Human resource data visual management method and system based on semantic analysis technology

Abstract

The invention discloses a human resource data visual management method and system based on a semantic analysis technology, and relates to the technical field of human resource management.The method comprises the steps that digital production data and text production data of employees are collected and analyzed, and employee evaluation indexes are determined; performing normalization processing on the digital production data and the text production data, calculating expected output, and determining an employee ability evaluation level; constructing a neural network evaluation model, and training the neural network evaluation model by using the digital production data and the text production data until the model converges; and evaluating the employee ability by using the converged neural network evaluation model, and outputting an employee evaluation level. According to the method, the digital production data and the text production data of the employees are collected at the same time, preprocessing and semantic analysis are carried out, dominant and implicit information in the human resource data can be more comprehensively mined, and the problem that key information is difficult to rapidly and accurately extract in a traditional data processing mode is solved.

Description

Technical Field

[0001] This invention relates to the field of human resource management technology, and in particular to a human resource data visualization management method and system based on semantic analysis technology. Background Technology

[0002] As businesses expand and their operations become more complex, human resource management faces the challenge of processing and analyzing massive amounts of data. Traditional data processing methods struggle to extract key information quickly and accurately, leading to inefficient decision-making. Semantic analysis technology can deeply understand and analyze textual data to extract valuable information, while data visualization technology can present complex data in an intuitive way, facilitating rapid understanding and decision-making for managers.

[0003] Among the existing publicly available technologies, such as the patent application publication number CN119887139A, an intelligent information screening method and system for human resources systems are disclosed. By using a two-way semantic matching network to intelligently screen candidates and by using a blockchain communication link to achieve two-way recommendation, it can automatically process massive recruitment data and select applicants who meet the requirements of enterprises.

[0004] The aforementioned existing technologies have the drawback of only being able to perform data analysis based on applicants' resumes, and cannot comprehensively evaluate the company's internal human resources data to be applicable to scenarios such as employee satisfaction analysis, turnover intention prediction, and competency assessment. Summary of the Invention

[0005] The technical problem to be solved by the present invention is to overcome the shortcomings of the prior art and provide a human resources data visualization management method and system based on semantic analysis technology.

[0006] To solve the above technical problems, the technical solution of the present invention is as follows: A human resources data visualization management method based on semantic analysis technology includes: Collect digital and text production data from employees, analyze the collected production data, and determine employee evaluation indicators. The digital production data and text production data are normalized, and the expected output of the digital production data and text production data is calculated to determine the employee's competency evaluation level. Construct a neural network evaluation model and train it using the digital production data and text production data until the model converges. The converged neural network evaluation model is used to evaluate employee capabilities and output employee evaluation levels.

[0007] As a preferred embodiment of the human resource data visualization management method based on semantic analysis technology described in this invention, the step of collecting employee digital production data and text production data, and analyzing the collected production data to determine employee evaluation indicators includes: The collected digital production data of employees is preprocessed, and the following are used as explicit evaluation indicators for employees: company knowledge, professional and technical knowledge, engineering category, efficiency, equipment management ability, problem-solving ability, on-time delivery ability, organizational and coordination ability, discipline, innovation, and objective contribution. The collected employee text production data is preprocessed, and semantic analysis is performed on the preprocessed text production data to determine the implicit evaluation indicators of employees. Calculate the correlation coefficient between evaluation indicators and determine employee evaluation indicators based on the correlation coefficient.

[0008] As a preferred embodiment of the human resource data visualization management method based on semantic analysis technology described in this invention, the semantic analysis of the preprocessed text production data includes: Scan employees' personal evaluation reports and perform word frequency statistics on the reports to extract keywords as hidden rating indicators; Construct a sentiment lexicon and establish text semantic matching rules to calculate the sentiment value of employee personal evaluation reports; Based on the percentage of each employee's contribution to the team, the sentiment value of the entire report is allocated to each employee in the team, resulting in a data value for each employee under the indicator, and the contribution of each employee is calculated based on the data value.

