A salary intelligent management method and system for the human resource industry

Abstract

The application discloses a kind of salary intelligent management method and system for human resources industry, it is related to data processing technical field, the method includes: abnormal screening processing is carried out to the salary data to be counted;Performance evaluation is carried out based on the salary data to be counted;Based on the salary data to be counted and performance evaluation result, salary accounting is carried out using fluctuation anomaly analysis, and target salary data is obtained;Based on target salary data, performance target correction data and work target correction data are obtained using salary influence analysis;The data attribute of the desensitization data in target salary data, performance target correction data and work target correction data is acquired to determine desensitization mode, and it is desensitized based on desensitization mode;The target salary data, performance target correction data and work target correction data are stored in storage node.The present application greatly improves the intelligent degree of salary management, and makes salary management achieve more ideal effect.

Description

Technical Field

[0001] This invention relates to the field of data processing technology, and in particular to an intelligent payroll management method and system for the human resources industry. Background Technology

[0002] With the increasing demand for labor and the rise of flexible employment models, the human resources industry is also constantly developing. For the human resources industry, personnel compensation management is a crucial task. Currently, in human resources compensation management, performance appraisals based on salary data are mostly conducted by relevant personnel according to analyzed data. However, this method relies too heavily on the professional competence of the personnel, and it is both costly and inefficient. Furthermore, current compensation management often involves only simple verification after compiling personnel salary data based on performance appraisal results and salary data, without considering anomalies in salary fluctuations, leading to a large error rate in compensation calculations. Moreover, for compensation management, adjusting corresponding performance and job content goals based on personnel compensation is also a critical aspect, impacting both the company's and individual's development. After obtaining compensation data and revised job target data, the data needs to be anonymized to protect the privacy of relevant personnel. Currently, data anonymization is often achieved through data perturbation methods, but this method lacks sufficient precision, failing to adequately protect personnel privacy and compromising the reliability of compensation management. Summary of the Invention

[0003] The purpose of this invention is to overcome the shortcomings of the prior art. This invention provides a method and system for intelligent salary management in the human resources industry, which greatly improves the level of intelligence in salary management and enables salary management to achieve more ideal results.

[0004] To address the aforementioned technical problems, this invention provides an intelligent payroll management method for the human resources industry, the method comprising:

[0005] Obtain the salary data to be statistically analyzed for each relevant person, perform anomaly filtering on the salary data to be statistically analyzed, and obtain the salary data to be statistically analyzed after anomaly filtering.

[0006] Performance evaluation is conducted based on the salary data to be statistically analyzed after anomaly removal, and the performance evaluation results are obtained.

[0007] Based on the salary data to be statistically analyzed after anomaly screening and the performance evaluation results, fluctuation anomaly analysis is used to calculate salary and obtain the target salary data for each relevant person.

[0008] Based on the target salary data, salary impact analysis is used to revise the performance goals and work goals of relevant personnel, and to obtain revised performance goal data and revised work goal data.

[0009] Obtain the data attributes of the data to be anonymized from the target salary data, performance target revision data, and work target revision data. Determine the anonymization method based on the data attributes and attribute feature analysis. Perform anonymization processing on the target salary data, performance target revision data, and work target revision data based on the anonymization method to obtain the anonymized target salary data, performance target revision data, and work target revision data.

[0010] The anonymized target salary data, performance target revision data, and work target revision data are stored in the corresponding storage nodes of the database.

[0011] Optionally, the step of performing anomaly filtering on the salary data to be statistically analyzed to obtain anomaly-filtered salary data to be statistically analyzed includes:

[0012] Obtain individual data anomaly rules and global data anomaly rules, and perform identification processing on the salary data to be counted based on the individual data anomaly rules and global data anomaly rules to obtain the identified salary data to be counted.

[0013] A binary tree is constructed, and anomaly removal processing is performed on the identified salary data to be statistically analyzed based on the binary tree to obtain the anomaly-removed salary data to be statistically analyzed.

[0014] Optionally, the step of performing anomaly filtering on the identified salary data based on the binary tree to obtain anomaly-filtered salary data includes:

[0015] The binary tree is converted into a balanced binary tree, and the training dataset is divided based on the balanced binary tree to obtain a normal training dataset and an abnormal training dataset.

[0016] The abnormal character recognition model is trained based on the normal training dataset and the abnormal training dataset to obtain a trained abnormal character recognition model.

[0017] Based on the trained abnormal character recognition model, abnormal characters are identified in the salary data to be statistically analyzed after the identification process, and abnormal character filtering is performed on the salary data to be statistically analyzed after abnormal character recognition, so as to obtain the salary data to be statistically analyzed after abnormal character filtering.

[0018] Optionally, the performance evaluation based on the salary data to be statistically analyzed after anomaly screening to obtain the performance evaluation results includes:

[0019] Extract working hour data, personnel cooperation data, and evaluation text data from the salary data to be statistically analyzed after anomaly screening;

[0020] Obtain the degree of correlation between the working hour data and performance data, and conduct an initial performance evaluation based on the degree of correlation to obtain the initial performance evaluation result;

[0021] The evaluation text data is segmented into words to obtain several segmented word groups, and a word vector matrix is ​​generated based on the several segmented word groups. The ability evaluation level is determined based on the word vector matrix.

[0022] Based on the aforementioned personnel cooperation data, a cooperation network model is used to analyze cooperation contribution and obtain cooperation contribution data.

[0023] The initial performance evaluation results are adjusted based on the aforementioned competency evaluation level and cooperation contribution data to obtain the final performance evaluation result.

[0024] Optionally, the salary calculation based on the anomaly-filtered salary data and performance evaluation results is performed using fluctuation anomaly analysis to obtain target salary data for each relevant individual, including:

[0025] Obtain the salary calculation rules for each relevant person, and perform salary statistics based on the salary calculation rules using the salary data to be statistically analyzed after anomaly screening and performance evaluation results to obtain the salary data for each relevant person.

[0026] The call address of the corresponding accounting engine is obtained based on the load balancing algorithm, and the salary data of each relevant person is transmitted to the corresponding accounting engine based on the call address.

[0027] The calculation engine obtains the same-level salary data of several peers corresponding to each relevant person, and performs data fluctuation characteristic analysis based on the same-level salary data to obtain fluctuation characteristic data.

