A data sharing processing method and system applied to human resources

Abstract

The application discloses a data sharing processing method and system applied to human resources, relates to the technical field of data processing, and comprises the following steps: collecting multi-source data, pre-processing the data, classifying and sharing the data, and storing the data in a public storage area and an authorized access area; managing roles and access permissions of the stored data, performing data sharing compliance auditing, dynamically adjusting access permissions based on an access environment, storing permission change records through a Fabric alliance chain, performing abnormality detection on real-time access data and executing security strategies, and performing bidirectional matching of job seekers and enterprises based on the stored data. The application not only effectively improves the sharing efficiency and security of human resource data, but also enhances the scalability of data storage and the intelligence of permission management, is suitable for application scenarios such as enterprise human resource management, a recruitment platform and government data sharing, and greatly improves the value and application potential of human resource data.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of data processing, in particular to a data sharing processing method and system applied to human resources. BACKGROUND

[0002] With the acceleration of enterprise digital transformation, the demand for data-driven in the field of human resource management (HRM) is increasing. Modern human resource management systems (HRMS) are widely used in recruitment, performance evaluation, salary management, employee development planning and other aspects, and use big data, artificial intelligence (AI), blockchain and other technologies to optimize data management and decision-making processes. Among them, the research on human resource data sharing technology mainly focuses on data storage, secure sharing, intelligent matching and other aspects, so that human resource management develops towards more accurate, more efficient and more secure direction.

[0003] However, although a variety of advanced technologies have been applied to human resource data management, existing systems still face many challenges. For example, the heterogeneity of data sources leads to non-uniform data formats, affecting data analysis and sharing efficiency; privacy protection and compliance requirements are increasingly stringent, especially under the constraints of GDPR, CCPA and other data protection laws, enterprises need to ensure the security and compliance of data in the sharing process. In addition, the lack of dynamic permission management based on access environment makes it difficult to adjust permissions in real time according to the user's network environment, device characteristics and access behavior, increasing the risk of data leakage. SUMMARY

[0004] In view of the above existing problems, the present application is proposed.

[0005] Therefore, the present application provides a data sharing processing method and system applied to human resources, which solves the problem that enterprises need to ensure the security and compliance of data in the sharing process. In addition, the lack of dynamic permission management based on access environment makes it difficult to adjust permissions in real time according to the user's network environment, device characteristics and access behavior, increasing the risk of data leakage.

[0006] To solve the above technical problems, the present application provides the following technical solutions:

[0007] In a first aspect, the present application provides a data sharing processing method applied to human resources, which comprises,

[0008] After collecting and preprocessing multi-source data, the data is classified and shared for storage, including a public storage area and an authorized access area;

[0009] The multi-source data includes recruitment data, employee data, salary and welfare data and performance data;

[0010] The stored data is managed in terms of roles and access rights, and data sharing compliance audit is performed;

[0011] The access rights are dynamically adjusted based on the access environment, the permission change records are stored through the Fabric alliance chain, the real-time access data is used for abnormal detection and security policy execution;

[0012] Based on the stored data, the job seekers and enterprises are matched in two directions, a matching ranking table is generated for real-time pushing, and the data is stored in the database.

[0013] As a preferred scheme of the data sharing processing method applied to human resources, wherein: the dynamic adjustment of access rights based on the access environment comprises:

[0014] When the visitor logs in the system, the device hardware information is automatically obtained;

[0015] Based on the device hardware information, high-trust devices, medium-trust devices and low-trust devices, high-trust IP, medium-trust IP and low-trust IP, and high-trust network, medium-trust network and low-trust network are obtained;

[0016] The collected device, IP and network information is integrated into the access environment database;

[0017] Based on the trust degree marks of the device, IP and network information in the environment database, the corresponding trust degree scores are set, and based on the trust degree scores, the final security score of the user access environment is calculated using a weighting method;

