key element feature data set of the current position of the employee, as shown in Table 1 below:

[0042] Table 1

[0043]

[0044]

[0045]

[0046]

[0047] 2) After preprocessing the human resource data set, select the "job title", "salary", "occupational skill certificate requirements", "age requirements", "gender requirements", "educational requirements", and "job content" of the internal gap positions in the enterprise as the seven key element features of the gap positions, and establish a data set of key element features of the gap positions, as shown in Table 2 below:

[0048] Table 2

[0049]

[0050] Step 3: Classify the seven key element features of the employee's current position and the gap position respectively, and use different matching mechanisms for matching according to the characteristics of the elements to calculate the matching degree between each key element of the employee's current position and the corresponding key element of the gap position.

[0051] 1) The "gender requirement N 1 " and "occupational skill certificate requirement N 2 " element features of the gap position and the employee's current "gender L 1 " and "existing occupational skill certificates L 2 " are corresponding element features respectively, and their characteristic values are fixed words, and the direct text matching mechanism is applicable for matching. The calculation formula is as follows:

[0052]

[0053]

[0054] Among them, if there is an intersection between the employee's current "gender L 1 " and the "gender requirement N 1 " element feature of the gap position, the matching degree s 1 is 1, otherwise it is 0; if there is an intersection between the employee's current "existing occupational skill certificates L 2 " and the "occupational skill certificate requirement N 2 " element feature of the gap position, the matching degree s 2 is 1, otherwise it is 0.

[0055] 2) The "age requirement N 3 " and "salary N4 ” and “Educational requirement N 5 ” element feature and the employee's current “Age L 3 ”, “Salary L 4 ” and “Educational level L 5 ” are corresponding element features respectively. There are magnitude relationships among these element features during the matching process, and a calculation matching mechanism is applied for matching.

[0056] That is, the “Age requirement N 3 ” of the vacant position should be greater than the employee's current “Age L 3 ”; the “Salary N 4 ” of the vacant position should be greater than or equal to the employee's current “Salary L 4 ”; the “Educational requirement N 5 ” of the vacant position should be lower than or equal to the employee's current “Educational level L 5 ”. For example, the “Educational requirement” of the vacant position is “Bachelor's degree”, while the employee's current “Educational level” is “Master's degree”. However, the magnitude relationship cannot be directly reflected. Therefore, the feature values of “Educational level L 5 ” and “Educational requirement N 5 ” need to be converted into numbers to apply the calculation matching mechanism for matching. “High school or below”, “Technical secondary school”, “College diploma”, “Bachelor's degree”, “Master's degree” and “Doctoral degree” are assigned values of “1”, “2”, “3”, “4”, “5” and “6” respectively. The processed data is shown in Table 3 and Table 4 below:

[0057] Table 3

[0058]

[0059]

[0060]

[0061]

[0062] Table 4

[0063]

[0064] Therefore, the calculation formula of the calculation matching mechanism is:

[0065]

[0066]

[0067]

[0068]

[0069] Among them, if the "age L 3 " element feature of the employee is less than the "age requirement N 3 " element feature of the vacant position, the matching degree s 31 is 1, otherwise it is 0; if the "salary L 4 " element feature of the employee is less than or equal to the "salary N 4 " element feature of the vacant position, the matching degree s 32 is 1, otherwise it is 0; if the "education level L 5 " element feature of the employee is greater than or equal to the "education level requirement N 5 " element feature of the vacant position, the matching degree s 33 is 1, otherwise it is 0; finally, if s 31 、s 32 and s 33 are all 1, then the matching degree s 3 (the first comprehensive matching degree result) is 1, otherwise it is 0.

[0070] 3) The "position name N 6 " and "work content N7" element features of the vacant position and the "position name L 6 " and "work content L 7 " of the employee's current position are corresponding element features respectively. "Position name" and "work content" are key reference element features for adjusting positions, which largely reflect the similarity between the two positions. However, it cannot be based on the previous two matching mechanisms and needs to use the text similarity matching mechanism to calculate the matching degree through natural language processing methods;

[0071] Among them, the unsupervised similarity calculation of the directed graph matrix neural network algorithm (GCN) is used in this embodiment. By using the jieba Chinese word segmentation tool to perform word segmentation operations on a large-scale corpus; then training to obtain word vectors to obtain the word vectors corresponding to each word, and finally summing the word vectors of all words in the text to obtain the sentence vector of the text; obtaining the similarity value between the texts by measuring the distance between the sentence vectors of the two texts.