[0009] As a preferred embodiment of the human resource data visualization management method based on semantic analysis technology described in this invention, the calculation of the correlation coefficient between evaluation indicators includes: The correlation coefficient between the evaluation indicators is calculated using Formula 1, which is: ,in, This represents the value of the k-th employee under the i-th indicator. This represents the average value of the i-th indicator. This represents the correlation coefficient between each pair of indicators.

[0010] As a preferred embodiment of the human resource data visualization management method based on semantic analysis technology described in this invention, the employee evaluation indicators include knowledge indicators, ability indicators, and professional quality indicators.

[0011] As a preferred embodiment of the human resource data visualization management method based on semantic analysis technology described in this invention, the normalization processing of the digital production data and text production data includes: The digital production data and text production data are normalized using Formula 2, which is: ,in, This is the average of the data. denoted as the standard deviation of the data.

[0012] As a preferred embodiment of the human resource data visualization management method based on semantic analysis technology described in this invention, the step of constructing a neural network evaluation model and training it using the digital production data and text production data until the model converges includes: Construct a neural network evaluation model, which includes an input layer, an output layer, and an intermediate layer; 70% of the digital production data and text production data were used as training data, and the remaining 30% were used as test data to train the neural network evaluation model. The output of the neural network model is compared with the expected output, the error between the model's output and the expected output is calculated, and training stops when the error value is less than the preset maximum error value.

[0013] As a preferred embodiment of the human resource data visualization management method based on semantic analysis technology described in this invention, the number of output layer nodes is 1, and the formula for calculating the number of intermediate layer nodes is: , where k is the number of neurons in the intermediate layer, m is the number of neurons in the input layer, n is the number of neurons in the output layer, and a is a constant between 1 and 10.

[0014] This invention also provides a human resources data visualization management system based on semantic analysis technology, comprising: The data acquisition and processing module is used to collect digital and text production data from employees, analyze the collected production data, and determine employee evaluation indicators. The evaluation level determination module is used to normalize digital production data and text production data, calculate the expected output of the digital production data and text production data, and determine the employee's ability evaluation level. The model building module is used to build a neural network evaluation model and train it using the digital production data and text production data until the model converges. The employee evaluation module is used to evaluate employee capabilities using a converged neural network evaluation model and output employee evaluation levels.

[0015] The beneficial effects of this invention are: (1) By simultaneously collecting employees’ digital production data and text production data and performing preprocessing and semantic analysis, this invention can more comprehensively mine explicit and implicit information in human resources data, avoid the limitations of analyzing only employee resume data, solve the problem that traditional data processing methods cannot quickly and accurately extract key information, improve the efficiency of enterprises in processing human resources data, and provide a richer and more accurate data foundation for subsequent decision-making.

[0016] (2) The neural network evaluation model constructed in this invention can effectively learn the complex relationships and patterns in employee data through the reasonable design of the input layer, intermediate layer and output layer and the use of a large amount of data for training and testing, so as to evaluate the employee's ability more accurately. The model is continuously optimized by comparing the error between the model output and the expected output, ensuring the reliability of the evaluation results and providing strong support for enterprise human resource decision-making.

[0017] (3) Based on the different evaluation levels of employees, the present invention generates corresponding development direction suggestions, providing specific employee training and management references for enterprise managers. It helps enterprises to formulate personalized incentive, training and promotion plans for different employees, fully tap the potential of employees, improve employee satisfaction and loyalty, and thus improve the overall human resource management level and competitiveness of enterprises. Attached Figure Description

[0018] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0019] Figure 1 This is a flowchart illustrating the human resource data visualization management method based on semantic analysis technology provided by the present invention. Detailed Implementation

[0020] To make the content of this invention easier to understand, the invention will be further described in detail below with reference to specific embodiments and accompanying drawings.

[0021] This application provides a human resource data visualization management method based on semantic analysis technology, which specifically includes the following steps: Step S101: Collect employee digital production data and text production data, analyze the collected production data, and determine employee evaluation indicators.

[0022] Specifically, the operation process for this step is as follows: Step S101a: Perform data preprocessing on the collected employee digital production data, and use company knowledge, professional technical knowledge, engineering category, efficiency, equipment management ability, problem-solving ability, on-time delivery ability, organizational and coordination ability, discipline, innovation, and objective contribution as explicit evaluation indicators for employees.