[0028] Based on the fluctuation characteristic data, singularity identification is performed on the salary data to obtain singularity identification results;

[0029] Based on the singularity identification results and fluctuation characteristic data, the salary data of each relevant person is verified and processed to obtain the verification and processing results, and the target salary data of each relevant person is determined based on the verification and processing results.

[0030] Optionally, the step of using salary impact analysis based on the target salary data to revise the performance goals and work goals of relevant personnel, and obtaining revised performance goal data and revised work goal data, includes:

[0031] Acquire salary-related data, and based on the salary-related data and target salary data, conduct salary impact analysis using salary indicators and constraints to obtain salary impact analysis data;

[0032] Based on the salary impact analysis data, the value analysis model is used to calculate the job value, and the job value is compared with a preset threshold to obtain the comparison result;

[0033] Based on the comparison results, the performance goals and work goals of each relevant person are revised and analyzed to obtain revised performance goal data and revised work goal data.

[0034] Optionally, the step of obtaining the data attributes of the data to be anonymized from the target salary data, performance target revision data, and work target revision data, and determining the anonymization method based on the data attributes and attribute feature analysis, includes:

[0035] Acquire sensitive sample data and select the corresponding desensitization algorithm based on the data attributes of the sensitive sample data;

[0036] The corresponding desensitization algorithm is added as a desensitization method label to the sample sensitive data, and a desensitization method database is constructed based on the sample sensitive data after adding the desensitization method label;

[0037] Identify the data attributes of the data to be anonymized in the target salary data, performance target revision data, and work target revision data. Perform attribute feature analysis based on the data attributes of the data to be anonymized and the data in the anonymization method database to obtain the anonymization method label corresponding to the data to be anonymized. Determine the anonymization method of the data to be anonymized based on the corresponding anonymization method label.

[0038] Optionally, the step of anonymizing the target salary data, performance target revision data, and work target revision data based on the anonymization method to obtain anonymized target salary data, performance target revision data, and work target revision data includes:

[0039] Based on a programmable gate array comparator, mark the target location corresponding to the data to be de-identified in the target salary data, performance target correction data, and work target correction data;

[0040] Obtain the array of function stack pointers based on the target pointer, and then obtain the corresponding compact array of function name strings based on the array of function stack pointers;

[0041] The compact function name string array is converted into the corresponding standard function string based on the demangle function, and the target replacement character is obtained based on the corresponding standard function string;

[0042] Based on the aforementioned desensitization method, the target replacement character is used to perform character replacement processing at the target position corresponding to the data to be desensitized, thereby obtaining the desensitized target salary data, performance target correction data, and work target correction data.

[0043] Optionally, storing the anonymized target salary data, performance target revision data, and work target revision data into the corresponding storage node of the database includes:

[0044] Obtain the priority coefficient and encrypted digest value of each storage node in the database, and perform activity analysis on the anonymized target salary data, performance target correction data and work target correction data to obtain the corresponding activity information. Based on the activity information, priority coefficient and encrypted digest value of the node, determine the data storage strategy.

[0045] Based on the data storage strategy, the anonymized target salary data, performance target revision data, and work target revision data are stored in the corresponding storage nodes of the database.

[0046] In addition, the present invention also provides an intelligent payroll management system for the human resources industry, the system comprising:

[0047] Data anomaly removal module: used to obtain the salary data to be statistically analyzed for each relevant person, perform anomaly removal processing on the salary data to be statistically analyzed, and obtain the salary data to be statistically analyzed after anomaly removal processing.

[0048] Performance Appraisal Module: Used to perform performance appraisals based on the salary data to be statistically analyzed after anomaly screening, and obtain performance appraisal results;

[0049] Payroll Calculation Module: This module is used to perform payroll calculation based on the salary data to be statistically analyzed after anomaly screening and performance evaluation results, and to obtain the target salary data for each relevant person.

[0050] Target Revision Analysis Module: This module is used to revise the performance and work goals of relevant personnel based on the target salary data and using salary impact analysis, thereby obtaining revised performance and work goal data.

[0051] Data anonymization module: used to acquire the data attributes of the data to be anonymized in the target salary data, performance target revision data and work target revision data, determine the anonymization method based on the data attributes and attribute feature analysis, and perform anonymization processing on the target salary data, performance target revision data and work target revision data based on the anonymization method to obtain the anonymized target salary data, performance target revision data and work target revision data;

[0052] Data storage module: Used to store the anonymized target salary data, performance target revision data, and work target revision data into the corresponding storage nodes of the database.

[0053] In this embodiment of the invention, anomaly screening of the salary data to be statistically analyzed is performed using binary trees and data anomaly rules, which can more accurately remove abnormal data and ensure the accuracy of subsequent data statistical analysis. Performance evaluation based on the working hour data, personnel cooperation data, and evaluation text data in the anomaly-screened salary data yields more comprehensive performance evaluation results, avoiding the bias and limitations caused by a single factor. Using fluctuation anomaly analysis for salary calculation based on the anomaly-screened salary data and performance evaluation results, more accurate salary data can be obtained, significantly reducing the error rate of salary calculation. Using salary impact analysis based on target salary data to revise the performance and work goals of relevant personnel can optimize the rationality of personnel work, achieve closed-loop salary management for human resources, and improve the intelligence level of salary management. Determining the desensitization method based on data attributes and attribute feature analysis, and then desensitizing the target salary data, performance goal revision data, and work goal revision data based on the desensitization method, can improve the accuracy of data desensitization, protect relevant privacy data, and effectively improve the reliability of salary management. Attached Figure Description

[0054] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. 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.

[0055] Figure 1 This is a flowchart illustrating an intelligent payroll management method for the human resources industry according to an embodiment of the present invention.

[0056] Figure 2 This is a flowchart illustrating an intelligent payroll management method for the human resources industry according to another embodiment of the present invention.

[0057] Figure 3 This is a schematic diagram of the structural composition of an intelligent payroll management system for the human resources industry, as described in an embodiment of the present invention. Detailed Implementation

[0058] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0059] Example 1

[0060] Please see Figure 1 , Figure 1 This is a flowchart illustrating an intelligent payroll management method for the human resources industry according to an embodiment of the present invention. The method includes:

[0061] S11: Obtain the salary data to be statistically analyzed for each relevant person, perform anomaly filtering on the salary data to be statistically analyzed, and obtain the salary data to be statistically analyzed after anomaly filtering.

