Post allocation human resource management system based on personnel difference planning

Abstract

The invention relates to the technical field of human resource management, and provides a post allocation human resource management system based on personnel difference planning, an employee ability evaluation module collects and quantifies employee ability data based on an improved clustering algorithm, and performs center selection, iterative clustering and category determination to obtain an employee ability evaluation result; an evaluation result is obtained through classification in combination with contour coefficient evaluation and parameter adjustment, a post demand analysis module builds a knowledge graph for texts such as post specifications by means of a natural language processing technology, key information is analyzed, skills required by posts are clarified, and a demand result is obtained; the intelligent matching module matches employees and posts by using an algorithm according to the two data, and calculates an optimal distribution scheme; the display module regularly collects employee performance data, compares and displays the employee performance data to a manager for evaluation; according to the method, quantitative combination with improved clustering algorithm calculation is performed by integrating employee capabilities, and matching solution is performed by combining mixed similarity calculation with linear programming, so that accurate classification of employee capabilities and accurate matching of employees and posts are realized.

Description

Technical Field

[0001] The invention relates to the technical field of human resource management, in particular to a human resource management system for job allocation based on personnel difference planning. Background Art

[0002] With the diversified development of corporate business and the increasing complexity of projects, the role of personnel differences in team collaboration has become increasingly prominent. Team collaboration human resource management based on personnel differences aims to accurately identify and make full use of the diverse characteristics of employees in terms of skills, knowledge, experience, personality, work style, etc. Through scientific analysis and planning, suitable personnel are combined into efficient and collaborative teams, and the most suitable tasks are assigned to individuals to achieve the overall corporate goals and optimize performance.

[0003] In today's highly competitive and complex business environment, companies face a wide variety of tasks and projects, covering multiple fields such as technology research and development, marketing, customer service, and operations management, and each field has unique and strict requirements for professional skills and comprehensive qualities. This poses a huge challenge to team building and task allocation in corporate human resource management.

[0004] The prior art still has the following deficiencies:

[0005] 1. Most companies currently rely on limited and relatively single evaluation indicators. For example, the common evaluation method may simply record the employee's skill certificate and years of work, but ignore the importance of deep-level characteristics such as the application effect of skills in actual projects, the depth and breadth of training courses, and the ability to acquire new knowledge for the classification of employee capabilities. Moreover, in terms of classification methods, most of them use subjective judgment or simple linear classification models, which cannot accurately mine employees' potential specialty categories based on data characteristics.

[0006] 2. In the current environment, employees have different abilities. There are some all-round talents with outstanding abilities and some professional talents with their own strengths. However, the existing human resource management system lacks an effective mechanism to accurately allocate positions reasonably based on these ability differences. In the process of matching employees with positions, usually only a single matching indicator is considered, such as simple skill matching or work experience matching, while ignoring the complex adaptation relationship between the overall structure of employee abilities and job requirements. Summary of the invention

[0007] In order to solve the above technical problems, the present invention provides a human resource management system for job allocation based on personnel difference planning, which realizes accurate classification of employee capabilities and accurate matching of employees and jobs by integrating multi-dimensional capability quantitative analysis with improved clustering algorithm and adopting hybrid similarity calculation combined with linear programming solution, so as to solve the problems in the prior art.

[0008] A human resource management system for job allocation based on personnel difference planning, comprising:

[0009] An employee competency assessment module, which is based on an improved clustering algorithm, collects and quantifies employee competency data, selects initial cluster centers, iterates clustering and determines categories, combines silhouette coefficient assessment and parameter adjustment, and implements employee competency classification to obtain employee competency assessment results;

[0010] Position demand analysis module, which establishes a position knowledge graph, analyzes the internal text materials of the enterprise including position descriptions, workflow documents and project cases through natural language processing technology, extracts key information, clarifies the professional skills, knowledge reserves, work experience and learning ability required for each position, and obtains the position demand analysis results;

[0011] An intelligent matching module, which is connected to the employee capability assessment module and the job requirement analysis module, and is used to match employees with jobs using an algorithm based on the employee capability assessment results and job requirement analysis data, and calculate the most appropriate job allocation plan;

[0012] The display module is connected to the data of the intelligent matching module and is used to regularly collect the performance data of employees in their positions, including information on work results, work quality and work efficiency, compare them with the expected performance goals, and display them to relevant managers for evaluating the effectiveness of the job allocation plan.

