Professional and industrial talent demand matching method and system based on AI identification

Abstract

The invention belongs to the field of talent market analysis, and particularly relates to a professional and industrial talent demand matching method and system based on AI identification. The method comprises the following steps: acquiring published student employment data in each university, classifying the acquired student employment data, adjusting the employment rate of student recruitment events in combination with previous student employment data classification, analyzing and matching posts for previous graduates, and outputting post matching degrees. And visual employment recommendation data is output in combination with the employment rate of the previous students and the employment matching degree of the previous students. Students waiting for employment can make employment preparation more quickly and can be helped to be refined, and meanwhile talents can be conveyed to enterprises more accurately.

Description

TECHNICAL FIELD

[0001] The application belongs to the field of talent market analysis, and particularly relates to a professional and industry talent demand matching method and system based on AI recognition. BACKGROUND

[0002] At present, in the industry and higher education institutions, the goal of talent training in higher education institutions is not clear enough, and the education mode mostly focuses on theoretical teaching, so the talents trained are difficult to adapt to the rapid development of the current society, thereby the precise human resources cannot be output for the industry, resulting in the problem that it is difficult for college students to find employment in the corresponding major, and the enterprise has a large talent gap. There is a large information gap between the information of industry recruitment talents and student employment when students are in school, so that students cannot understand the nature of the work and the work matched with the major, and the enterprise can only start from the student's educational background when recruiting graduates, and the student's school experience has less reference.

[0003] In the prior art, talent model information is mostly generated from statistical data and experience summary, and talent model information storage and updating is not digitized, which requires a large amount of manual research and sorting and is slow to update, and manual research and sorting need to be performed again when updating talent model information. The skill evaluation corresponding to the existing talent model information has no quantitative standard, and the information has certain post-position, the model information reference rate is low, based on the above problems, an AI recognition based professional and industry talent demand matching method and system is urgently needed. SUMMARY

[0004] The purpose of the present application is to provide an AI recognition based professional and industry talent demand matching method and system to solve one or more technical problems existing in the prior art, and at least provide a beneficial choice or create conditions.

[0005] An AI recognition based professional and industry talent demand matching method, the method comprises the following steps: Obtain student employment data of a plurality of majors of a plurality of schools in an education system within a graduation cycle, wherein the student employment data is used to represent the employment industry selected by students of a plurality of majors in each graduation cycle, and the type of student employment data includes: professional employment proportion and industry employment proportion, the professional employment proportion is used to represent the proportion of employment industry in the major, and the industry employment proportion is used to represent the professional proportion of employees in the industry; The obtained student employment data is screened and classified and optimized by a logistic regression algorithm; Obtaining the professional information and employment tendency of the students who have not graduated, identifying and judging the suitable employment positions and industry development direction of the students through AI, preliminarily matching and recommending the occupation and position, combining the preliminarily matched and recommended occupation and position with the student development variable, twice calculating the recommended weight, obtaining the development direction and recommended occupation and position, and the student development variable is represented as the skill certificate obtained by the student through school learning or the award certificate obtained by participating in the competition for weight addition; Through calculation, the final score is output by combining the preliminary recommendation and the secondary recommendation after excluding the interference factors.

[0006] Further, the student employment data published in each university is obtained through the education system, and the professional employment proportion and the industry employment proportion are subdivided, and the professional employment proportion and the industry employment proportion are respectively subdivided and defined as A_bak and B_bak, the A_bak represents the proportion of student employment industry in the same major, and the B_bak represents the proportion of majors recruited by the graduates of the current year in each industry, the data C_bak is evolved from the data B_bak, and the C_bak represents the proportion of professionals in each industry; The obtained data A_bak and B_bak are respectively obtained through the DFS algorithm to obtain the data intersection node, and the multiple data redundancies in the node are simplified.

[0007] Further, the multiple data redundancies in the node include the following steps: The same school data in the A_bak and B_bak is classified, and the classification rule is that the data A_bak is classified according to the school to obtain the proportion of the employment industry of the students in the major of each school, the data B_bak is classified according to the school to obtain the proportion of the students in each industry, and the refined data is defined, the data A_bak is classified according to the school to obtain the data D_bak, and the data B_bak is classified according to the school to obtain the data E_bak. The obtained data is the data that can be obtained from the existing talent model information, the data is classified in detail according to the data obtained from the existing talent model information, and the classification is based on the CNN model.