[0062] In a specific implementation of this invention, the step of performing anomaly screening on the salary data to be statistically analyzed to obtain anomaly-screened salary data includes: acquiring individual data anomaly rules and global data anomaly rules, and performing identification processing on the salary data to be statistically analyzed based on the individual data anomaly rules and global data anomaly rules to obtain identification-processed salary data to be statistically analyzed; establishing a binary tree, and performing anomaly screening on the identification-processed salary data to be statistically analyzed based on the binary tree to obtain anomaly-screened salary data to be statistically analyzed.

[0063] Furthermore, the step of performing anomaly filtering on the identified salary data based on the binary tree to obtain anomaly-filtered salary data includes: converting the binary tree into a balanced binary tree; dividing the training dataset based on the balanced binary tree to obtain a normal training dataset and an abnormal training dataset; training an abnormal character recognition model based on the normal training dataset and the abnormal training dataset to obtain a trained abnormal character recognition model; performing anomaly character recognition on the identified salary data based on the trained abnormal character recognition model to obtain anomaly-recognized salary data; and performing anomaly filtering on the anomaly-recognized salary data to obtain anomaly-filtered salary data.

[0064] Specifically, the process involves acquiring salary data for relevant personnel, including those dispatched by human resources companies through recruitment software and those recruited internally. This data includes working hours, collaboration data, evaluation text data, project commission data, and work completion data. Individual and global data anomaly rules are then obtained. Individual rules identify anomalies in a single data analysis object, while global rules identify anomalies in the overall data. These rules are set by administrators based on actual conditions. The salary data is then processed using these individual and global rules to obtain an anomaly-processed version. These rules help prevent errors in anomaly identification. Finally, a binary tree is constructed using a binary tree template. This binary tree is then converted into a balanced binary tree. An inorder traversal is performed to obtain an ordered array. A recursive method is then used to construct a balanced binary tree from this ordered array. The training dataset is partitioned based on the balanced binary tree. For each data point in the training dataset, the depth of each data point within the balanced binary tree is calculated, and a target score is calculated based on this depth. Based on the target score and a preset partitioning threshold, the training dataset is divided into a normal training dataset and an abnormal training dataset. An abnormal character recognition model is trained on these datasets. This abnormal character recognition model can employ a deep neural network model. The trained abnormal character recognition model is then used to identify abnormal characters in the processed salary data to be analyzed. Abnormal characters refer to characters in the dataset that do not conform to standards, such as special symbols or letters in numeric fields. Abnormal characters may cause significant deviations in the statistical analysis results. Anomaly removal processing is then performed on the salary data to be analyzed after abnormal character identification. Abnormal data refers to data that does not exhibit actual patterns. This anomaly removal process preserves valid data, ensuring the reliability of subsequent statistical analysis.

[0065] S12: Based on the salary data to be statistically analyzed after anomaly screening, conduct performance evaluation to obtain performance evaluation results;

[0066] In the specific implementation of this invention, the performance evaluation based on the salary data to be statistically analyzed after anomaly screening to obtain the performance evaluation result includes: extracting working hour data, personnel cooperation data, and evaluation text data from the salary data to be statistically analyzed after anomaly screening; obtaining the degree of correlation between the working hour data and the performance data, and performing an initial performance evaluation based on the degree of correlation to obtain an initial performance evaluation result; segmenting the evaluation text data to obtain several segmented word groups, generating a word vector matrix based on the several segmented word groups, and determining the ability evaluation level based on the word vector matrix; performing cooperation contribution analysis using a cooperation network model based on the personnel cooperation data to obtain cooperation contribution data; and adjusting the initial performance evaluation result based on the ability evaluation level and cooperation contribution data to obtain the performance evaluation result.

[0067] Specifically, after anomaly filtering, the salary data to be statistically analyzed is used to extract working hour data, personnel collaboration data, and evaluation text data. Working hour data includes normal working hours and overtime hours. Personnel collaboration data includes collaborative project information, personnel information, project module responsibility information, and completion information. Evaluation text data consists of management personnel's work evaluations of relevant personnel, containing phrases related to task completion, job responsibility performance, and employee development. The correlation between the working hour data and performance data is determined. Performance data includes performance evaluation standards and performance payment data. Based on the correlation, an initial performance evaluation is conducted. Preliminary performance evaluations are performed based on the correlation between normal working hours and overtime hours in the working hour data to obtain initial performance evaluation results. The evaluation text data is segmented into words. Coarse-grained segmentation is performed to obtain coarse-grained segmentation results, followed by fine-grained segmentation to obtain fine-grained segmentation results. The coarse-grained segments in the initial segmentation results are iterated to determine their character length. Based on the character length of each coarse-grained segment, the segmentation is corrected to obtain corrected coarse-grained segmentation results. The segmentation of the evaluation text data is determined based on the corrected and fine-grained segmentation results, resulting in several segmented word groups. A word vector matrix is ​​generated based on these word groups. Each word group is vectorized to obtain several word vectors. These word vectors are concatenated to obtain a word vector matrix. The ability evaluation level is determined based on this word vector matrix. The ability evaluation model is a convergent model obtained by training a deep neural network with the sample dataset. The ability evaluation levels include excellent and good assessment abilities. Based on the personnel cooperation data, a cooperation network model is used to analyze cooperation contribution. Personnel information is used as nodes, and project information as edges. Weights are assigned to each node according to preset rules, and the mutual influence weights between nodes are calculated based on these weights. A cooperation network model is constructed based on the nodes, edges, weights, and mutual influence weights. The cooperation contribution of each relevant person is calculated using the personnel cooperation data, thus obtaining the cooperation contribution data. The initial performance evaluation results are adjusted based on the ability evaluation level and cooperation contribution data. The ability evaluation level is matched with the corresponding ability performance evaluation level, and the cooperation contribution data is matched with the corresponding cooperation performance evaluation level. The initial performance evaluation results are then adjusted based on the ability performance evaluation level and cooperation performance evaluation level to obtain the final performance evaluation result. Generating the final performance evaluation result using initial performance evaluation based on work hour data, ability evaluation level, and cooperation contribution data improves the accuracy and comprehensiveness of the performance evaluation, avoids the bias and limitations of performance evaluation based on a single factor, and makes the obtained performance evaluation results more objective.