[0013] Preferably, the employee capability assessment module includes a data collection unit, a model building unit and a model optimization unit;

[0014] The data collection unit collects employee capability data including professional skills, knowledge reserves, work experience, learning ability and communication ability, quantifies the data, and combines them into feature vectors to form an employee capability data set, which serves as the basis for the subsequent classification model.

[0015] The model building unit uses an improved clustering algorithm to first select an initial clustering center by a specific probability method, iteratively assigns employee data to corresponding clusters and updates the clustering center, and finally divides employee ability categories according to the clustering results;

[0016] The model optimization unit uses the silhouette coefficient to evaluate the clustering effect of the model establishment unit, thereby achieving effective division of the comprehensive ability differences of employees.

[0017] Preferably, the data quantification standard of the data acquisition unit is as follows;

[0018] Professional skills are quantified based on a percentage-based scoring mechanism, such as professional skills test scores and evaluation of the effectiveness of skill application in actual projects. Knowledge reserves are assigned a percentage based on academic qualifications and completion of training courses. Work experience is divided based on years of work experience, with the employee with the longest years of work among employees being assigned 100 points, and the remaining years of work are converted into percentage scores for scoring;

[0019] Learning ability: based on the speed at which employees master new skills and the frequency of obtaining new knowledge certificates, the percentages are converted into a percentage system for quantitative scoring.

[0020] Combine these rating data for each employee into a feature vector Among them, x1 represents the professional skill score, x2 represents the knowledge reserve score, x3 represents the work experience score, and x4 represents the learning ability score, forming the employee ability data set i and j are feature vectors composed of two different employee data, and n is the number of employees.

[0021] Preferably, the model building unit specifically constructs the model process as follows:

[0022] Initial cluster center selection: randomly select a data point from the data set as the first initial cluster center c1;

[0023] For each data point in the dataset Calculate the minimum sum of squared distances from the selected cluster center The number of cluster centers selected initially is 1, and the formula is:

[0024]

[0025] in, express The Euclidean distance of

[0026] By probability Select the next cluster center:

[0027]

[0028] Repeat the above calculation until m initial cluster centers are selected, where m is the number of pre-set employee specialty categories;

[0029] Iterate the distribution and assign each data point Assign to the cluster to which the nearest cluster center belongs. The assignment rule is:

[0030]

[0031] in, Represents data points The identifier of the cluster to which it is assigned, represents the kth cluster center, and the value of k ranges from 1 to m;

[0032] Update the cluster center, for each cluster E k , where E k is the set of all data points assigned to the kth cluster center, and its cluster center is recalculated The formula is:

[0033]

[0034] Among them, |E k | is the number of data points in this cluster;

[0035] Repeat the above iterative allocation and update process of cluster centers until the cluster centers no longer change. It can also be determined that the iteration stops when a threshold is set or the maximum number of iterations is reached.

[0036] Category determination: Analyze the final clustering results and determine the category of employee expertise represented by each cluster based on the average scores of employees in each ability dimension and the score distribution characteristics;

[0037] Preferably, the model optimization unit uses the silhouette coefficient To evaluate the clustering effect, the formula is:

[0038]

[0039] in, It is a data point The average distance to other data points in the cluster to which it belongs, yes The average distance to data points in other clusters, the silhouette coefficient The closer the value is to 1, the better the clustering effect is; the closer it is to -1, the worse the clustering effect is; the closer it is to 0, the less obvious the boundaries between clusters are;

[0040] If the silhouette coefficient is not ideal, that is, it is between 0 and -1, different weights are assigned to the features of different ability dimensions when calculating the distance or the initial cluster center is reselected, and the algorithm is run again until the employee specialty classification result with a silhouette coefficient between 0 and 1 is obtained.