[0008] Based on the data categorized by school, the industry coverage rate of different majors across schools was obtained through comparison. This coverage rate includes the coverage rate of majors within industries and the coverage rate of industries within majors. Analysis of each school's coverage rate was conducted to determine if it falls within a reasonable fluctuation range. This reasonable fluctuation range is defined as the periodic data for recruitment events and recruitment categories of each school in cooperation with corresponding companies within each quarterly time interval. The recruitment categories are categorized by company type, and a mapping network is constructed. This mapping network maps data between schools, majors, companies, and students. Students are sorted alphabetically by the first letter of their pinyin names. Schools are defined as the main node, and companies as the child nodes. A recruitment event is defined when a student from a particular major in a school goes to a company for recruitment. The number of recruitment event visits, P, is defined as the number of times the main node performs data analysis for the recruitment event. When a student successfully joins a company through recruitment, the recruitment event load, Y, is defined as the recruitment event load. The effective parameter PT is calculated from the recruitment event load of each recruitment event in the main node and the number of recruitment event visits, P. When the number of recruitment event visits, P > 0, the effective parameter PT = Y / P can be calculated. When the number of recruitment event visits, P ≤ 0, the effective parameter PT = 0. Construct a recruitment classification sequence [PT] based on each valid parameter value, where [PT] = { , , ..., , ..., }, the Here, represents the valid parameter value in the k-th recruitment category, and n is the total number of recruitment categories. For a recruitment event, each recruitment event has a level category Dg, which is the level category of the recruitment event based on the school. = D_bak(i), where D_bak(i) is the data D_bak in the i-th school, and each recruitment event has a grade classification Dg, which is the ranking of the recruitment event in the database of the master node in the historical recruitment grade. Define a set of level events, Dgr, through the level classification Dg, where Dgr = { }, where k×i is the total number of the graded event set Dgr. The larger the graded classification value of Dgr, the higher the employment rate of that major in that master node. For a recruitment event, each recruitment event has a master node weight level S related to that recruitment event. The formula for calculating the weight level S is: ; Where m is the cumulative independent variable. This represents the level of the m-th recruitment visit within the same master node under the same recruitment event; an arithmetic mean of the grade Dgr of the recruitment interview times in the same main node for the same recruitment event; the weight grade of the main node represents an adjustment weight according to an influence of each recruitment event on the employment rate of students in the school, and a sequence S of recruitment event authorities S is constructed according to the weight grades S of the respective main nodes, S = [S1, S2, …, SN], wherein S1, S2, …, SN represent the weight grades of the Nth main node to which the recruitment event is allocated, and log represents a logarithm function; the elements in the sequence S of the authorities represent weight adjustment coefficients of the recruitment event on the employment rate of the main node. , ,… ], wherein represents the weight grade of the Nth main node to which the recruitment event is allocated, and log is a logarithm function; the elements in the sequence S of the authorities represent weight adjustment coefficients of the recruitment event on the employment rate of the main node.

[0009] Preferably, the employment rates of the schools are adjusted in weight according to the recruitment events, the adjusted employment rates are more valuable for the students to be employed, and the students to be employed can clearly see the employment difficulties of the majors and industries, and the recruitment events include the certificates and honors obtained by the students when they are employed, so that the students in school have a clearer learning direction.

[0010] Further, the major information and employment tendency of the graduating students are obtained, the employment positions and industry development directions suitable for the students are identified and judged by AI, and the primary matching recommended positions and occupations are obtained. The specific steps of the primary matching recommended positions and occupations are as follows: The primary matching recommended positions and occupations are obtained by AI identification and judgment, the teacher evaluation, score data and school behavior data of the students during the school period are obtained, the teacher evaluation, score data and school behavior data are defined as first evaluation data, the school behavior data includes a school comprehensive test score, a participation in competition and award honor score and a student discipline score, the obtained first evaluation data is standardized, and a first evaluation score is calculated according to a set weight value; second evaluation data is obtained from the same data as the first evaluation data of the employed graduated students during the school period; Key words of the text data of the first evaluation data and the second evaluation data are extracted, the key words are obtained based on AI identification, the key words are defined as weights, the content in the text is filtered according to the key words, and a student subjective score is obtained according to the proportion of the key words in the text and the weight of the key words; The key words obtained by the graduating students are defined as a first key word group, the key words obtained by the employed students are defined as a second key word group, the first key word group and the second key word group are de-duplicated, the first key word group and the second key word group are matched, and a similar key word proportion is determined; The employment position matching degree of the graduating students with the similar proportion of the employed students is calculated by obtaining the employment positions of the employed students, judging the subjective score of the graduating students with the similar proportion, and calculating the first evaluation score, and the calculation method of the matching degree is: Obtain keywords from job postings, match these keywords with the keywords in job postings for recent graduates to obtain the keyword matching degree D for recent graduates, and obtain the keyword matching degree U between employed students and their job postings. Calculate the job matching degree for recent graduates. ; in, To ensure a good match between the job positions of employed students and the proportion of those with the most similar keywords to recent graduates, This refers to the matching degree of job positions between recent graduates and employed students with the same major and the most similar keywords. The matching degree between recent graduates and employed students in the same positions as those with the same major and a similar keyword ratio to recent graduates is calculated. , , It can be the same data. , , Recent graduates and , , Keyword matching degree for the corresponding job position , , respectively employed students and , , Keyword matching degree for the corresponding job position, where A is the total number of keywords obtained, and log is the logarithmic function; The matching degree of the three matching degrees is used to recalculate the matching degree of the job positions of employed students with the second most similar keyword ratio to the recent graduates. This process is repeated until the matching degree of the job position is less than twice the maximum matching degree. At this point, the matching degree calculation stops, and the recommended job position and the matching degree of the job position are output.