[0068] S13: Based on the salary data to be statistically analyzed after anomaly screening and the performance evaluation results, use fluctuation anomaly analysis to calculate salary and obtain the target salary data for each relevant person;

[0069] In the specific implementation of this invention, the step of using fluctuation anomaly analysis to perform salary calculation based on the salary data to be statistically analyzed after anomaly screening and performance evaluation results to obtain target salary data for each relevant person includes: obtaining the salary calculation rules corresponding to each relevant person; performing salary statistics based on the salary calculation rules using the salary data to be statistically analyzed after anomaly screening and performance evaluation results to obtain salary data for each relevant person; obtaining the call address of the adapted calculation engine based on the load balancing algorithm, and transmitting the salary data of each relevant person to the corresponding calculation engine based on the call address; the calculation engine obtaining the same-level salary data of several peers corresponding to each relevant person, and performing data fluctuation feature analysis based on the same-level salary data to obtain fluctuation feature data; identifying singularities in the salary data based on the fluctuation feature data to obtain singularity identification results; verifying the salary data of each relevant person based on the singularity identification results and fluctuation feature data to obtain verification processing results, and determining the target salary data for each relevant person based on the verification processing results.

[0070] Specifically, the system obtains the corresponding salary calculation rules for each relevant personnel. These rules include salary calculation for normal working hours, overtime hours, and commission-based pay for collaborative projects. Based on these rules, salary statistics are performed using the anomaly-filtered salary data and performance evaluation results. Preliminary salary data is calculated using the salary calculation rules based on the anomaly-filtered salary data. Performance-based pay for each relevant personnel is determined based on the performance evaluation results. The combination of these two factors yields the final salary data for each relevant personnel. The system also obtains the call address of the appropriate accounting engine based on a load balancing algorithm. The load balancer determines the corresponding accounting engine (the appropriate accounting engine) based on weighted round-robin and minimum connection count, and obtains its call address. Based on this call address, the salary data for each relevant personnel is transmitted to the corresponding accounting engine, effectively improving the execution efficiency of salary calculation. The calculation engine acquires the peer salary data of several peers corresponding to each relevant personnel. Peer personnel refer to employees with the same labor contract as the personnel conducting the salary verification. The peer salary data represents the actual salaries of these peers. Based on this peer salary data, data fluctuation characteristic analysis is performed, analyzing the mean, standard deviation, and variance of the peer salary data to form fluctuation characteristic data. Based on this fluctuation characteristic data, singularity identification is performed on the salary data. The salary data is decomposed using a non-negative matrix factorization algorithm to obtain the basic feature vector and corresponding weight coefficients. Multidimensional feature mapping data is generated using latent space feature mapping based on the basic feature vector and corresponding weight coefficients. Singularity identification is then performed on the multidimensional feature mapping data using the fluctuation characteristic data based on the local anomaly factor algorithm, obtaining singularities or outliers in the salary data. This singularity identification result can more accurately identify data points that clearly deviate from the conventional salary model. Based on the singularity identification results and fluctuation characteristic data, the salary data of each relevant person is verified. The error probability of salary statistics is determined according to the singularity identification results and fluctuation characteristic data. If the error probability is greater than or equal to a preset value, the salary statistics are repeated until the obtained error probability is less than the preset value. If the error probability is less than the preset value, the salary data of each relevant person is correct, and the verification result is obtained. Based on the verification result, the target salary data of each relevant person is determined, which can ensure the reliability of salary calculation.

[0071] S14: Based on the target salary data, use salary impact analysis to conduct a correction analysis of the performance goals and work goals of relevant personnel, and obtain performance goal correction data and work goal correction data;

[0072] In the specific implementation of this invention, the step of using salary impact analysis based on the target salary data to perform a correction analysis on the performance goals and work goals of relevant personnel, and obtaining performance goal correction data and work goal correction data, includes: acquiring salary correlation data; performing salary impact analysis based on the salary correlation data and the target salary data using salary indicators and constraints to obtain salary impact analysis data; calculating work value based on the salary impact analysis data using a value analysis model, and comparing the work value with a preset threshold to obtain a comparison result; and performing a correction analysis on the performance goals and work goals of relevant personnel based on the comparison result to obtain performance goal correction data and work goal correction data.

[0073] Specifically, the process involves acquiring salary-related data, including attendance information, performance settlement results, and valid overtime information. Based on this data and target salary data, a salary impact analysis is conducted using salary indicators and constraints. Salary indicators may include performance target percentages and project completion percentages. An objective function is constructed based on the salary indicators, and constraints are established based on the salary-related data. An impact analysis model is then built using the objective function and constraints. The cost data incurred by relevant personnel to achieve the target salary data is analyzed using the impact analysis model, thus obtaining salary impact analysis data. Based on this salary impact analysis data, a value analysis model is used to calculate the work value. The salary impact analysis data is input into the value analysis model to obtain the work value, which is the value obtained by achieving the salary target. This work value is then compared with a preset threshold to obtain the comparison result. Based on the comparison results, the performance goals and work goals of each relevant person are revised and analyzed. If the work value obtained is greater than or equal to the preset threshold, the performance goals and work goals can be appropriately increased according to the proportion exceeding the preset threshold. If the work value obtained is less than the preset threshold, the performance goals and work goals can be appropriately reduced according to the proportion below the preset threshold. The revised performance goal data and revised work goal data are obtained to ensure the rationality of the performance goals and work goals assigned to relevant persons.

[0074] S15: Obtain the data attributes of the data to be anonymized in the target salary data, performance target revision data, and work target revision data; determine the anonymization method based on the data attributes and attribute feature analysis; and perform anonymization processing on the target salary data, performance target revision data, and work target revision data based on the anonymization method to obtain the anonymized target salary data, performance target revision data, and work target revision data.

[0075] In the specific implementation of this invention, the step of obtaining the data attributes of the data to be anonymized in the target salary data, performance target revision data, and work target revision data, and determining the anonymization method based on the data attributes and attribute feature analysis, includes: obtaining sample sensitive data, and selecting a corresponding anonymization algorithm based on the data attributes of the sample sensitive data; adding the corresponding anonymization algorithm as an anonymization method label to the sample sensitive data, and constructing an anonymization method database based on the sample sensitive data after adding the anonymization method label; identifying the data attributes of the data to be anonymized in the target salary data, performance target revision data, and work target revision data, performing attribute feature analysis based on the data attributes of the data to be anonymized and the data in the anonymization method database, obtaining the anonymization method label corresponding to the data to be anonymized, and determining the anonymization method of the data to be anonymized based on the corresponding anonymization method label.