[0041] Preferably, the job requirement analysis module first uses Chinese word segmentation technology including rule-based, statistical or deep learning, including Jieba word segmentation and HanLP word segmentation, to process the job description, workflow documents and project case text materials within the enterprise, and divide the text into meaningful word blocks;

[0042] Next, use Stanford Part-of-Speech Tagger or Natural Language Toolkit NLTK to tag the part of speech for each word block. Then, use named entity recognition technology including Baidu Named Entity Recognition NER or Language Technology Platform LTP's NER function to identify key entities in the text, such as job titles, professional skills, knowledge areas, work experience descriptions, and learning ability.

[0043] On this basis, we use technologies including the Latent Dirichlet Allocation (LDA) model and the Hierarchical Dirichlet Process (HDP) model to mine potential topics in the text and extract key information related to job requirements. At the same time, we use word vector generation models such as Word2Vec and the global vector representation model GloVe to convert words in the text into vector form.

[0044] Preferably, the intelligent matching module uses a hybrid similarity calculation method based on a combination of cosine similarity and Manhattan distance to match employees with positions;

[0045] Assume that the number of employees is equal to the number of positions, and transform the employee ability evaluation vector and job requirements vector The key skills and ability requirements extracted through the job knowledge graph are quantified and compared, where y1 represents professional skill requirements, y2 represents knowledge reserve requirements, y3 represents work experience requirements, and y4 represents learning ability requirements;

[0046] First, calculate the cosine similarity sim cos :

[0047]

[0048] Among them, sim cos It is used to measure the consistency between the employee's ability evaluation vector and the job requirement vector in terms of direction. It focuses on whether the relative proportions of the employee's various abilities are similar to the relative proportions of the various abilities required for the job.

[0049] At the same time, calculate the Manhattan distance F man :

[0050]

[0051] Among them, F man It is used to reflect the absolute difference between the employee's ability evaluation vector and the job requirement vector in each dimension, focusing on the gap between the employee's specific ability and the job requirement;

[0052] Then, the comprehensive similarity is obtained by weighted average:

[0053]

[0054] Among them, α is the weight coefficient, which is determined according to the emphasis of the enterprise on the direction and degree of capability matching;

[0055] The comprehensive similarity data of employees and positions are constructed into a two-dimensional matrix, in which rows represent employees and columns represent positions. The elements in the matrix are the corresponding comprehensive similarity values. The objective function is set to maximize the sum of the comprehensive similarities of all employees and the assigned positions. At the same time, constraints are added: each employee can only be assigned to one position, and each position can only accept one employee. By inputting these parameters and constraints into the linear programming solver for solving, the optimal allocation plan for employees and positions is obtained.

[0056] Preferably, when comparing the work results with the expected performance goals, the display module uses the target achievement analysis method to compare the achievement ratio, and uses the quality deviation assessment method to compare the work quality; the work efficiency is compared using the time efficiency comparison method, and these comparison results are displayed to relevant managers in the form of intuitive charts including bar charts, line charts, pie charts and data reports.

[0057] Compared with the prior art, the present invention has the following beneficial effects:

[0058] 1. The present invention integrates and quantifies professional skills, knowledge reserves, work experience, learning ability and communication ability through the employee ability assessment module, uses improved clustering algorithm analysis to achieve accurate classification of employee capabilities, and determines employee specialty categories with objective data and scientific algorithms, providing accurate basis for job allocation and team building, achieving the effect of fine management and optimal allocation of human resources, and improving team collaboration efficiency.

[0059] 2. The present invention uses natural language processing technology to mine multiple types of text data through the job requirement analysis module to construct a job knowledge graph and clarify the key job requirements; it uses hybrid similarity calculation combined with linear programming to achieve accurate matching of employees and jobs, fully considers the adaptation relationship for global optimization, avoids mismatching, and achieves the effect of improving employee job fit, enhancing team collaboration fluency, improving enterprise operating efficiency and competitiveness, and releasing the maximum value of human resources. BRIEF DESCRIPTION OF THE DRAWINGS

[0060] Figure 1 It is a schematic diagram of the working process of the system of the present invention;

[0061] Figure 2 It is a schematic diagram of the calculation flow of the employee capability evaluation module of the present invention;

[0062] Figure 3 It is a schematic diagram of the calculation flow of the intelligent matching module of the present invention. DETAILED DESCRIPTION

[0063] The following embodiments of the present invention are described in further detail in conjunction with the accompanying drawings and examples. The following examples are used to illustrate the present invention, but are not intended to limit the scope of the present invention.