[0011] Furthermore, an event matrix is ​​established based on recruitment events, and this event matrix is ​​used to filter out data redundancy. Create an empty array E1, add all the recruitment data in the mapping relationship network to the data E1, the arrangement order of the recruitment events in the array E1 is from small to large according to the classification Dg of the recruitment events, define the subscript of the element with the smallest value in the array E1 as s, create an empty array E2, add the matching degree of the fresh graduate employment positions to the array E2 from high to low, record E2f as the fth element in the array E2, f = 1, 2, …, N, create an empty array KE, add E21, E22, …, E2s-1, E2s+1, E2s+2, …, E2N to the array KE in turn, record KEj as the jth element in the array KE, j = 1, 2, …, N-1; From j = 1, traverse the value range of j, update the value of KEj in the array KE as KEj-E2s in turn, save the updated array KE, and compress the data similar to the similar recruitment data in the array E1 by changing the order; Select the subscript of the element with the smallest value from the array KE and record it as v, record E1r as the rth element in the array E1, delete the elements with the subscripts s and v in the array E1, define E1s and E1v as the same data, create an empty test set, initialize the integer variable k = 1, k ∈ [1, N-1]; From the first element of the test array, subtract the value of KEk from the value of each element in the test array to obtain N-2 values u(1), u(2), …, u(w), …, u(N-2), w = 1, 2, …, N-2, create a variable DW, and assign DW by the following formula: DW= ; Wherein, L is the proportion of the same value and the total value obtained by comparing the array E1 and the array E2; N-2 values D1, D2, …, DN-2 are sequentially composed into a weighted array, record the element with the smallest value in the weighted array as n(min), record the subscript of n(min) in the weighted array as T, compare the sizes of n(min), E2s and E1v, when n(min) < min{CVs, KUv}, add the current T value to the test set, wherein min() represents the element with the smallest value in the set; When the value of k is less than N-1, increase the value of k by 1, and return to the test data calculation; when the value of k is equal to N-1; Screening all high-weight elements in the array, record all high-weight element index as high-weight index, in the array KA, the index of the high-weight index is recorded as redundant data, all redundant data is saved in the data path storage to the database, and the redundant data is deleted, and the index of the high-weight element outside the index is recorded as ordinary index, in the array KA, the index of the ordinary index is recorded as effective data; Create a set D and a set E, the set D includes the array E2 corresponding to the array E1 recruitment enterprise enterprise matching degree, the set E includes the array E1 corresponding to the array E2 enterprise matching degree manager recruitment event, initialize the distance event matrix through set D and set E, define the matrix as Set the initial distance of the data in set D to infinity, repeat the traversal of all data in set D, calculate the shortest distance from all data to the initial data, the distance is quantified by matching degree, traverse set E, judge whether the distance between all data in set D and the initial data and the number of recruitment events in set E is proportional, if proportional, output matrix If not proportional, update matrix Output matrix Compare the newly output matrix With the array E1 and the array E2 before input, delete the repeated data in the array E1 and the array E2, and ensure that the effective data is screened in the matrix .

[0012] Preferably, by arranging and compressing the array E1 and the array E2, the data effective value is recalculated, and the large amount of redundant array is more intuitive through the matrix Output, the recruitment event of the previous graduate student and the matching degree of the graduate student and each post are arranged to obtain clear and clear comparison array output.

[0013] A professional and industry talent demand matching system based on AI recognition, comprising: An acquisition module is configured to acquire student employment data of a plurality of majors of a plurality of schools in an education system within a graduation cycle; An AI recognition module is configured to identify student majors and employment tendencies through AI recognition, and generate employment recommendations and development directions; A processing module is configured to process data acquired by the acquisition module and calculate recommended data; A feedback module is configured to give feedback recommendations to student input information based on data acquired by the processing module.

[0014] A non-volatile storage medium having stored therein a computer program, wherein a device in which the non-volatile storage medium is located implements the AI recognition-based professional and industry talent demand matching method in any one of the above methods by running the computer program.

[0015] A computer program product, comprising: a computer program, wherein the computer program, when executed by a processor, implements the AI recognition-based professional and industry talent demand matching method in any one of the above methods.

[0016] The beneficial effects of the present application are: obtaining the employment data of the employed students and analyzing the recruitment events when the students are recruited, obtaining a more real college employment rate, and also obtaining the employment success rate of each industry and each school, combining the score data and competition honor of the graduates with the subjective score given by the teachers with the recruitment and employment success rate of the previous graduates, so that the employment matching rate can be more clearly seen, the students to be employed can better prepare for employment and can improve the direction, and the enterprises can more accurately deliver talents. BRIEF DESCRIPTION OF DRAWINGS

[0017] The above and other features of the present application will become more apparent from the following detailed description of an embodiment thereof, taken in conjunction with the accompanying drawings, in which like reference numerals refer to like elements in the several views. As will be obvious from the drawings and detailed description, the embodiments described in the following description are only by way of explanations of the present application. It will be readily apparent to those skilled in the art that various changes and modifications of the embodiments described herein can be made without departing from the scope of the present application, and that the scope of the application includes all such modifications and changes that come within the scope of the claims. Figure 1 A flowchart of an AI recognition-based professional and industry talent demand matching method is shown. DETAILED DESCRIPTION

[0018] The concept, specific structure and resulting technical effects of the present application will be described in detail below with reference to the embodiments and drawings, so as to fully understand the purpose, scheme and effects of the present application. It should be noted that the embodiments and features in the embodiments in the present application can be combined with each other without conflict.