[0018] The user identity information is obtained, the RBAC rule is used to allocate default access rights to the user according to the role type of the user, and different permissions are mapped to numerical values;

[0019] The user's access behavior data is collected, the access time score, the location score and the access frequency score are calculated, and the calculated scores are normalized and then weighted to obtain the environment adaptability factor;

[0020] The user access right score A is calculated based on the permission, security score and environment adaptability factor, the score is compared with the preset threshold, and the high-level permission, the medium-level permission and the low-level permission are obtained;

[0021] All user sets U, role category sets R and permission sets P in the enterprise are defined;

[0022] The mapping of roles and permissions is established based on the sets using a role permission mapping matrix;

[0023] The matching degree S of the role r on the permission p is calculated, the coverage degree Z of the role r on the permission set P is calculated, and the management cost c caused by the role adjustment is measured.

[0024] The statistical role permission change log is optimized by using the non-dominant sorting genetic algorithm to balance the role coverage and the role change cost, and a Pareto optimal solution set is obtained;

[0025] The Pareto optimal solution set is traversed to obtain the role r and the associated user u and the permission p, the frequency of user access p is counted, the frequency is compared with the preset threshold I, if the frequency is greater than or equal to the threshold I, it is a high access frequency user, then the permission is preferentially allocated, otherwise it is a low access frequency user, and the access is limited;

[0026] The Q-learning access reward value is calculated to measure the optimization effect of different role-permission allocation schemes, the Q-learning update formula is defined, and the reward function E is defined;

[0027] The calculated Q-learning update value is stored in the reinforcement learning optimization database, and based on the Q-learning calculation result, the cross point with the highest Q value is selected as the target point of the genetic algorithm cross operation;

[0028] Starting from the cross point calculated by Q-learning, the best access permission allocation scheme is searched forward, starting from the tail of the role permission allocation scheme, the role allocation scheme with the highest matching degree is searched, and the final permission allocation scheme is generated combining the results of forward search and reverse search.

[0029] As a preferred scheme of the data sharing processing method applied to human resources, the method comprises the following steps:

[0030] The skill keywords in the resume of the job seeker and the skill keywords in the job position description of the enterprise are extracted by using TF-IDF, and the matching degree of the skills of the job seeker and the job requirements is calculated by using cosine similarity;

[0031] The experience years of the job requirements and the actual experience years of the job seeker are extracted for rule matching to obtain the experience matching degree;

[0032] The expected salary of the job seeker and the salary provided by the enterprise are matched according to rules to obtain the salary matching degree;

[0033] The distribution probability of the job seeker in the past post is calculated based on Bayesian classification, and the post matching degree is calculated according to the category of the current post;

[0034] The enterprise recruitment priority is obtained based on the decision tree and the average recruitment time of the industry;

[0035] The applicant adaptation degree is obtained by weighted calculation of the skill matching degree, experience matching degree and salary expectation matching degree of the job seeker, and the enterprise adaptation degree is obtained by weighted calculation of the post demand matching degree and enterprise recruitment priority;

[0036] The final matching score is obtained by mean calculation of the job seeker adaptation degree and the enterprise adaptation degree, the job seekers are sorted in descending order according to the matching score, a matching list is generated and pushed to the enterprise HR.

[0037] As a preferred scheme of the data sharing processing method applied to human resources, the data is classified and shared and stored, including:

[0038] The index tree is used for hierarchical management of data, each piece of data is indexed according to the data category, the Kubernetes+Hadoop HDFS is used for distributed storage, and the hot and cold data hierarchical storage is used.

[0039] As a preferred scheme of the data sharing processing method applied to human resources, the role and access permission management of the stored data is performed, and the data sharing compliance audit is executed, including:

[0040] According to the role of the user, the basic access permission is granted, the data sharing request submitted by the user is analyzed, the data category is identified, the zero-knowledge proof is used when sharing data, and the data is filtered according to the sharing party;

[0041] The evidence of all data sharing requests is stored, the Merkle Tree technology is used to calculate the hash of the data sharing record, and the Fabric block chain is used to store the hash value of the shared data.