[0064] The present invention provides a human resource management system for job allocation based on personnel difference planning, comprising:

[0065] An employee competency assessment module, which is based on an improved clustering algorithm, collects and quantifies employee competency data, selects initial cluster centers, iterates clustering and determines categories, combines silhouette coefficient assessment and parameter adjustment, and implements employee competency classification to obtain employee competency assessment results;

[0066] Position demand analysis module, which establishes a position knowledge graph, analyzes the internal text materials of the enterprise including position descriptions, workflow documents and project cases through natural language processing technology, extracts key information, clarifies the professional skills, knowledge reserves, work experience and learning ability required for each position, and obtains the position demand analysis results;

[0067] An intelligent matching module, which is connected to the employee capability assessment module and the job requirement analysis module, and is used to match employees with jobs using an algorithm based on the employee capability assessment results and job requirement analysis data, and calculate the most appropriate job allocation plan;

[0068] The display module is connected to the data of the intelligent matching module and is used to regularly collect the performance data of employees in their positions, including information on work results, work quality and work efficiency, compare them with the expected performance goals, and display them to relevant managers for evaluating the effectiveness of the job allocation plan.

[0069] Example:

[0070] like Figure 1-Figure 3 As shown, in this embodiment, a large technology company undertook a complex software development project involving multiple technical fields, including front-end development, back-end architecture design, database management, and artificial intelligence algorithm application. At the beginning of the project, the human resources department adopted a job allocation human resources management system based on personnel difference planning in the present invention, and assigned different tasks to each employee based on the employee's job description and past general work experience.

[0071] First, the employee competency assessment module collects various competency data of project employees, including professional skills test scores, actual skill application results in past projects, educational background, relevant training courses attended, years of work experience, and ability to learn new knowledge.

[0072] And quantify and combine these scoring data of each employee into a feature vector Among them, x1 represents the professional skill score, x2 represents the knowledge reserve score, x3 represents the work experience score, and x4 represents the learning ability score, forming the employee ability data set i and j are feature vectors composed of two different employee data, and n is the number of employees.

[0073] The model building unit receives the quantified data and uses the improved clustering algorithm to first select the initial cluster center through a specific probability method, iteratively assigns the employee data to the corresponding cluster and updates the cluster center, and finally divides the employee ability categories according to the clustering results. The specific model building process of the model building unit is as follows:

[0074] Initial cluster center selection: randomly select a data point from the data set as the first initial cluster center c1;

[0075] For each data point in the dataset Calculate the minimum sum of squared distances from the selected cluster center The number of cluster centers selected initially is 1, and the formula is:

[0076]

[0077] in, express The Euclidean distance of

[0078] By probability Select the next cluster center:

[0079]

[0080] Repeat the above calculation until m initial cluster centers are selected, where m is the number of pre-set employee specialty categories;

[0081] Iterate the distribution and assign each data point Assign to the cluster to which the nearest cluster center belongs. The assignment rule is:

[0082]

[0083] in, Represents data points The identifier of the cluster to which it is assigned, represents the kth cluster center, and the value of k ranges from 1 to m;

[0084] Update the cluster center. For each cluster Ek, ​​where Ek is the set of all data points assigned to the kth cluster center, recalculate its cluster center. The formula is:

[0085]

[0086] Among them, |E k | is the number of data points in this cluster;

[0087] Repeat the above iterative allocation and update process of cluster centers until the cluster centers no longer change. It can also be determined that the iteration stops when a threshold is set or the maximum number of iterations is reached.