[0072] The processing methods of the "position name" and "work content" element features are the same. Taking the "position name" element feature as an example, compare the "position name N 6 " of the vacant position with the "position name L 6The degree of similarity. Among them, the "job title" is not composed of a single word. Generally, it is composed of multiple minimum unit words. For example, "software development engineer" is composed of 3 minimum unit words: "software", "development", and "engineer". Through Distributed representation, each minimum unit word is mapped to a word vector, and all word vectors form a vector space. Next, statistical methods can be used to study the relationships between words. The specific operation steps are as follows:

[0073] ① Establish a word vector model and a sentence vector model.

[0074] Collect the job title catalogs within the enterprise, job title catalogs in the same industry, job title catalogs on major recruitment websites, Baidu Encyclopedia, and Chinese Wikipedia, etc. as the training data set for "job titles", and then combine with the Word2Vec algorithm for training to obtain a 128-dimensional word vector model;

[0075] Divide the "job title" into multiple minimum unit words, and query whether the established word vector model contains all the divided minimum unit words. If so, take the word vectors corresponding to each minimum unit word in the word vector model as the word vector of this "job title"; if not, set the word vectors of each minimum unit word to 0;

[0076] Add up the word vectors of each minimum unit word, then divide the sum of the obtained word vectors by the number of minimum unit words to get the average value of the word vectors, and use this average value as the sentence vector of the job title.

[0077] ② Use the GCN algorithm to determine the similarity of the sentence vectors of "Job Title L" of the employee's current position 6 " and "Job Title N" of the gap position 6 ".

[0078] First, construct the node representation of the graph. Suppose there are n positions, and each position is represented as a node. Use the sentence vectors calculated above to represent the characteristics of the nodes, denoted as H = |h 1 ,h 2 ,…,h i ,…,h n , h i where represents the feature vector of the i-th position;

[0079] Then use the point mutual information method (PMI) to calculate the weights of the edges. Suppose the edge weight matrix is A = [a ij , where a ij represents the weight of the edge from node i to node j;

[0080] Then, based on the weights of the nodes and edges of the graph, and through a multi-layer neural network model for graph propagation, an updated representation of the nodes is obtained. Assume that the node representation after t layers of propagation is H t =[h 1t ,h 2t ,…,h it ,…,h nt ;The node of "Job Name L 6 " for the employee's current position is L 6 H t =[L 6 h 1t ,L 6 h 2t ,…,L 6 h it ,…,L 6 h nt ,The node of "Job Name N 6 " for the vacant position is N 6 H t =[N 6 h 1t ,N 6 h 2t ,…,N 6 h jt ,…,N 6 h mt ;Among them, according to experimental experience, setting the number of layers to 7 layers has the best effect. If the number of layers exceeds 7 layers, there is a risk of overfitting;

[0081] Finally, determine the cosine similarity s 6 between the sentence vector of "Job Name L 6 " for the employee's current position and the sentence vector of "Job Name N 4 " for the vacant position, and use it as the matching degree of "Job Name". The specific calculation formula is as follows:

[0082] s 4 (N 6 ,L 6 )=cos_sim(N 6 h jt ,L 6 h it )

[0083] Among them, cos_sim is the cosine similarity calculation function; N 6 h jt is the node representation of the job name of the j-th job in the vacant position after t layers of propagation; L 6 h it is the node representation of the job name of the i-th job in the employee's current position after t layers of propagation.

[0084] 4) According to the above step 3, similarly, the matching degree s of "job content" can be obtained. 5 The calculation formula is as follows:

[0085] s 5 (N 7 , L 7 ) = cos_sim(N 7 h jt , L 7 h it )

[0086] where cos_sim is the cosine similarity calculation function; N 7 h jt is the node representation of the i-th sentence of the "job content" of the vacant position after t-layer propagation; L 7 h jt is the node representation of the i-th sentence of the "job content" of the employee's current position after t-layer propagation.