[0042] As a preferred scheme of the data sharing processing method applied to human resources, the abnormality detection according to the real-time access data and the execution of the security strategy include that the user access behavior data is collected by using the log analysis, the access log is stored by using the time series database, the historical access time series of the user is established by using the time series analysis, the average value and the access frequency standard deviation of the historical access of the user are calculated, and the access behavior deviation M is calculated.

[0043] The abnormal threshold N and are set, and N< If M is less than the threshold N, it is low risk, the abnormal log is recorded, and the user continues to access, if M is less than the threshold and greater than the threshold N, it is medium risk, manual audit is performed, if M is greater than or equal to the threshold , it is high risk, the access permission is limited, the account is frozen, and the administrator is notified.

[0044] As a preferred embodiment of the data sharing and processing method for human resources described in this invention, the step of storing data in the database refers to storing the collected real-time data, permission adjustment data, and talent recommendation data in the database, marking the stored data with timestamps, and uploading the stored data to the cloud for backup.

[0045] Secondly, the present invention provides a data sharing and processing system for human resources, comprising,

[0046] The data acquisition module is used to collect data from multiple sources and perform data cleaning and standardization.

[0047] The access control module is used to assign initial user permissions and employs zero-knowledge proofs and Fabric notarization for data compliance management.

[0048] The permission adjustment module is used to automatically adjust access permissions based on the user's access environment and detect abnormal behavior;

[0049] The matching module is responsible for the two-way matching of job seekers and companies, and generates a matching ranking table.

[0050] The data storage module is used to store matching data, permission adjustment data, and shared logs, and supports cloud backup.

[0051] Thirdly, the present invention provides a computer device including a memory and a processor, wherein the memory stores a computer program, wherein when the computer program is executed by the processor, it implements any step of the data sharing processing method for human resources as described in the first aspect of the present invention.

[0052] Fourthly, the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein: when the computer program is executed by a processor, it implements any step of the data sharing processing method for human resources as described in the first aspect of the present invention.

[0053] The application has the advantages that the privacy protection of data sharing is ensured through zero-knowledge proof and Merkle Tree hash storage, fine-grained permission management is realized by combining role-based and attribute-based access control, all data access behaviors are recorded based on the Fabric alliance chain storage to ensure the security and compliance of shared data, the multi-factor authentication mechanism of SHA-256 device identification calculation, GeoIP positioning and Wi-Fi BSSID analysis is introduced, dynamic permission adjustment, time series analysis and Q-learning reinforcement learning are combined to realize dynamic adjustment of permissions based on access environment, improve the accuracy and response speed of abnormal access detection, and avoid the risk of data leakage caused by static access control mode. In the aspect of intelligent matching of talents, the algorithms such as TF-IDF skill keyword extraction, cosine similarity calculation, Bayesian classification post adaptation degree and decision tree recruitment priority calculation are combined to accurately match job seekers and enterprise post demands, and a matching ranking table is generated to realize real-time push, thereby improving the recruitment efficiency. BRIEF DESCRIPTION OF DRAWINGS

[0054] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings needed in the embodiment description will be briefly introduced. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0055] Fig. 1 The flowchart of the data sharing processing method applied to human resources in embodiment 1.

[0056] Fig. 2 The structural diagram of the data sharing processing system applied to human resources in embodiment 1.

[0057] Fig. 3 The permission division flowchart in embodiment 1. DETAILED DESCRIPTION

[0058] In order to make the above-mentioned purposes, features and advantages of the present application more obvious and easy to understand, the specific embodiments of the present application will be described in detail in conjunction with the drawings of the specification.

[0059] In the following description, many specific details are set forth in order to provide a thorough understanding of the present application, but the present application can also be implemented in other ways different from those described herein, and those skilled in the art can make similar generalizations without departing from the connotation of the present application, therefore the present application is not limited to the specific embodiments disclosed below.