[0023] Step S101b: Perform data preprocessing on the collected employee text production data, and perform semantic analysis on the preprocessed text production data to determine the implicit evaluation indicators for employees.

[0024] The semantic analysis steps for preprocessed text production data are as follows: First, scan employee personal evaluation reports and perform word frequency statistics to extract keywords as implicit rating indicators. Next, construct a sentiment lexicon and establish text semantic matching rules to calculate the sentiment score of each employee's personal evaluation report. Finally, based on the employee's percentage contribution to the team, allocate the sentiment score of the entire report to each employee in the team, obtaining the data value for each employee under the indicator, and calculate each employee's contribution based on the data value.

[0025] Step S101c: Calculate the correlation coefficient between the evaluation indicators and determine the employee evaluation indicators based on the correlation coefficient.

[0026] Specifically, the correlation coefficient between the evaluation indicators is calculated using Formula 1, which is: ,in, This represents the value of the k-th employee under the i-th indicator. This represents the average value of the i-th indicator. This represents the correlation coefficient between each pair of indicators.

[0027] Based on the results of correlation calculations, employee evaluation indicators are divided into: knowledge indicators, ability indicators, and professional quality indicators.

[0028] In addition to using the correlation calculation method mentioned above, the effectiveness of evaluation indicators can also be calculated by their discriminative power. A larger coefficient of variation indicates a stronger discriminative power and a greater impact on the results; conversely, a smaller coefficient of variation indicates a weaker discriminative power and a smaller impact on the results. The calculation formula is as follows: ,in, The standard deviation of each indicator is . This represents the average value of each indicator.

[0029] Evaluation indicator systems are generally divided into primary indicators: knowledge, skills / abilities, and professional ethics; and secondary indicators: company knowledge, professional technical knowledge, engineering category, efficiency, equipment management ability, problem-solving ability, on-time delivery ability, product quality, product lifespan, product performance, discipline, innovation, sense of responsibility, initiative, and teamwork. By selecting primary and secondary indicators, both explicit and implicit abilities of employees can be evaluated.

[0030] Step S102: Normalize the digital production data and text production data, calculate the expected output of the digital production data and text production data, and determine the employee competency evaluation level.

[0031] Specifically, the method for normalizing data is as follows: The digital production data and text production data are normalized using Formula 2, which is: ,in, This is the average of the data. denoted as the standard deviation of the data.

[0032] Besides the methods mentioned above, another method can be used for data normalization, and its calculation method is as follows: ,in, For normalized data, For the original data, The minimum value in the original data. This is the maximum value in the original data.

[0033] After calculating the expected output of digital and text production data, the evaluation levels are divided into five categories based on employee capabilities: Excellent, Good, Average, Pass, and Fail. Each level is assigned a specific numerical range, as follows: Excellent ([0.9, 1.0]), Good ([0.8, 0.9]), Average ([0.7, 0.8]), Pass ([0.6, 0.7]), and Fail ([0, 0.6]).

[0034] Step S103: Construct a neural network evaluation model and train it using the digital production data and text production data until the model converges.

[0035] Specifically, the operation process for this step is as follows: Step S103a: Construct a neural network evaluation model. The constructed neural network evaluation model includes an input layer, an output layer, and intermediate layers. The number of nodes in the input layer is determined by the number of knowledge indicators, ability indicators, and professional competence indicators. The output layer has one node. The formula for calculating the number of nodes in the intermediate layer is: , where k is the number of neurons in the intermediate layer, m is the number of neurons in the input layer, n is the number of neurons in the output layer, and a is a constant between 1 and 10.

[0036] In addition to the methods mentioned above, a trial-and-error method can also be used to calculate the number of intermediate layer nodes in a neural network. The trial-and-error method is based on the empirical formula method. That is, it first uses the empirical formula method to calculate the range of neurons that meet the conditions, and then starts training from the smallest number of neurons within that range. After training the smallest number of neurons, the number of neurons is increased sequentially until the largest neuron in the range is trained. During the trial-and-error method, it is essential to keep the training data and other network parameters constant to ensure the accuracy of the results. During training, the number of iterations required to meet the training requirements and the mean squared error of the network at that point are recorded. After all neurons within the range have been trained, the number of iterations and the mean squared error are compared and analyzed. The number of neurons with the smaller number of iterations and network error is then selected as the number of neurons in the intermediate layer of the model.