[0087] Step 4: Determine the total matching degree score s of the key element features in the employee's current position and the key element features in the vacant position. total .

[0088] Since "gender requirement N 1", "occupational skill certificate requirement N 2 ", "age requirement N 3 ", "salary N 4 ", "education requirement N 5 " belong to the basic conditions for job transfer, the matching degrees s 1 , s 2 , s 3 are merged into s p (the second comprehensive matching degree result), and then by accumulating the values of s p , s 4 and s 5 , the total matching degree score s total is obtained. The specific calculation formula is as follows:

[0089]

[0090] s total = s p + 0.5s 4 + 0.5s 5

[0091] Taking the vacant position as "backend software development engineer" as an example, according to the calculation formulas obtained in step 3 and step 4, the total matching degree scores of 2000 employees to be transferred within the enterprise are calculated (the data with the current position of "backend software development engineer" has been excluded), as shown in Table 5 below:

[0092] Table 5

[0093]

[0094]

[0095]

[0096] According to the actual situation of the unit, by adjusting the interval of the total matching score, a list of recommended candidates for positions with corresponding scores is obtained. Then, according to the satisfaction of employees and managers with the results of the adjusted positions, the interval of the total matching score is flexibly adjusted to obtain the optimal list of recommended personnel for the vacant positions within the enterprise. For example, the interval of the total matching score is set as [1.2, 2.0], and the specific list is shown in Table 6 below:

[0097] Table 6

[0098]

[0099]

[0100]

[0101]

[0102] Transfer the positions of a human resources administrator with 15 years of work experience in 5 years, and compare it with the method of transferring positions using the method of the present invention. Moreover, compare from two aspects: the transfer period and the number of satisfied people among managers with the list of recommended personnel for position transfer. The specific structure is shown in Table 7 and Table 8 below:

[0103] Table 7

[0104]

[0105] Table 8

[0106] Number of recommended job transfers Average job transfer cycle of the method of the present invention / minute Average number of satisfied managers / person 5 4.03 4.88 10 4.03 8.34 15 4.03 12.63 20 4.03 16.01 25 4.03 21.59 30 4.03 24.83 35 4.03 24.53

[0107] It can be seen through comparison that the method of the present invention can quickly obtain the optimal list of recommended personnel for vacant positions, and the effect of the position matching degree before and after position transfer is close to that of a human resources administrator with 5 years of work experience.

[0108] The present invention provides an automatic matching method for internal position transfer in an enterprise. The method collects the characteristic data of employees' current positions and vacant positions, screens the key characteristic data, and forms a data set of employees' current positions and vacant positions. For different types of data, a targeted matching mechanism is designed, which improves the matching degree of position recommendation, achieves the effect of reducing the influence of subjective factors of human resources management personnel, shortening the time for employees to adapt to new positions after position transfer, and breaking through the scenario limitations of traditional position recommendation methods.

Claims

1. An automatic matching method for internal job transfers in an enterprise, characterized in that, it includes the following steps: 1) Preprocess the constructed human resources dataset to obtain the current job key element feature dataset of the employee to be transferred, and the corresponding key element feature dataset of the vacant job; 2) Calculate the matching degree of each key element feature corresponding to the current job and the vacant job of the employee to be transferred using the corresponding matching mechanism, obtain the matching degree results of each key element feature corresponding to the current job and the vacant job of the employee to be transferred, and process the matching degree results of each key element feature to obtain the total matching degree result of the current job and the vacant job of the employee to be transferred; 3) Determine whether the employee to be transferred meets the requirements of the vacant job according to the total matching degree result of the current job and the vacant job of the employee to be transferred.

2. The automatic matching method for internal job transfers in an enterprise according to claim 1, characterized in that, the key element features include: job name, salary, professional skill certificate, age, gender, education level, and job content.

3. The automatic matching method for internal job transfers in an enterprise according to claim 1, characterized in that, the matching mechanism includes a direct text matching mechanism, a calculation matching mechanism, and a text similarity matching mechanism; among them, the direct text matching mechanism refers to directly comparing whether a certain key element feature corresponding to the current job and the vacant job is the same; the calculation matching mechanism refers to comparing whether there is a size relationship between a certain key element feature corresponding to the current job and the vacant job; the text similarity matching mechanism refers to calculating the cosine similarity between a certain key element feature corresponding to the current job and the vacant job.