[0076] Furthermore, the step of performing anonymization processing on the target salary data, performance target correction data, and work target correction data based on the anonymization method to obtain anonymized target salary data, performance target correction data, and work target correction data includes: marking the target positions corresponding to the data to be anonymized in the target salary data, performance target correction data, and work target correction data using a programmable gate array comparator; obtaining a function stack pointer array based on the target pointer, and obtaining a corresponding compact function name string array based on the function stack pointer array; converting the compact function name string array into a corresponding standard function string based on the demangle function, and obtaining a target replacement character based on the corresponding standard function string; and performing character replacement processing at the target positions corresponding to the data to be anonymized using the target replacement character based on the anonymization method to obtain anonymized target salary data, performance target correction data, and work target correction data.

[0077] Specifically, sensitive sample data is acquired, and corresponding de-identification algorithms are selected based on the data attributes of the sensitive sample data. Data attributes can be key fields of the data, such as performance, total compensation, and work goals. De-identification algorithms include data encryption, privacy data masking, and transformation of random functions. The corresponding de-identification algorithm is added as a de-identification method label to the sensitive sample data, and a de-identification method database is constructed based on the sensitive sample data with added de-identification method labels. The data attributes of the data to be de-identified in the target compensation data, performance target correction data, and work goal correction data are identified. Attribute feature analysis is performed based on the data attributes of the data to be de-identified and the data in the de-identification method database. Feature value matching is performed in the de-identification method database according to the data attributes of the data to be de-identified. The de-identification method label is determined based on the matched feature values, obtaining the de-identification method label corresponding to the data to be de-identified. The de-identification method for the data to be de-identified is determined based on the corresponding de-identification method label. The target position corresponding to the data to be de-identified is marked in the target compensation data, performance target correction data, and work goal correction data using a programmable gate array comparator. A target pointer can be determined at the target location. Based on the target pointer, a function stack pointer array is obtained, and then a corresponding compact function name string array is obtained from the function stack pointer array. The relevant arrays are matched against the database. The compact function name string array is converted into a corresponding standard function string using the demangle function. The demangle function can be used to parse character names. Based on the corresponding standard function string, the target replacement character is obtained. The target replacement character is determined based on the corresponding standard function string and the desensitization method. Based on the desensitization method, the target replacement character is used to perform character replacement processing at the target position corresponding to the data to be desensitized. If the desensitization method is irreversible, character replacement processing is used; if the desensitization method is reversible, encryption processing is used. This yields the desensitized target salary data, performance target correction data, and work target correction data. Data desensitization is an important part of salary information management in human resources companies. For sensitive information, such as salary and performance, reasonable masking is required. Therefore, desensitizing target salary data, performance target correction data, and work target correction data can protect the privacy of relevant personnel and prevent data leakage.

[0078] S16: Store the anonymized target salary data, performance target revision data, and work target revision data into the corresponding storage nodes of the database.

[0079] In the specific implementation of this invention, storing the anonymized target salary data, performance target revision data, and work target revision data into the corresponding storage nodes of the database includes: obtaining the priority coefficient and node encrypted digest value of each storage node in the database, performing activity analysis on the anonymized target salary data, performance target revision data, and work target revision data to obtain corresponding activity information, determining a data storage strategy based on the activity information, priority coefficient, and node encrypted digest value, and storing the anonymized target salary data, performance target revision data, and work target revision data into the corresponding storage nodes of the database based on the data storage strategy.

[0080] Specifically, the priority coefficient and node cryptographic digest value of each storage node in the database are obtained; the node identifier and node capacity of each storage node are obtained; and the node cryptographic digest value is generated based on the node identifier and node capacity. The number of data reads, the number of data reconstruction reads, the average data read latency, and the bandwidth of each storage node are obtained; and the priority coefficient of each storage node is calculated based on the number of data reads, the number of data reconstruction reads, the average data read latency, and the bandwidth. The expression for calculating the priority coefficient is as follows:

[0081] y=(a*ω a +b*ω b +c*ω c +d*ω d )σ,

[0082] Where y is the priority coefficient, a is the number of data reads, and ω is the number of data reads. a Here, b represents the weight corresponding to the number of data reads, and ω represents the number of data reconstruction reads. b The weights corresponding to the number of data reconstruction reads, where c is the average data read latency, and ω is the weight corresponding to the number of data reads. c The weights corresponding to the average data read latency are d, where d is the bandwidth and ω is the weight. d σ represents the weight corresponding to bandwidth, and σ is the correction coefficient. Activity level analysis is performed on the anonymized target salary data, performance target correction data, and work target correction data. Based on the data access frequency judgment rule and creation time judgment rule in the preset time sequence rules, and Apache Doris technology, the activity level analysis of the anonymized target salary data, performance target correction data, and work target correction data is obtained to obtain corresponding activity level information. Based on the activity level information, priority coefficient, and node cryptographic digest value, a data storage strategy is determined, that is, the required storage nodes are determined according to the activity level information, priority coefficient, and node cryptographic digest value. Based on the data storage strategy, the anonymized target salary data, performance target correction data, and work target correction data are stored in the corresponding storage nodes of the database, ensuring the reliability of data storage.

[0083] In this embodiment of the invention, anomaly screening of the salary data to be statistically analyzed is performed using binary trees and data anomaly rules, which can more accurately remove abnormal data and ensure the accuracy of subsequent data statistical analysis. Performance evaluation based on the working hour data, personnel cooperation data, and evaluation text data in the anomaly-screened salary data yields more comprehensive performance evaluation results, avoiding the bias and limitations caused by a single factor. Using fluctuation anomaly analysis for salary calculation based on the anomaly-screened salary data and performance evaluation results, more accurate salary data can be obtained, significantly reducing the error rate of salary calculation. Using salary impact analysis based on target salary data to revise the performance and work goals of relevant personnel can optimize the rationality of personnel work, achieve closed-loop salary management for human resources, and improve the intelligence level of salary management. Determining the desensitization method based on data attributes and attribute feature analysis, and then desensitizing the target salary data, performance goal revision data, and work goal revision data based on the desensitization method, can improve the accuracy of data desensitization, protect relevant privacy data, and effectively improve the reliability of salary management.