[0019] As shown in Figure 1 obtaining student employment data of a plurality of majors of a plurality of schools in an education system within a graduation cycle, wherein the student employment data is used to represent the employment industry selected by students of a plurality of majors in each graduation cycle, and the types of student employment data include: professional employment proportion and industry employment proportion, the professional employment proportion is used to represent the proportion of employment industry in the major, and the industry employment proportion is used to represent the proportion of professionals in the industry; The obtained student employment data is screened and classified by a logistic regression algorithm; The major information and employment tendency of the non-graduated students are obtained, and the suitable employment positions and industry development directions of the students are identified and judged by AI, the recommended occupations and positions are preliminarily matched, the recommended occupations and positions preliminarily matched are combined with the student development variables, the recommended weight is calculated twice, and the development direction and the recommended occupation and position are obtained, wherein the student development variable represents the skill certificate obtained by the student through school learning or the award certificate obtained by participating in the competition. The final score is output by calculating the interference factors, combining the preliminary recommendation and the secondary recommendation.

[0020] Further, the student employment data published by each university is obtained through the education system, and the professional employment proportion and the industry employment proportion are subdivided, and the professional employment proportion and the industry employment proportion are respectively subdivided and defined as A_bak and B_bak, wherein A_bak represents the proportion of student employment industry in the same major, and B_bak represents the proportion of majors recruited by the graduates of the current year in each industry, and the data C_bak is evolved from the data B_bak, and C_bak represents the proportion of employees in each industry. The obtained data A_bak and B_bak are respectively obtained by the DFS algorithm to obtain the data intersection nodes, and the multiple data redundancies in the nodes are simplified.

[0021] Further, the multiple data redundancies in the nodes include the following steps: The same school data in A_bak and B_bak is classified, and the classification rule is that data A_bak is classified according to schools to obtain the proportion of student employment industry in each school major, and data B_bak is classified according to schools to obtain the proportion of students in each industry in each school, and the refined data is defined, and the data A_bak classified according to schools is D_bak, and the data B_bak classified according to schools is E_bak. The obtained data is the data that can be obtained from the existing talent model information, and the data is classified in detail according to the data obtained from the existing talent model information, and the classification is based on the CNN model.

[0022] Based on the data categorized by school, the industry coverage rate of different majors across schools was obtained through comparison. This coverage rate includes the coverage rate of majors within industries and the coverage rate of industries within majors. Analysis of each school's coverage rate was conducted to determine if it falls within a reasonable fluctuation range. This reasonable fluctuation range is defined as the periodic data for recruitment events and recruitment categories of each school in cooperation with corresponding companies within each quarterly time interval. The recruitment categories are categorized by company type, and a mapping network is constructed. This mapping network maps data between schools, majors, companies, and students. Students are sorted alphabetically by the first letter of their pinyin names. Schools are defined as the main node, and companies as the child nodes. A recruitment event is defined when a student from a particular major in a school goes to a company for recruitment. The number of recruitment event visits, P, is defined as the number of times the main node performs data analysis for the recruitment event. When a student successfully joins a company through recruitment, the recruitment event load, Y, is defined as the recruitment event load. The effective parameter PT is calculated from the recruitment event load of each recruitment event in the main node and the number of recruitment event visits, P. When the number of recruitment event visits, P > 0, the effective parameter PT = Y / P can be calculated. When the number of recruitment event visits, P ≤ 0, the effective parameter PT = 0. Construct a recruitment classification sequence [PT] based on each valid parameter value, where [PT] = { , , ..., , ..., }, the Here, represents the valid parameter value in the k-th recruitment category, and n is the total number of recruitment categories. For a recruitment event, each recruitment event has a level category Dg, which is the level category of the recruitment event based on the school. = D_bak(i), where D_bak(i) is the data D_bak in the i-th school, and each recruitment event has a grade classification Dg, which is the ranking of the recruitment event in the database of the master node in the historical recruitment grade. Define a set of level events, Dgr, through the level classification Dg, where Dgr = { }, where k×i is the total number of the graded event set Dgr. The larger the graded classification value of Dgr, the higher the employment rate of that major in that master node. For a recruitment event, each recruitment event has a master node weight level S related to that recruitment event. The formula for calculating the weight level S is: ; Where m is the cumulative independent variable. This represents the level of the m-th recruitment visit within the same master node under the same recruitment event; The weight level Dgr is the arithmetic mean of the number of recruitment visits within the same master node under the same recruitment event; the weight level of the master node represents the adjustment weight based on the impact of each recruitment event on the employment rate of students in the school, and a recruitment event permission sequence S is constructed based on the weight level S of each master node, S = [ , , ..., ],in The weight level of the recruitment event assigned to the Nth master node is represented by log, where log is the logarithmic function; the elements in the permission sequence S represent the weight adjustment coefficients of the recruitment event on the employment rate of the master node.

[0023] Preferably, the employment rates of each school are weighted and adjusted according to the recruitment events. The adjusted employment rates are more valuable for students seeking employment, and at the same time, they allow students to see more clearly the employment difficulty of each major and industry. In addition, the recruitment events include the certificates and honors obtained by students when they are employed, which can also give students in school a clearer direction for their studies.