[0060] Second, the "one embodiment" or "an embodiment" referred to herein means a specific feature, structure, characteristic, or combination of features and / or characteristics described herein that can be included in at least one implementation of the present application. The various appearances of "in one embodiment" or "an embodiment" in the specification do not all refer to the same embodiment, although they can.

[0061] Embodiment 1, Reference Figs. 1-3 For the first embodiment of the present application, the embodiment provides a data sharing processing method applied to human resources, including the following steps:

[0062] S1, after collecting multi-source data and preprocessing, the data is classified and shared for storage, including a public storage area and an authorized access area;

[0063] Specifically, the multi-source data includes recruitment data (job demand, resume matching data, interview evaluation), employee data (employee basic information and corresponding position information), salary and benefit data (including basic salary, job salary, performance salary, salary history, etc.), performance data (performance evaluation indicators, evaluation period, evaluation results, reward and punishment records, performance improvement plan, etc.), training data (employee training records, training courses, career development planning, etc.);

[0064] After cleaning and standardizing the data, the data is managed in layers using an index tree (Index Tree) to improve query efficiency:

[0065] Each piece of data generates an index according to the data category (recruitment / employee / salary and benefit / performance / training):

[0066] Distributed storage using Kubernetes + Hadoop HDFS improves storage reliability;

[0067] Cold and hot data layer storage is used: high-frequency access data (such as job demand, resume matching data) -> cache layer (Redis), to improve query efficiency, low-frequency access data (such as career development records, salary history) -> blockchain storage (Fabric), to ensure data security and traceability.

[0068] Kubernetes (K8s) is a container orchestration system that automates the deployment, scaling, and management of applications. In the invention, Kubernetes is used for distributed storage management, combined with Hadoop HDFS, which can efficiently manage the load balancing of data storage nodes and improve the scalability and high availability of data storage. HDFS is a distributed file system in the Hadoop ecosystem, which is good at handling large-scale data sets. It adopts a master-slave architecture and uses a block storage mechanism to improve storage reliability. In the invention, the use of HDFS can effectively store and manage massive recruitment, enterprise, and industry data, ensuring the stability of data storage and improving disaster recovery capabilities in combination with Kubernetes. The data sharing processing method applied to human resources proposed in the invention significantly improves the reliability, query efficiency, and security of human resource data storage and management through key technologies such as data cleaning, index tree hierarchical management, Kubernetes+HDFS distributed storage, hot and cold data hierarchical storage, Fabric blockchain evidence storage, and Redis cache optimization.

[0069] S2, role and access permission management of the stored data is performed, and data sharing compliance audit is executed;

[0070] Specifically, the role and access permission management of the stored data is performed, and the data sharing compliance audit includes:

[0071] According to the user's role, basic access permissions are granted, including HR manager → can access complete performance data and salary, department manager → can only access direct subordinate employee training and attendance data, and recruitment agency → can only access job matching data and cannot view candidate detailed information;

[0072] Analyzing the data sharing request submitted by the user, identifying the data category: recruitment data: resume matching degree, interview evaluation, employee data: basic identity information;

[0073] When sharing data, zero-knowledge proof is used to ensure that the data meets the desensitization standard, and data filtering is performed according to the sharing party (HR, recruitment agency, government regulatory agency): recruitment agency: only can view the matching degree of job seekers and positions, cannot access name, contact information, etc. External enterprises: only can view anonymized career development data, cannot access personal information. Government agencies: can access full data, but need to be stored and strictly approved;

[0074] All data sharing requests are stored, including: access user ID, shared data category, sharing timestamp, and shared access IP;

[0075] The Merkle Tree technology is used to calculate the hash of the data sharing record, and the Fabric blockchain is used to store the hash value of the shared data. Fabric is a consortium blockchain suitable for enterprise-level applications, which adopts a modular architecture and supports smart contracts (Chaincode) and private data channels. The invention uses the Fabric blockchain to store the hash value of the data sharing record, ensuring that the data is tamper-proof and achieving traceability of data access. Compared with public chains, the permission control mechanism of Fabric is more powerful, supporting the compliance requirements of inter-enterprise data sharing.