[0084] Example 2

[0085] Please see Figure 2 , Figure 2 This is a flowchart illustrating an intelligent payroll management method for the human resources industry according to another embodiment of the present invention, the method comprising:

[0086] S201: Obtain the salary data to be statistically analyzed for each relevant person, perform anomaly filtering on the salary data to be statistically analyzed, and obtain the salary data to be statistically analyzed after anomaly filtering.

[0087] S202: Performance evaluation is conducted based on the salary data to be statistically analyzed after anomaly screening, and the performance evaluation results are obtained;

[0088] S203: Obtain the salary calculation rules for each relevant person, and perform salary statistics based on the salary calculation rules using the salary data to be counted after anomaly screening and performance evaluation results to obtain the salary data for each relevant person;

[0089] S204: Obtain the call address of the adapted accounting engine based on the load balancing algorithm, and transmit the salary data of each relevant person to the corresponding accounting engine based on the call address;

[0090] S205: The calculation engine obtains the same-level salary data of several peers corresponding to each relevant person, and performs data fluctuation characteristic analysis based on the same-level salary data to obtain fluctuation characteristic data;

[0091] S206: Based on the fluctuation characteristic data, perform singularity identification on the salary data to obtain singularity identification results;

[0092] S207: Based on the singularity identification results and fluctuation characteristic data, the salary data of each relevant person is checked and processed, and the error probability of salary statistics is determined according to the singularity identification results and fluctuation characteristic data.

[0093] S208: Determine whether the error probability is greater than or equal to a preset threshold;

[0094] S209: Determine the target salary data for each relevant person based on the verification and processing results;

[0095] S210: Based on the target salary data, use salary impact analysis to conduct a correction analysis of the performance goals and work goals of relevant personnel, and obtain performance goal correction data and work goal correction data;

[0096] S211: Obtain the data attributes of the data to be anonymized in the target salary data, performance target revision data, and work target revision data; determine the anonymization method based on the data attributes and attribute feature analysis; and perform anonymization processing on the target salary data, performance target revision data, and work target revision data based on the anonymization method to obtain the anonymized target salary data, performance target revision data, and work target revision data.

[0097] S212: Store the anonymized target salary data, performance target revision data, and work target revision data into the corresponding storage nodes of the database.

[0098] In this embodiment of the invention, anomaly screening of the salary data to be statistically analyzed is performed using binary trees and data anomaly rules, which can more accurately remove abnormal data and ensure the accuracy of subsequent data statistical analysis. Performance evaluation based on the working hour data, personnel cooperation data, and evaluation text data in the anomaly-screened salary data yields more comprehensive performance evaluation results, avoiding the bias and limitations caused by a single factor. Using fluctuation anomaly analysis for salary calculation based on the anomaly-screened salary data and performance evaluation results, more accurate salary data can be obtained, significantly reducing the error rate of salary calculation. Using salary impact analysis based on target salary data to revise the performance and work goals of relevant personnel can optimize the rationality of personnel work, achieve closed-loop salary management for human resources, and improve the intelligence level of salary management. Determining the desensitization method based on data attributes and attribute feature analysis, and then desensitizing the target salary data, performance goal revision data, and work goal revision data based on the desensitization method, can improve the accuracy of data desensitization, protect relevant privacy data, and effectively improve the reliability of salary management.

[0099] Example 3

[0100] Please see Figure 3 , Figure 3 This is a schematic diagram of the structural composition of an intelligent payroll management system for the human resources industry, as described in this embodiment of the invention. The system includes:

[0101] Data anomaly removal module 31: used to obtain the salary data to be statistically analyzed for each relevant person, perform anomaly removal processing on the salary data to be statistically analyzed, and obtain the salary data to be statistically analyzed after anomaly removal processing.

[0102] Performance Appraisal Module 32: Used to perform performance appraisals based on the salary data to be statistically analyzed after anomaly screening, and to obtain performance appraisal results;

[0103] Payroll Calculation Module 33: This module is used to perform payroll calculation based on the salary data to be statistically analyzed after anomaly screening and performance evaluation results, and to obtain the target salary data for each relevant person.

[0104] Target Revision Analysis Module 34: Used to perform revision analysis on the performance targets and work targets of relevant personnel based on the target salary data and using salary impact analysis, to obtain performance target revision data and work target revision data;

[0105] Data anonymization module 35: used to acquire the data attributes of the data to be anonymized in the target salary data, performance target revision data and work target revision data, determine the anonymization method based on the data attributes and attribute feature analysis, and perform anonymization processing on the target salary data, performance target revision data and work target revision data based on the anonymization method to obtain the anonymized target salary data, performance target revision data and work target revision data;

[0106] Data storage module 36: Used to store the anonymized target salary data, performance target revision data, and work target revision data into the corresponding storage nodes of the database.

[0107] In the specific implementation of this invention, the specific implementation methods of the system items can be referred to the above embodiments, and will not be repeated here.

[0108] In this embodiment of the invention, anomaly screening of the salary data to be statistically analyzed is performed using binary trees and data anomaly rules, which can more accurately remove abnormal data and ensure the accuracy of subsequent data statistical analysis. Performance evaluation based on the working hour data, personnel cooperation data, and evaluation text data in the anomaly-screened salary data yields more comprehensive performance evaluation results, avoiding the bias and limitations caused by a single factor. Using fluctuation anomaly analysis for salary calculation based on the anomaly-screened salary data and performance evaluation results, more accurate salary data can be obtained, significantly reducing the error rate of salary calculation. Using salary impact analysis based on target salary data to revise the performance and work goals of relevant personnel can optimize the rationality of personnel work, achieve closed-loop salary management for human resources, and improve the intelligence level of salary management. Determining the desensitization method based on data attributes and attribute feature analysis, and then desensitizing the target salary data, performance goal revision data, and work goal revision data based on the desensitization method, can improve the accuracy of data desensitization, protect relevant privacy data, and effectively improve the reliability of salary management.

[0109] Those skilled in the art will understand that all or part of the steps in the various methods of the above embodiments can be implemented by a program instructing related hardware. The program can be stored in a computer-readable storage medium, which may include: read-only memory (ROM), random access memory (RAM), disk or optical disk, etc.

[0110] Furthermore, the above provides a detailed description of an intelligent salary management method and system for the human resources industry provided by the embodiments of the present invention. Specific examples have been used to illustrate the principles and implementation methods of the present invention. The description of the above embodiments is only for the purpose of helping to understand the method and core ideas of the present invention. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of the present invention. Therefore, the content of this specification should not be construed as a limitation of the present invention.