[0024] Furthermore, the program obtains the professional information and employment preferences of graduating students, uses AI to identify suitable job positions and industry development directions for them, and initially matches and recommends careers and positions. The specific steps for this initial matching and recommendation of careers and positions are as follows: The system acquires AI-generated preliminary matching and recommendations for careers and positions. It also acquires teacher evaluations, academic performance data, and on-campus behavior data of students during their time at school. These teacher evaluations, academic performance data, and on-campus behavior data are defined as the first evaluation data. The on-campus behavior data includes: on-campus comprehensive test scores, scores for participation in competitions and awards, and scores for student disciplinary violations. The acquired first evaluation data is standardized, and a first evaluation score is calculated based on set weight values. Finally, the system acquires data from previous graduates who have already entered the workforce, and this data is used to identify and recommend similar jobs and positions. This data is defined as the second evaluation data. Keywords are extracted from the text data of the first evaluation data and the second evaluation data. The keywords are obtained based on AI recognition. Weights are defined for the keywords. The text content is filtered according to the keywords. The subjective scores of students are obtained based on the proportion and weight of the keywords in the text. The keywords obtained by recent graduates are defined as the first keyword group, and the keywords obtained by employed students are defined as the second keyword group. The first keyword group and the second keyword group are deduplicated, and the first keyword group and the second keyword group are matched to determine the proportion of similar keywords. By obtaining the job positions of already employed students, the job matching degree between the current year's graduates and the employed students of a similar proportion is calculated by comparing the subjective scores and first evaluation scores of recent graduates with those scores. The calculation method for the matching degree is as follows: Obtain keywords from job postings, match these keywords with the keywords in job postings for recent graduates to obtain the keyword matching degree D for recent graduates, and obtain the keyword matching degree U between employed students and their job postings. Calculate the job matching degree for recent graduates. ; in, To ensure a good match between the job positions of employed students and the proportion of those with the most similar keywords to recent graduates, This refers to the matching degree of job positions between recent graduates and employed students with the same major and the most similar keywords. The matching degree between recent graduates and employed students in the same positions as those with the same major and a similar keyword ratio to recent graduates is calculated. , , It can be the same data. , , Recent graduates and , , Keyword matching degree for the corresponding job position , , respectively employed students and , , Keyword matching degree for the corresponding job position, where A is the total number of keywords obtained, and log is the logarithmic function; Preferably, the , , Normally, the data might be the same. If different data exist, it indicates that the student may excel in a particular area, or that the student acquired skills in different professional fields during their studies. , , Data mean (mean) , , ), calculate the , , Data variance ; p takes values ​​of d1, d2, and d3. Calculate the mean variance of all students in this school. pp=[1,M], where pp is an integer, analyze the variance of the data. When data variance ≤ If a student is defined as a stable student, then the matching degree is directly output. > If the student's comprehensive test score, participation in competitions and awards are weighted and the matching degree is recalculated, the matching degree is recalculated only once and then the matching degree is output.

[0025] The matching degree of the three matching degrees is used to recalculate the matching degree of the job positions of employed students with the second most similar keyword ratio to the recent graduates. This process is repeated until the matching degree of the job position is less than twice the maximum matching degree. At this point, the matching degree calculation stops, and the recommended job position and the matching degree of the job position are output.

[0026] Furthermore, an event matrix is ​​established based on recruitment events, and this event matrix is ​​used to filter out data redundancy. Create an empty array E1 and add all recruitment data from the mapping relationship network to E1. The recruitment events in array E1 are sorted from smallest to largest according to the recruitment event level classification Dg. Define the index of the smallest element in array E1 as s. Create an empty array E2 and add the job matching degree of fresh graduates to array E2 from high to low. Let E2f be the f-th element in array E2, f=1,2,…,N. Create an empty array KE and add E21,E22,...,E2s-1,E2s+1,E2s+2,...,E2N to array KE in sequence. Let KEj be the j-th element in array KE, j=1,2,...,N-1. Starting from j=1, iterate through the range of j values, update the value of KEj in array KE to KEj-E2s in turn, save the updated array KE, and compress the data by sequentially changing the data in array E1 to be the same as similar recruitment data. Select the index of the smallest element from array KE and denote it as v. Let E1r be the r-th element in array E1. Delete the elements with index s and index v from array E1. Define the elements with deleted E1s and E1v as completely identical data. Create a blank test set and initialize the integer variable k=1, k∈[1,N-1]. Starting from the first element of the test array, subtract the value of KEk from the value of each element in the test array to obtain N-2 values ​​u(1), u(2), ..., u(w), ..., u(N-2), where w = 1, 2, ..., N-2. Create a variable DW and assign a value to DW using the following formula: DW= ; Where L is the ratio of the same value to the total value between array E1 and array E2, obtained by comparing the two arrays; Arrange the N - 2 values D1, D2,..., DN - 2 in sequence to form a weighted array. Denote the element with the smallest value in the weighted array as n(min), and denote the index of n(min) in the weighted array as T. Compare the magnitudes of the three values n(min), E2s, and E1v. When n(min) < min{CVs, KUv}, add the current T value to the test set, where min() represents the element with the smallest value in the set. When the value of k is less than N - 1, increment the value of k by 1 and return to calculate the test data. When the value of k is equal to N - 1; Select all high - weight elements in the array. Denote the indices of all high - weight elements as high - weight indices. In the array KA, mark the elements with high - weight indices as redundant data. Save all redundant data to the database through the data path, delete the redundant data, and mark the indices other than high - weight elements as ordinary indices. In the array KA, mark the elements with ordinary indices as valid data. Create a set D and a set E. The set D includes the enterprise matching degrees corresponding to the recruitment enterprises in array E1 in array E2, and the set E includes the recruitment events corresponding to the enterprise matching degree managers in array E1 in array E2. Initialize the distance event matrix through the sets D and E, and define the matrix as , set the initial distance of the data in set D to infinity. Repeatedly traverse all the data in set D, calculate the shortest distance from all data to the initial data. The distance is quantified by the matching degree. Traverse set E, and determine whether the distance from all data in set D to the initial data is proportional to the number of recruitment events in set E. If it is proportional, output the matrix , if it is not proportional, update the matrix , through the output matrix , compare the newly output matrix with the arrays E1 and E2 before input, and delete the duplicate data in arrays E1 and E2 to ensure the selection of valid data in the matrix .