[0037] Before using the trial-and-error method, the number of intermediate layers in the neural network should be calculated using a formula. With 15 input layers and 1 output layer, the number of intermediate layer nodes in this invention ranges from 5 to 14. Then, while ensuring that the sample data and network parameters remain unchanged, the trial-and-error method is used to train each of the 5 to 14 neurons. The optimal number of intermediate layer neurons is selected by comprehensively considering the mean square error of network training with different numbers of intermediate layer neurons and the number of iterations.

[0038] The transfer function, also known as the activation function, has different effects depending on the specific function. Common transfer functions in backpropagation (BP) neural networks can be categorized into three types: linear transfer functions, nonlinear transfer functions, and threshold transfer functions. This invention selects the Tansig transfer function between the input and intermediate layers, the Purelin function for the output layer, and the trainlm function for training.

[0039] In the training process of a neural network, the accurate selection of the number of intermediate layer neurons is crucial for the successful establishment of the neural network model. This application uses a formula to determine that the number of intermediate layer neurons in the neural network model should be between 5 and 14. Then, a trial-and-error method is used to determine the optimal number of intermediate layer neurons. Using the same sample data, maintaining a maximum training iteration count of 10,000, a minimum training error of 0.00001, and a learning rate of 0.51, intermediate layer neurons between 5 and 14 are trained sequentially. After training, the mean squared error and number of iterations are compared for different numbers of intermediate layer neurons, and the number of intermediate layer neurons with fewer iterations and smaller network error is selected.

[0040] Step S103b: Use 70% of the digital production data and text production data as training data and the remaining 30% as test data to train the neural network evaluation model; Step S103c: Compare the output of the neural network model with the expected output, calculate the error between the model's output and the expected output, and stop training when the error value is less than the preset maximum error value.

[0041] Step S104: Use the converged neural network evaluation model to evaluate the employee's ability and output the employee evaluation level.

[0042] It should be noted that the model is considered accurate when the accuracy of the employee evaluation rating and the expected output reaches 85%.

[0043] For employees whose evaluation results are excellent, the suggestion is to provide material and spiritual rewards to recognize and reward their work, which can be used as a basis for the selection and appointment of cadres.

[0044] For employees whose evaluation results are good, the suggestion generated is that the company should help employees set goals, fully tap their work potential, and promote their growth and all-round development.

[0045] For employees whose evaluation results are average, the recommended approach is for the company to improve its compensation mechanism, optimize its performance appraisal mechanism, and enhance employee enthusiasm and initiative.

[0046] For employees whose evaluation results are satisfactory, suggestions are generated such as providing training for the employees, conducting targeted training on their weaknesses, and introducing mentorship to correct deviations in a timely manner.

[0047] For employees whose evaluation results are unsatisfactory, the suggestion generated is that employees of this level have significant deficiencies in knowledge, skills, abilities, and professional qualities.

[0048] In addition, this application also provides a human resources data visualization management system based on semantic analysis technology. The system includes: a data acquisition and processing module, an evaluation level determination module, a model building module, and an employee evaluation module.

[0049] Specifically, the data acquisition and processing module is used to collect employees' digital and text production data, analyze the collected production data, and determine employee evaluation indicators.

[0050] The evaluation level determination module is used to normalize digital production data and text production data, calculate the expected output of the digital production data and text production data, and determine the employee's competence evaluation level.

[0051] The model building module is used to build a neural network evaluation model and train it using the digital production data and text production data until the model converges.

[0052] The employee evaluation module is used to evaluate employee capabilities using a converged neural network evaluation model and output employee evaluation levels.

[0053] Therefore, the technical solution of this application, by simultaneously collecting employees' digital production data and text production data, and performing preprocessing and semantic analysis, can more comprehensively mine explicit and implicit information in human resource data, avoiding the limitations of analyzing only employee resume data, solving the problem that traditional data processing methods are unable to quickly and accurately extract key information, improving the efficiency of enterprises in processing human resource data, and providing a richer and more accurate data foundation for subsequent decision-making.