[0076] The invention adopts a role-based access control method to ensure that different users can only access data related to their responsibilities. Users only need to access a limited data set, reducing the system's computing burden and improving data query speed. Different role users can only access necessary data, reducing the risk of data misuse. Zero-knowledge proof is used to ensure that data sharing meets the desensitization standard, avoiding sensitive information leakage. Shared data is stored in the Fabric blockchain, ensuring data traceability and preventing data tampering. Fabric blockchain combines Merkle Tree to store data hash values, ensuring data integrity. Any data sharing behavior can be traced, improving data access compliance. Data sharing records can be verified, enhancing the trust between enterprises and regulatory agencies.

[0077] S3, dynamically adjust access permissions based on access environment, and store permission change records through Fabric consortium chain, detect anomalies in real-time access data and execute security policies;

[0078] Specifically, dynamically adjusting access permissions based on access environment includes:

[0079] When the visitor logs in to the system, automatically obtain device hardware information, including device type, operating system, browser type, and whether it is an enterprise registered device;

[0080] Based on the device hardware information, use the SHA-256 hash algorithm to calculate the unique identifier of the device, and compare it with the enterprise registered device database. If the unique identifier matches, mark the device as a high-trust device. If the unique identifier does not match but the IP is consistent, it is a medium-trust device, triggering secondary authentication. If the unique identifier does not match and there are multiple abnormal logins, it is a low-trust device, limiting some permissions;

[0081] Get the access IP through the HTTP request header, use the GeoIP API to parse the country, city, and operator information corresponding to the IP. If the IP is an enterprise intranet IP, it is a high-trust device. If the IP is an enterprise VPN, it is a medium-trust device. If the IP is unknown, it is a low-trust device and needs additional verification;

[0082] Obtain the BSSID of the current Wi-Fi through Windows, analyze the BSSID (Basic Service Set Identifier) to obtain the Wi-Fi connection mode, if the BSSID belongs to the enterprise registration range, it is a high-trust network, if the BSSID belongs to the home network provided by the ISP, it is a medium-trust network, and if the BSSID belongs to a public hotspot such as a coffee shop or an airport, it is a low-trust network;

[0083] Integrate the collected device, IP, and network information into the access environment database;

[0084] Based on the trust level marks of the device, IP, and network information in the environment database, set the corresponding trust level score, and based on the trust level score, use a weighted method to calculate the final security score of the user access environment;

[0085] Obtain user identity information, and use RBAC rules to assign default access permissions to users according to their role types, including: HR managers -> access sensitive data such as salary, performance, and attendance, department managers -> access subordinate employee data such as performance and attendance, and ordinary employees -> only access personal data such as attendance records and training data; and map different permissions to numerical values;

[0086] Collect user access behavior data, including access time (whether within normal working hours?), access location (whether within the company range?), and access frequency (whether frequently accessing sensitive data in a short period of time?), use a time window function to calculate the access time score, determine the location score based on the user coordinates, including 1 if located in the company and 0 otherwise, calculate the access frequency score using the exponential weighted moving average method, and normalize the calculated scores and perform weighted calculation to obtain the environment adaptability factor;

[0087] Based on the permissions, security score, and environment adaptability factor, calculate the user access permission score A by weighted calculation, compare the score with the preset threshold, and obtain high-level permissions, medium-level permissions, and low-level permissions, specifically: set the threshold q and W by statistical analysis of historical data, and q>W, if the score A is greater than or equal to q, it is a high-level permission, if A is less than q and greater than or equal to W, it is a medium-level permission, and if A is less than W, it is a low-level permission;