4. The automatic matching method for internal job transfers in an enterprise according to claim 3, characterized in that, the matching degree results of each key element feature are matching degree scores. The key element features using the direct text matching mechanism for matching include gender and professional skill certificate, and its calculation formula is: Among them, s 1 is the gender matching score between the employee's current position and the vacant position; L 1 is the gender of the employee; N 1 is the gender requirement of the vacant position; s 2 is the vocational skill certificate matching score between the employee's current position and the vacant position; L 2 is the vocational skill certificate that the employee already has; N 2 is the vocational skill certificate requirement of the vacant position.

5. The automatic matching method for internal job transfers in an enterprise according to claim 3, characterized in that, the matching degree results of each key element feature are matching degree scores. The key element features using the calculation matching mechanism for matching include salary, age, and education level, and its calculation formula is: Among them, s 31 is the age matching score between the employee's current position and the vacant position; L 3 is the age of the employee; N 3 is the age requirement of the vacant position; s 32 is the salary matching score between the employee's current position and the vacant position; L 4 is the salary of the employee's current position; N 4 is the salary of the vacant position; s 33 is the education matching score between the employee's current position and the vacant position; L 5 is the education of the employee; N 5 is the education requirement of the vacant position.

6. The automatic matching method for internal job transfers in an enterprise according to claim 3, characterized in that, the matching degree results of each key element feature are matching degree scores. The key element features using the text similarity matching mechanism for matching include job name and job content, and its calculation formula is: s 4 (N 6 ,L 6 ) = cos_sim(N 6 h jt ,L 6 h it ) s 5 (N 7 ,L 7 ) = cos_sim(N 7 h jt ,L 7 h it ) Among them, L 6 is the job title of the employee's current position; N 6 is the job title of the vacant position; L 6 h it is the node after t-layer propagation of the name of the i-th position in the employee's current position; N 6 h jt is the node after t-layer propagation of the name of the j-th position in the vacant position; s 4 is the matching score of the job titles between the employee's current position and the vacant position; cos_sim is the cosine similarity calculation function; L 7 is the job content of the employee's current position; L 7 h it is the node after t-layer propagation of the i-th sentence in the job content of the employee's current position; N 7 is the job content of the vacant position; N 7 h jt is the node after t-layer propagation of the j-th sentence in the job content of the vacant position; s 5 is the matching score of the job contents between the employee's current position and the vacant position.

7. The automatic matching method for internal job transfers in an enterprise according to claim 2, characterized in that, the specific steps for processing the matching degree results of each key element feature are: first integrate the matching degree results of salary, age, and education level to obtain the first comprehensive matching degree result, then integrate the first comprehensive matching degree result with the matching degree results of gender and professional skill certificate to obtain the second comprehensive matching degree result, and finally perform a weighted sum of the second comprehensive matching degree result and the matching degree results of job name and job content to obtain the total matching degree result of the current job of the employee to be transferred. Its calculation formula is as follows: s total = k 1 s p + k 2 s 4 + k 3 s 5 Among them, s 31 is the age matching degree result between the employee's current position and the vacant position; s 32 is the salary matching degree result between the employee's current position and the vacant position; s 33 is the educational background matching degree result between the employee's current position and the vacant position; s 3 is the first comprehensive matching degree result; s 1 is the gender matching degree result between the employee's current position and the vacant position; s 2 is the vocational skill certificate matching degree result between the employee's current position and the vacant position; s p is the second comprehensive matching degree result; s 4 is the position name matching degree result between the employee's current position and the vacant position; s 5 is the job content matching degree result between the employee's current position and the vacant position; s total is the total matching degree result between the current position of the employee to be transferred and the vacant position; k 1 , k 2 and k 3 are the weights of the second comprehensive matching degree result, the weight of the position name matching degree result between the employee's current position and the vacant position, and the weight of the job content matching degree result between the employee's current position and the vacant position, respectively.

8. The method for automatically matching internal job transfers in an enterprise according to claim 5, characterized in that, in the process of calculating the educational background matching score between the employee's current position and the vacant position, it is necessary to first convert each educational background into a corresponding score and calculate using the scores.

Images