[0054] In addition to the above embodiments, the present invention may have other implementation methods; all technical solutions formed by equivalent substitution or equivalent transformation fall within the protection scope claimed by the present invention.

Claims

1. A human resource data visualization management method based on semantic analysis technology, characterized in that: include: Collect digital and text production data from employees, analyze the collected production data, and determine employee evaluation indicators. The digital production data and text production data are normalized, and the expected output of the digital production data and text production data is calculated to determine the employee's competency evaluation level. Construct a neural network evaluation model and train it using the digital production data and text production data until the model converges. The converged neural network evaluation model is used to evaluate employee capabilities and output employee evaluation levels.

2. The human resource data visualization management method based on semantic analysis technology according to claim 1, characterized in that: The process of collecting employee digital and text production data, analyzing the collected production data, and determining employee evaluation indicators includes: The collected digital production data of employees is preprocessed, and the following are used as explicit evaluation indicators for employees: company knowledge, professional and technical knowledge, engineering category, efficiency, equipment management ability, problem-solving ability, on-time delivery ability, organizational and coordination ability, discipline, innovation, and objective contribution. The collected employee text production data is preprocessed, and semantic analysis is performed on the preprocessed text production data to determine the implicit evaluation indicators of employees. Calculate the correlation coefficient between evaluation indicators and determine employee evaluation indicators based on the correlation coefficient.

3. The human resource data visualization management method based on semantic analysis technology according to claim 2, characterized in that: The semantic analysis of the preprocessed text production data includes: Scan employees' personal evaluation reports and perform word frequency statistics on the reports to extract keywords as hidden rating indicators; Construct a sentiment lexicon and establish text semantic matching rules to calculate the sentiment value of employee personal evaluation reports; Based on the percentage of each employee's contribution to the team, the sentiment value of the entire report is allocated to each employee in the team, resulting in a data value for each employee under the indicator, and the contribution of each employee is calculated based on the data value.

4. The human resource data visualization management method based on semantic analysis technology according to claim 2, characterized in that: The correlation coefficients between the calculated evaluation indicators include: The correlation coefficient between explicit and implicit evaluation indicators is calculated using Formula 1, which is: ,in, This represents the value of the k-th employee under the i-th indicator. This represents the average value of the i-th indicator. This represents the correlation coefficient between each pair of indicators.

5. The human resource data visualization management method based on semantic analysis technology according to claim 2, characterized in that: The employee evaluation indicators include knowledge indicators, ability indicators, and professional quality indicators.

6. The human resource data visualization management method based on semantic analysis technology according to claim 1, characterized in that: Normalization of the digital production data and text production data includes: The digital production data and text production data are normalized using Formula 2, which is: ,in, This is the average of the data. denoted as the standard deviation of the data.

7. The human resource data visualization management method based on semantic analysis technology according to claim 1, characterized in that: The step of constructing a neural network evaluation model and training it using the digital production data and text production data until the model converges includes: Construct a neural network evaluation model, which includes an input layer, an output layer, and an intermediate layer; 70% of the digital production data and text production data were used as training data, and the remaining 30% were used as test data to train the neural network evaluation model. The output of the neural network model is compared with the expected output, the error between the model's output and the expected output is calculated, and training stops when the error value is less than the preset maximum error value.

8. The human resource data visualization management method based on semantic analysis technology according to claim 7, characterized in that: The number of output layer nodes is 1, and the formula for calculating the number of intermediate layer nodes is: , where k is the number of neurons in the intermediate layer, m is the number of neurons in the input layer, n is the number of neurons in the output layer, and a is a constant between 1 and 10.

9. A human resources data visualization management system based on semantic analysis technology, characterized in that: include: The data acquisition and processing module is used to collect digital and text production data from employees, analyze the collected production data, and determine employee evaluation indicators. The evaluation level determination module is used to normalize digital production data and text production data, calculate the expected output of the digital production data and text production data, and determine the employee's ability evaluation level. The model building module is used to build a neural network evaluation model and train it using the digital production data and text production data until the model converges. The employee evaluation module is used to evaluate employee capabilities using a converged neural network evaluation model and output employee evaluation levels.

Images