Claims

1. A method for intelligent payroll management in the human resources industry, characterized in that, The method includes: Obtain the salary data to be statistically analyzed for each relevant person, perform anomaly filtering on the salary data to be statistically analyzed, and obtain the salary data to be statistically analyzed after anomaly filtering. Performance evaluation is conducted based on the salary data to be statistically analyzed after anomaly removal, and the performance evaluation results are obtained. Based on the salary data to be statistically analyzed after anomaly screening and the performance evaluation results, fluctuation anomaly analysis is used to calculate salary and obtain the target salary data for each relevant person. Based on the target salary data, salary impact analysis is used to revise the performance goals and work goals of relevant personnel, and to obtain revised performance goal data and revised work goal data. Obtain the data attributes of the data to be anonymized from the target salary data, performance target revision data, and work target revision data. Determine the anonymization method based on the data attributes and attribute feature analysis. Perform anonymization processing on the target salary data, performance target revision data, and work target revision data based on the anonymization method to obtain the anonymized target salary data, performance target revision data, and work target revision data. The anonymized target salary data, performance target revision data, and work target revision data are stored in the corresponding storage nodes of the database; The process involves using fluctuation anomaly analysis to calculate salary data for relevant personnel based on the anomaly-filtered salary data and performance evaluation results, thereby obtaining target salary data for each individual. This includes: acquiring salary calculation rules for each relevant individual; performing salary statistics based on these rules using the anomaly-filtered salary data and performance evaluation results to obtain salary data for each relevant individual; obtaining the call address of the appropriate calculation engine based on a load balancing algorithm; and transmitting the salary data for each relevant individual to the corresponding calculation engine based on these call addresses. The calculation engine then acquires the corresponding salary data for several peers of each relevant individual and performs calculations based on this peer salary data. Data fluctuation characteristic analysis is performed to obtain fluctuation characteristic data. Salary data is decomposed using a non-negative matrix factorization algorithm to obtain basic feature vectors and corresponding weight coefficients. Multidimensional feature mapping data is generated using latent space feature mapping based on the basic feature vectors and corresponding weight coefficients. Singularity identification is performed on the multidimensional feature mapping data using the fluctuation characteristic data and a local anomaly factor algorithm to obtain singularities or outliers in the salary data, thus obtaining singularity identification results. Based on the singularity identification results and fluctuation characteristic data, the salary data of relevant personnel is verified to obtain verification results. Target salary data for each relevant personnel is then determined based on the verification results. The performance evaluation is conducted based on the salary data to be statistically analyzed after anomaly screening, and the following steps are taken: 1) Extracting working hour data, personnel collaboration data, and evaluation text data from the salary data to be statistically analyzed after anomaly screening; 2) Obtaining the correlation between the working hour data and performance data, and conducting an initial performance evaluation based on the correlation to obtain an initial performance evaluation result; 3) Segmenting the evaluation text data to obtain several segmented word groups, generating a word vector matrix based on these word groups, and determining the competency evaluation level based on the word vector matrix; 4) Assigning weights to each node according to preset rules, using personnel information as nodes and project information as edges, and calculating the mutual influence weights between nodes based on their weights; 5) Constructing a collaboration network model based on nodes, edges, weights, and mutual influence weights; 6) Analyzing the collaboration contribution based on the personnel collaboration data using the collaboration network model to obtain collaboration contribution data; 7) Adjusting the initial performance evaluation result based on the competency evaluation level and collaboration contribution data to obtain the final performance evaluation result. The anonymized target salary data, performance target revision data, and work target revision data are stored in the corresponding storage nodes of the database. This includes: obtaining the node encryption digest value of each storage node in the database; calculating the priority coefficient of each storage node based on the number of data reads, the number of data reconstruction reads, the average data read latency, and the bandwidth; the expression for calculating the priority coefficient is as follows: ， Where y is the priority coefficient and a is the number of data reads. Here, b represents the weight corresponding to the number of data reads, and b represents the number of data reconstruction reads. The weights corresponding to the number of data reconstruction reads are assigned, and c is the average data read latency. The weights corresponding to the average data read latency are d, where d is the bandwidth. The weights corresponding to the bandwidth. To correct the coefficients, and to analyze the activity level of the anonymized target salary data, performance target correction data, and work target correction data, corresponding activity level information is obtained. Based on the activity level information, priority coefficients, and node encryption digest values, a data storage strategy is determined. Based on the data storage strategy, the anonymized target salary data, performance target correction data, and work target correction data are stored in the corresponding storage nodes of the database.

2. The intelligent payroll management method for the human resources industry according to claim 1, characterized in that, The step of performing anomaly filtering on the salary data to be statistically analyzed, to obtain anomaly-filtered salary data to be statistically analyzed, includes: Obtain individual data anomaly rules and global data anomaly rules, and perform identification processing on the salary data to be counted based on the individual data anomaly rules and global data anomaly rules to obtain the identified salary data to be counted. A binary tree is constructed, and anomaly removal processing is performed on the identified salary data to be statistically analyzed based on the binary tree to obtain the anomaly-removed salary data to be statistically analyzed.

3. The intelligent payroll management method for the human resources industry according to claim 2, characterized in that, The step of performing anomaly filtering on the identified salary data based on the binary tree to obtain anomaly-filtered salary data includes: The binary tree is converted into a balanced binary tree, and the training dataset is divided based on the balanced binary tree to obtain a normal training dataset and an abnormal training dataset. The abnormal character recognition model is trained based on the normal training dataset and the abnormal training dataset to obtain a trained abnormal character recognition model. Based on the trained abnormal character recognition model, abnormal characters are identified in the salary data to be statistically analyzed after the identification process, and abnormal character filtering is performed on the salary data to be statistically analyzed after abnormal character recognition, so as to obtain the salary data to be statistically analyzed after abnormal character filtering.

4. The intelligent payroll management method for the human resources industry according to claim 1, characterized in that, The process involves using salary impact analysis based on the target salary data to revise the performance and work goals of relevant personnel, obtaining revised performance and work goal data, including: Acquire salary-related data, and based on the salary-related data and target salary data, conduct salary impact analysis using salary indicators and constraints to obtain salary impact analysis data; Based on the salary impact analysis data, the value analysis model is used to calculate the job value, and the job value is compared with a preset threshold to obtain the comparison result; Based on the comparison results, the performance goals and work goals of each relevant person are revised and analyzed to obtain revised performance goal data and revised work goal data.