[0088] Category determination: Analyze the final clustering results and determine the category of employee expertise represented by each cluster based on the average scores of employees in each ability dimension and the score distribution characteristics;

[0089] Model optimization unit using silhouette coefficient To evaluate the clustering effect, the formula is:

[0090]

[0091] in, It is a data point The average distance to other data points in the cluster to which it belongs, yes The average distance to data points in other clusters, the silhouette coefficient The closer the value is to 1, the better the clustering effect is; the closer it is to -1, the worse the clustering effect is; the closer it is to 0, the less obvious the boundaries between clusters are;

[0092] If the silhouette coefficient is not ideal, that is, it is between 0 and -1, different weights are assigned to the features of different ability dimensions when calculating the distance or the initial cluster center is reselected, and the algorithm is run again until the employee specialty classification result with a silhouette coefficient between 0 and 1 is obtained.

[0093] Next, the job requirements analysis module uses Chinese word segmentation technologies based on rules, statistics or deep learning, including Jieba word segmentation and HanLP word segmentation, to process the company's internal job descriptions, workflow documents and project case text materials and segment the text into meaningful word blocks;

[0094] Next, use Stanford Part-of-Speech Tagger or Natural Language Toolkit NLTK to tag the part of speech for each word block. Then, use named entity recognition technology including Baidu Named Entity Recognition NER or Language Technology Platform LTP's NER function to identify key entities in the text, such as job titles, professional skills, knowledge areas, work experience descriptions, and learning ability.

[0095] On this basis, we use the latent Dirichlet allocation (LDA) model and the hierarchical Dirichlet process (HDP) model to mine potential topics in the text and extract key information related to job requirements. At the same time, we use the word vector generation model Word2Vec and the global vector representation model GloVe to convert the words in the text into vector form.

[0096] The intelligent matching module matches employees with positions using a hybrid similarity calculation method based on a combination of cosine similarity and Manhattan distance;

[0097] Assume that the number of employees is equal to the number of positions, and transform the employee ability evaluation vector and job requirements vector The key skills and ability requirements extracted through the job knowledge graph are quantified and compared, where y1 represents professional skill requirements, y2 represents knowledge reserve requirements, y3 represents work experience requirements, and y4 represents learning ability requirements;

[0098] First, calculate the cosine similarity sim cos :

[0099]

[0100] Among them, sim cos It is used to measure the consistency between the employee's ability evaluation vector and the job requirement vector in terms of direction. It focuses on whether the relative proportions of the employee's various abilities are similar to the relative proportions of the various abilities required for the job.

[0101] At the same time, calculate the Manhattan distance F man :

[0102]

[0103] Among them, F man It is used to reflect the absolute difference between the employee's ability evaluation vector and the job requirement vector in each dimension, focusing on the gap between the employee's specific ability and the job requirement;

[0104] Then, the comprehensive similarity is obtained by weighted average:

[0105]

[0106] Among them, α is the weight coefficient, which is determined according to the emphasis of the enterprise on the direction and degree of capability matching;

[0107] The comprehensive similarity data of employees and positions are constructed into a two-dimensional matrix, in which rows represent employees and columns represent positions. The elements in the matrix are the corresponding comprehensive similarity values. The objective function is set to maximize the sum of the comprehensive similarities of all employees and the assigned positions. At the same time, constraints are added: each employee can only be assigned to one position, and each position can only accept one employee. By inputting these parameters and constraints into the linear programming solver for solving, the optimal allocation plan for employees and positions is obtained.

[0108] For example, employees who are good at database performance tuning and have rich practical experience can be assigned to database management positions. Through this scientific and intelligent matching method, employees can maximize their value in the positions that best suit them, improve their job satisfaction, and thus improve work efficiency and quality, laying a solid foundation for the efficient advancement and successful delivery of projects.

[0109] During the project execution, the display module regularly collects employee performance data, uses the target achievement analysis method to compare work results to obtain the achievement ratio, uses the quality deviation assessment method to compare work quality, and uses the time efficiency comparison method to compare work efficiency. These comparison results are displayed to relevant managers in the form of intuitive charts, including bar charts, line charts, pie charts, and data reports.