[0027] Preferably, through arranging and compressing arrays E1 and E2, recalculate the valid data values, making the arrays with a large amount of redundancy more intuitively output through the matrix , and output a clear and understandable comparison array by organizing the recruitment events of往届毕业生 (previous graduates) and the matching degrees of应届毕业生 (fresh graduates) with each position.

[0028] A matching system for the needs of majors and industries based on AI recognition, comprising: An acquisition module for acquiring the employment data of students in several majors within several schools in the graduation cycle in the education system; An AI recognition module for identifying students' majors and employment tendencies through AI and generating employment recommendations and development directions; The processing module is used to process the data acquired by the acquisition module and calculate the recommended data. The feedback module uses the data obtained from the processing module to provide feedback and recommendations based on the student's input.

[0029] A non-volatile storage medium storing a computer program, wherein the device containing the non-volatile storage medium executes, by running the computer program, any one of the above methods, a matching method for professional and industry talent needs based on AI recognition.

[0030] A computer program product includes: a computer program, wherein, when the computer program is executed by a processor, it implements the AI-based matching method for professional and industry talent needs as described in any one of the above methods.

[0031] The computer program product may be a processor, which can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete component gate circuits, transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or any conventional processor. The processor is the control center of the system, connecting various sub-regions of the system via various interfaces and lines.

[0032] The non-volatile storage medium can be used to store the computer program and / or modules. The processor implements various system functions by running or executing the computer program and / or modules stored in the memory, and by calling data stored in the memory. The memory may mainly include a program storage area and a data storage area. The program storage area may store the operating system, at least one application program required for a function (such as sound playback function, image playback function, etc.), etc.; the data storage area may store data created according to the use of the mobile phone (such as audio data, phonebook, etc.). In addition, the memory may include high-speed random access memory, and may also include non-volatile memory, such as hard disk, memory, plug-in hard disk, smart media card (SMC), secure digital (SD) card, flash card, at least one disk storage device, flash memory device, or other volatile solid-state storage device.

[0033] Although the invention has been described in considerable detail and particularly with regard to several of the described embodiments, it is not intended to limit itself to any of these details or embodiments or any particular embodiment, thereby effectively covering the intended scope of the invention. Furthermore, the invention has been described above with respect to embodiments foreseeable by the inventors in order to provide a useful description, and non-substantial modifications to the invention that have not yet been foreseen may still represent equivalent modifications.

Claims

1. A method for matching professional and industry talent needs based on AI recognition, characterized in that, The method includes the following steps: The system obtains employment data of students from several majors in several schools within the education system during the graduation cycle. The student employment data is used to represent the employment industries chosen by students of several majors in each graduation cycle. The types of student employment data include: major employment ratio and industry employment ratio. The major employment ratio is used to represent the proportion of students employed in a particular major in a particular industry. The industry employment ratio is used to represent the proportion of students employed in a particular industry by their respective majors. The acquired student employment data was filtered and categorized using a logistic regression algorithm. The system obtains the professional information and employment preferences of students who have not yet graduated, uses AI to identify and determine suitable job positions and industry development directions for students, and initially matches and recommends occupations and positions. The initial matched and recommended occupations and positions are combined with student development variables, and the recommendation weights are calculated a second time to obtain development directions and recommended occupations and positions. The student development variables are represented by the weighting of skill certificates obtained by students through their studies at school or award certificates obtained through participating in competitions. After calculating and eliminating interfering factors, the final score is output by combining the initial recommendation and the secondary recommendation.

2. The method for matching professional and industry talent needs based on AI recognition according to claim 1, characterized in that, We obtain published student employment data from various universities through the education system, as well as two types of data: the proportion of students employed by a major and the proportion of students employed by an industry. These two types of data are then further subdivided and defined as A_bak and B_bak, respectively. A_bak represents the proportion of students employed by an industry within the same major, while B_bak represents the proportion of majors that recruit recent graduates in each industry. Data C_bak is derived from B_bak, and C_bak represents the proportion of employees by major in each industry. The acquired data A_bak and B_bak are used to obtain the data intersection nodes through the DFS algorithm, and the multiple data redundancies in the nodes are simplified.