5. The intelligent payroll management method for the human resources industry according to claim 1, characterized in that, The process of acquiring the data attributes of the data to be anonymized from the target salary data, performance target revision data, and work target revision data, and determining the anonymization method based on the data attributes and attribute feature analysis, includes: Acquire sensitive sample data and select the corresponding desensitization algorithm based on the data attributes of the sensitive sample data; The corresponding desensitization algorithm is added as a desensitization method label to the sample sensitive data, and a desensitization method database is constructed based on the sample sensitive data after adding the desensitization method label; Identify the data attributes of the data to be anonymized in the target salary data, performance target revision data, and work target revision data. Perform attribute feature analysis based on the data attributes of the data to be anonymized and the data in the anonymization method database to obtain the anonymization method label corresponding to the data to be anonymized. Determine the anonymization method of the data to be anonymized based on the corresponding anonymization method label.

6. The intelligent payroll management method for the human resources industry according to claim 1, characterized in that, The process of anonymizing the target salary data, performance target revision data, and work target revision data based on the aforementioned anonymization method to obtain anonymized target salary data, performance target revision data, and work target revision data includes: Based on a programmable gate array comparator, mark the target location corresponding to the data to be de-identified in the target salary data, performance target correction data, and work target correction data; Obtain the array of function stack pointers based on the target pointer, and then obtain the corresponding compact array of function name strings based on the array of function stack pointers; The compact function name string array is converted into the corresponding standard function string based on the demangle function, and the target replacement character is obtained based on the corresponding standard function string; Based on the aforementioned desensitization method, the target replacement character is used to perform character replacement processing at the target position corresponding to the data to be desensitized, thereby obtaining the desensitized target salary data, performance target correction data, and work target correction data.

7. An intelligent payroll management system for the human resources industry, characterized in that, The system includes: Data anomaly removal module: used to obtain the salary data to be statistically analyzed for each relevant person, perform anomaly removal processing on the salary data to be statistically analyzed, and obtain the salary data to be statistically analyzed after anomaly removal processing. Performance Appraisal Module: Used to perform performance appraisals based on the salary data to be statistically analyzed after anomaly screening, and obtain performance appraisal results; Payroll Calculation Module: This module is used to perform payroll calculation based on the salary data to be statistically analyzed after anomaly screening and performance evaluation results, and to obtain the target salary data for each relevant person. Target Revision Analysis Module: This module is used to revise the performance and work goals of relevant personnel based on the target salary data and using salary impact analysis, thereby obtaining revised performance and work goal data. Data anonymization module: used to acquire the data attributes of the data to be anonymized in the target salary data, performance target revision data and work target revision data, determine the anonymization method based on the data attributes and attribute feature analysis, and perform anonymization processing on the target salary data, performance target revision data and work target revision data based on the anonymization method to obtain the anonymized target salary data, performance target revision data and work target revision data; Data storage module: Used to store the anonymized target salary data, performance target revision data, and work target revision data into the corresponding storage nodes of the database; The process involves using fluctuation anomaly analysis to calculate salary data for relevant personnel based on the anomaly-filtered salary data and performance evaluation results, thereby obtaining target salary data for each individual. This includes: acquiring salary calculation rules for each relevant individual; performing salary statistics based on these rules using the anomaly-filtered salary data and performance evaluation results to obtain salary data for each relevant individual; obtaining the call address of the appropriate calculation engine based on a load balancing algorithm; and transmitting the salary data for each relevant individual to the corresponding calculation engine based on these call addresses. The calculation engine then acquires the corresponding salary data for several peers of each relevant individual and performs calculations based on this peer salary data. Data fluctuation characteristic analysis is performed to obtain fluctuation characteristic data. Salary data is decomposed using a non-negative matrix factorization algorithm to obtain basic feature vectors and corresponding weight coefficients. Multidimensional feature mapping data is generated using latent space feature mapping based on the basic feature vectors and corresponding weight coefficients. Singularity identification is performed on the multidimensional feature mapping data using the fluctuation characteristic data and a local anomaly factor algorithm to obtain singularities or outliers in the salary data, thus obtaining singularity identification results. Based on the singularity identification results and fluctuation characteristic data, the salary data of relevant personnel is verified to obtain verification results. Target salary data for each relevant personnel is then determined based on the verification results. The performance evaluation based on the salary data to be statistically analyzed after anomaly screening, to obtain the performance evaluation result, includes: extracting working hour data, personnel cooperation data, and evaluation text data from the salary data to be statistically analyzed after anomaly screening; obtaining the correlation between the working hour data and performance data, and performing an initial performance evaluation based on the correlation to obtain an initial performance evaluation result; segmenting the evaluation text data to obtain several segmented word groups, generating a word vector matrix based on the segmented word groups, and determining the ability evaluation level based on the word vector matrix; assigning weights to each node according to preset rules, using personnel information as nodes and project information as edges, calculating the mutual influence weights between nodes based on the weights of each node, constructing a cooperation network model based on nodes, edges, weights, and mutual influence weights, performing cooperation contribution analysis based on the personnel cooperation data using the cooperation network model to obtain cooperation contribution data; and adjusting the initial performance evaluation result based on the ability evaluation level and cooperation contribution data to obtain the final performance evaluation result. The step of storing the anonymized target salary data, performance target revision data, and work target revision data into the corresponding storage nodes of the database includes: obtaining the node encryption digest value of each storage node in the database; calculating the priority coefficient of each storage node based on the number of data reads, the number of data reconstruction reads, the average data read latency, and the bandwidth; the expression for calculating the priority coefficient is as follows: ， Where y is the priority coefficient and a is the number of data reads. Here, b represents the weight corresponding to the number of data reads, and b represents the number of data reconstruction reads. The weights corresponding to the number of data reconstruction reads are assigned, and c is the average data read latency. The weights corresponding to the average data read latency are d, where d is the bandwidth. The weights corresponding to the bandwidth. To correct the coefficients, and to analyze the activity level of the anonymized target salary data, performance target correction data, and work target correction data, corresponding activity level information is obtained. Based on the activity level information, priority coefficients, and node encryption digest values, a data storage strategy is determined. Based on the data storage strategy, the anonymized target salary data, performance target correction data, and work target correction data are stored in the corresponding storage nodes of the database.

Images