[0110] The embodiments of the present invention are provided for the purpose of illustration and description. Although the embodiments of the present invention have been shown and described above, it can be understood that the above embodiments are exemplary and cannot be understood as limitations of the present invention. Ordinary technicians in this field can change, modify, replace and modify the above embodiments within the scope of the present invention.

Claims

1. A human resource management system for job allocation based on personnel difference planning, characterized in that: include: An employee competency assessment module, which is based on an improved clustering algorithm, collects and quantifies employee competency data, selects initial cluster centers, iterates clustering and determines categories, combines silhouette coefficient assessment and parameter adjustment, and implements employee competency classification to obtain employee competency assessment results; Position demand analysis module, which establishes a position knowledge graph, analyzes the internal text materials of the enterprise including position descriptions, workflow documents and project cases through natural language processing technology, extracts key information, clarifies the professional skills, knowledge reserves, work experience and learning ability required for each position, and obtains the position demand analysis results; An intelligent matching module, which is connected to the employee capability assessment module and the job requirement analysis module, and is used to match employees with jobs using an algorithm based on the employee capability assessment results and job requirement analysis data, and calculate the most appropriate job allocation plan; The display module is connected to the data of the intelligent matching module and is used to regularly collect the performance data of employees in their positions, including information on work results, work quality and work efficiency, compare them with the expected performance goals, and display them to relevant managers for evaluating the effectiveness of the job allocation plan.

2. A human resource management system for job allocation based on personnel difference planning as claimed in claim 1, characterized in that: The employee capability assessment module includes a data collection unit, a model building unit and a model optimization unit; The data collection unit collects employee capability data including professional skills, knowledge reserves, work experience, learning ability and communication ability, quantifies the data, and combines them into feature vectors to form an employee capability data set, which serves as the basis for the subsequent classification model; The model building unit uses an improved clustering algorithm to first select an initial clustering center by a specific probability method, iteratively assigns employee data to corresponding clusters and updates the clustering center, and finally divides employee ability categories according to the clustering results; The model optimization unit uses the silhouette coefficient to evaluate the clustering effect of the model establishment unit, thereby achieving effective division of the comprehensive ability differences of employees.

3. A human resource management system for job allocation based on personnel difference planning as claimed in claim 2, characterized in that: The data quantification standards of the data acquisition unit are as follows; Professional skills are quantified based on a percentage-based scoring mechanism, including professional skills test scores and the effectiveness of skill application in actual projects; knowledge reserves are assigned a percentage based on academic qualifications and training course completion; work experience is divided according to years of work, with the employee with the longest years of work among the employees corresponding to 100 points, and the remaining years of work are converted into percentage scores for scoring; Learning ability: based on the speed at which employees master new skills and the frequency of obtaining new knowledge certificates, the percentages are converted into a percentage system for quantitative scoring. Combine these rating data for each employee into a feature vector Among them, x1 represents the professional skill score, x2 represents the knowledge reserve score, x3 represents the work experience score, and x4 represents the learning ability score, forming the employee ability data set i and j are feature vectors composed of two different employee data, and n is the number of employees.

4. A human resource management system for job allocation based on personnel difference planning as claimed in claim 3, characterized in that: The specific model building process of the model building unit is as follows: Initial cluster center selection: randomly select a data point from the data set as the first initial cluster center c1; For each data point in the dataset Calculate the minimum sum of squared distances from the selected cluster center The number of cluster centers selected initially is 1, and the formula is: in, express The Euclidean distance of By probability Select the next cluster center: Repeat the above calculation until m initial cluster centers are selected, where m is the number of pre-set employee specialty categories; Iterate the distribution and assign each data point Assign to the cluster to which the nearest cluster center belongs. The assignment rule is: in, Represents data points The identifier of the cluster to which it is assigned, represents the kth cluster center, and the value of k ranges from 1 to m; Update the cluster center, for each cluster E k , where E k is the set of all data points assigned to the kth cluster center, and its cluster center is recalculated The formula is: Among them, |E k | is the number of data points in this cluster; Repeat the above process of iterative allocation and updating of cluster centers until the cluster centers no longer change. It can also be determined that the iteration stops when a threshold is set or the maximum number of iterations is reached. Category determination: Analyze the final clustering results and determine the employee specialty category represented by the cluster based on the average scores of employees in each cluster in various ability dimensions and the score distribution characteristics.