3. The method for matching professional and industry talent needs based on AI recognition according to claim 2, characterized in that, The simplification of multiple data redundancy in the node includes the following steps: The data from the same school in A_bak and B_bak are classified according to the following rules: Data A_bak is classified by school into the proportion of students employed in each industry according to their majors at each school; the proportion of recent graduates in Data B_bak is classified by school to obtain the proportion of students from each school in each industry; the refined data is defined as follows: Data A_bak classified by school is designated as D_bak, and Data B_bak's proportion of recent graduates classified by school is designated as E_bak. Based on the data categorized by school, the industry coverage rate of different majors across schools was obtained through comparison. This coverage rate includes the coverage rate of majors within industries and the coverage rate of industries within majors. Analysis of each school's coverage rate was conducted to determine if it falls within a reasonable fluctuation range. This reasonable fluctuation range is defined as the periodic data for recruitment events and recruitment categories of each school in cooperation with corresponding companies within each quarterly time interval. The recruitment categories are categorized by company type, and a mapping network is constructed. This mapping network maps data between schools, majors, companies, and students. Students are sorted alphabetically by the first letter of their pinyin names. Schools are defined as the main node, and companies as the child nodes. A recruitment event is defined when a student from a particular major in a school goes to a company for recruitment. The number of recruitment event visits, P, is defined as the number of times the main node performs data analysis for the recruitment event. When a student successfully joins a company through recruitment, the recruitment event load, Y, is defined as the recruitment event load. The effective parameter PT is calculated from the recruitment event load of each recruitment event in the main node and the number of recruitment event visits, P. When the number of recruitment event visits, P > 0, the effective parameter PT = Y / P can be calculated. When the number of recruitment event visits, P ≤ 0, the effective parameter PT = 0. Construct a recruitment classification sequence [PT] based on each valid parameter value, where [PT] = { , , ..., , ..., }, the Here, represents the valid parameter value in the k-th recruitment category, and n is the total number of recruitment categories. For a recruitment event, each recruitment event has a level category Dg, which is the level category of the recruitment event based on the school. = D_bak(i), where D_bak(i) is the data D_bak in the i-th school, and each recruitment event has a grade classification Dg, which is the ranking of the recruitment event in the database of the master node in the historical recruitment grade. Define a set of level events, Dgr, through the level classification Dg, where Dgr = { }, where k×i is the total number of the graded event set Dgr. The larger the graded classification value of Dgr, the higher the employment rate of that major in that master node. For a recruitment event, each recruitment event has a master node weight level S related to that recruitment event. The formula for calculating the weight level S is: ； Where m is the cumulative independent variable. This represents the level of the m-th recruitment visit within the same master node under the same recruitment event; The weight level Dgr is the arithmetic mean of the number of recruitment visits within the same master node under the same recruitment event; the weight level of the master node represents the adjustment weight based on the impact of each recruitment event on the employment rate of students in the school, and a recruitment event permission sequence S is constructed based on the weight level S of each master node, S = [ , , ..., ],in The weight level of the recruitment event assigned to the Nth master node is represented by log, where log is the logarithmic function; the elements in the permission sequence S represent the weight adjustment coefficients of the recruitment event on the employment rate of the master node.

4. The method for matching professional and industry talent needs based on AI recognition according to claim 1, characterized in that, The process involves obtaining the professional information and employment preferences of graduating students, using AI to identify suitable job positions and industry development directions, and initially matching and recommending careers and positions. The specific steps for this initial matching and recommendation of careers and positions are as follows: The system acquires AI-generated preliminary matching and recommendations for careers and positions. It also acquires teacher evaluations, academic performance data, and on-campus behavior data of students during their time at school. These teacher evaluations, academic performance data, and on-campus behavior data are defined as the first evaluation data. The on-campus behavior data includes: on-campus comprehensive test scores, scores for participation in competitions and awards, and scores for student disciplinary violations. The acquired first evaluation data is standardized, and a first evaluation score is calculated based on set weight values. Finally, the system acquires data from previous graduates who have already entered the workforce, and this data is used to identify and recommend similar jobs and positions. This data is defined as the second evaluation data. Keywords are extracted from the text data of the first evaluation data and the second evaluation data. The keywords are obtained based on AI recognition. Weights are defined for the keywords. The text content is filtered according to the keywords. The subjective scores of students are obtained based on the proportion and weight of the keywords in the text. The keywords obtained by recent graduates are defined as the first keyword group, and the keywords obtained by employed students are defined as the second keyword group. The first keyword group and the second keyword group are deduplicated, and the first keyword group and the second keyword group are matched to determine the proportion of similar keywords. By obtaining the job positions of already employed students, the job matching degree between the current year's graduates and the employed students of a similar proportion is calculated by comparing the subjective scores and first evaluation scores of recent graduates with those scores. The calculation method for the matching degree is as follows: Obtain the keywords of job recruitment information, match the keywords with the keywords of job recruitment information of fresh graduate students to obtain the keyword matching degree D of fresh graduate students, obtain the keyword matching degree of employed students and their employed job recruitment information, defined as the keyword matching degree U of employed students, and calculate the job matching degree of fresh graduate students: ； in, To ensure a good match between the job positions of employed students and the proportion of those with the most similar keywords to recent graduates, This refers to the matching degree of job positions between recent graduates and employed students with the same major and the most similar keywords. The matching degree between recent graduates and employed students in the same positions as those with the same major and a similar keyword ratio to recent graduates is calculated. , , It can be the same data. , , Recent graduates and , , Keyword matching degree for the corresponding job position , , respectively employed students and , , Keyword matching degree for the corresponding job position, where A is the total number of keywords obtained, and log is the logarithmic function; Calculate the matching degree of the employed students' job positions with the ratio of the second closest keyword of fresh graduate students again through the three obtained matching degrees, and so on, until the obtained matching degree is less than twice the maximum matching degree of the job position, stop calculating the matching degree, and output the recommended job position and the job matching degree.

5. The method for matching professional and industry talent needs based on AI recognition according to claim 1, characterized in that, Establish an event matrix through recruitment events and use the event matrix to complete data redundancy screening: Create an array E1 as an empty array, add all recruitment data in the mapping relationship network to the data E1. The arrangement order of the recruitment events in the array E1 is sorted from small to large according to the level classification Dg of the recruitment events. Define the subscript of the element with the smallest value in the array E1 as s. Create an array E2 as an empty array, add the job matching degrees of fresh graduate students from high to low to the array E2. Denote E2f as the f-th element in the array E2, f = 1, 2, …, N. Create an empty array KE, and add E21, E22, …, E2s - 1, E2s + 1, E2s + 2, …, E2N to the array KE in turn. Denote KEj as the j-th element in the array KE, j = 1, 2, …, N - 1; Start from j = 1, traverse the value range of j, and update the value of KEj in the array KE to KEj - E2s in turn, save the updated array KE, and compress the data with the same order change in the array E1 to similar recruitment data; Screen out the subscript of the element with the smallest value from the array KE and denote it as v. Denote E1r as the r-th element in the array E1. Delete the elements with subscript s and subscript v in the array E1. Define the data after deleting E1s and E1v as exactly the same data. Create an empty test set, initialize the integer variable k = 1, k ∈ [1, N - 1]; Start from the first element of the test array, subtract the value of KEk from each element in the test array in turn to obtain N - 2 values u(1), u(2), …, u(w), …, u(N - 2), w = 1, 2, …, N - 2. Create a variable DW and assign a value to DW through the following formula: DW= ; where L is the ratio of the same numerical value to the total numerical value obtained by comparing the array E1 and the array E2; Form the N - 2 values D1, D2, …, DN - 2 into a weighted array in turn. Denote the element with the smallest value in the weighted array as n(min), and denote the subscript of n(min) in the weighted array as T. Compare the sizes of the three values n(min), E2s, and E1v. When n(min) < min{CVs, KUv}, add the current T value to the test set, where min() represents the element with the smallest value in the set; When the value of k is less than N-1, increment the value of k by 1 and return to the test data for calculation; when the value of k is equal to N-1; Filter out all high-weight elements in the array and denote the index of all high-weight elements as high-weight index. In array KA, denote the elements with high-weight index as redundant data. Save all redundant data to the database and delete the redundant data. Denote the indexes outside the high-weight elements as ordinary indexes. In array KA, denote the indexes with ordinary indexes as valid data. Create sets D and E. Set D includes the company matching scores of the companies in array E1 corresponding to the companies in array E2. Set E includes the recruitment events of the managers in array E2 corresponding to the company matching scores of array E1. Initialize a distance event matrix using sets D and E, and define the matrix as follows: Set the initial distance of data in set D to infinity. Repeatedly traverse all data in set D, calculating the shortest distance from all data to the initial data. This distance is quantified by the matching degree. Traverse set E, determining whether the distance between all data in set D and the initial data is proportional to the number of recruitment events in set E. If they are proportional, output the matrix. If they are not proportional, then update the matrix. Through the output matrix The newly output matrix Compare the input with arrays E1 and E2, remove duplicate data from arrays E1 and E2, and ensure that the matrix... Filter valid data.

6. A matching system for professional and industry talent needs based on AI recognition, characterized in that, include: The acquisition module is used to acquire employment data of students from several majors in several schools within the education system during their graduation cycle. The AI ​​recognition module is used to identify students' majors and employment preferences through AI, and generate job recommendations and development directions. The processing module is used to process the data acquired by the acquisition module and calculate the recommended data. The feedback module uses the data obtained from the processing module to provide feedback and recommendations based on the student's input.

7. A non-volatile storage medium, characterized in that, The non-volatile storage medium stores a computer program, wherein the device containing the non-volatile storage medium executes the matching method for professional and industry talent needs based on AI recognition as described in any one of claims 1 to 6 by running the computer program.

8. A computer program product, characterized in that, include: A computer program, wherein when executed by a processor, the computer program implements the AI-based matching method for professional and industry talent needs as described in any one of claims 1 to 6.

Images