5. A human resource management system for job allocation based on personnel difference planning as claimed in claim 2, characterized in that: The model optimization unit uses the silhouette coefficient To evaluate the clustering effect, the formula is: in, It is a data point The average distance to other data points in the cluster to which it belongs, yes The average distance to data points in other clusters, the silhouette coefficient The closer the value is to 1, the better the clustering effect is; the closer it is to -1, the worse the clustering effect is; the closer it is to 0, the less obvious the boundaries between clusters are; If the silhouette coefficient is not ideal, that is, it is between 0 and -1, different weights are assigned to the features of different ability dimensions when calculating the distance or the initial cluster center is reselected, and the algorithm is run again until the employee specialty classification result with a silhouette coefficient between 0 and 1 is obtained.

6. A human resource management system for job allocation based on personnel difference planning as claimed in claim 1, characterized in that: The job requirement analysis module first uses Chinese word segmentation technologies based on rules, statistics or deep learning, including Jieba word segmentation and HanLP word segmentation, to process the job descriptions, workflow documents and project case text materials within the enterprise and divide the text into meaningful word blocks; Next, use Stanford Part-of-Speech Tagger or Natural Language Toolkit NLTK to tag the part of speech for each word block. Then, use named entity recognition technology including Baidu Named Entity Recognition NER or Language Technology Platform LTP's NER function to identify key entities in the text, such as job titles, professional skills, knowledge areas, work experience descriptions, and learning ability. On this basis, we use technologies including the Latent Dirichlet Allocation (LDA) model and the Hierarchical Dirichlet Process (HDP) model to mine potential topics in the text and extract key information related to job requirements. At the same time, we use word vector generation models such as Word2Vec and the global vector representation model GloVe to convert words in the text into vector form.

7. A human resource management system for job allocation based on personnel difference planning as claimed in claim 1, characterized in that: The intelligent matching module matches employees with positions by using a hybrid similarity calculation method based on a combination of cosine similarity and Manhattan distance; Assume that the number of employees is equal to the number of positions, and transform the employee ability evaluation vector and job requirements vector The key skills and ability requirements extracted through the job knowledge graph are quantified and compared, where y1 represents professional skill requirements, y2 represents knowledge reserve requirements, y3 represents work experience requirements, and y4 represents learning ability requirements; First, calculate the cosine similarity sim cos : Among them, sim cos It is used to measure the consistency between the employee's ability evaluation vector and the job requirement vector in terms of direction. It focuses on whether the relative proportions of the employee's various abilities are similar to the relative proportions of the various abilities required for the job. At the same time, calculate the Manhattan distance F man : Among them, F man It is used to reflect the absolute difference between the employee's ability evaluation vector and the job requirement vector in each dimension, focusing on the gap between the employee's specific ability and the job requirement; Then, the comprehensive similarity is obtained by weighted average: Among them, α is the weight coefficient, which is determined according to the emphasis of the enterprise on the direction and degree of capability matching; The comprehensive similarity data of employees and positions are constructed into a two-dimensional matrix, in which rows represent employees and columns represent positions. The elements in the matrix are the corresponding comprehensive similarity values. The objective function is set to maximize the sum of the comprehensive similarities of all employees and the assigned positions. At the same time, constraints are added: each employee can only be assigned to one position, and each position can only accept one employee. By inputting these parameters and constraints into the linear programming solver for solving, the optimal allocation plan for employees and positions is obtained.

8. A human resource management system for job allocation based on personnel difference planning as claimed in claim 1, characterized in that: When comparing work results with expected performance targets, the display module uses the target achievement analysis method to compare work results to obtain the achievement ratio, and uses the quality deviation assessment method to compare work quality; the work efficiency is compared using the time efficiency comparison method, and these comparison results are displayed to relevant managers in the form of intuitive charts, including bar charts, line charts, pie charts and data reports.

Images