[0088] Define a set of all users U in the enterprise, including user ID, user attributes, and user behavior;

[0089] A set of role categories R, including but not limited to HR managers, department managers, ordinary employees, IT administrators, etc.;

[0090] A set of permissions P, including but not limited to user permissions to access salary and welfare data, performance data, and training data, etc.;

[0091] The mapping between the role and the permission is established based on a role-permission mapping matrix of a set;

[0092] The matching degree S of the role r on the permission p is calculated:

[0093]

[0094] In the formula, is the number of times that the user u accesses the permission p, is the total number of times that the user u accesses all permissions;

[0095] The coverage degree Z of the role r on the permission set P is calculated:

[0096]

[0097] In the formula, is the matching degree score of the role r on the permission set P, is the total number of the permission set;

[0098] The management cost c caused by the adjustment of the role is measured:

[0099]

[0100] In the formula, m is the total number of roles, is the number of times of permission adjustment required for role change;

[0101] The role permission change log is counted, and the trade-off between the role coverage degree and the role change cost is optimized using the non-dominated sorting genetic algorithm to obtain a Pareto optimal solution set, including:

[0102] The optimization target is defined as maximizing the role coverage degree and minimizing the role change cost, and an optimization target matrix is constructed according to the calculated role coverage degree and the management cost of role change, the rows of the matrix represent the roles, and the columns represent the optimization targets;

[0103] The role information is read as the initial population, and the population individuals represent a role permission configuration scheme. The population is sorted, and based on the concept of non-dominated solution, if solution A is better than solution B in all targets, then solution A supports solution B, if solution A and solution B do not dominate each other, they belong to the same Pareto front, the Pareto level is calculated, including: the first layer (F1): solutions that are not dominated by any other solution, the second layer (F2): solutions that are only dominated by F1 layer solutions, and so on, multiple Pareto levels are constructed;

[0104] The crowding distance U of solution i is calculated based on the Pareto level:

[0105]

[0106] In the formula, is the optimized value of the next solution i+1 adjacent to i in the Pareto hierarchy on the objective j, is the optimized value of the previous solution i−1 adjacent to i in the Pareto hierarchy on the objective j, and denote the maximum and minimum values in the current population;

[0107] select the solution set with the lowest Pareto hierarchy (e.g., F1), and within the same hierarchy, select the solution with the largest crowding distance as the Pareto optimal solution;

[0108] traverse the Pareto optimal solution set to obtain the role r and its associated user u and permission p, count the frequency of user access to p, compare the frequency with the preset threshold I, if the frequency is greater than or equal to the threshold I, it is a high access frequency user, then the permission is preferentially allocated, otherwise it is a low access frequency user, then the access is limited; the preset threshold I is set by the Z-Score statistical method;

[0109] calculate the Q-learning access reward value to measure the optimization effect of different role-permission allocation schemes, and define the Q-learning update formula:

[0110]

[0111] wherein, is the Q value of taking action a under the current state s, is the weight that controls the new and old Q values, which is set through experiments, is the immediate reward value after executing a in the current state s, is the influence of future rewards, the greater it is, the more attention is paid to long-term returns, is the maximum Q value of all possible actions a' in the new state s', representing the future reward estimate under the optimal strategy;

[0112] and define the reward function E:

[0113]

[0114] wherein, and are weight parameters for balancing coverage and change cost, which are determined by an adaptive method, X is the role coverage improved by the new role-permission allocation scheme, and L is the permission change cost reduced after role optimization;

[0115] store the calculated Q-learning update value in the reinforcement learning optimization database, and based on the Q-learning calculation result, select the cross point with the highest Q value as the target point for genetic algorithm cross operation;

[0116] From the intersection point calculated by the Q-learning, a forward search is performed for the optimal access right allocation scheme, and from the tail of the role right allocation scheme, a backward search is performed for the role allocation scheme with the highest matching degree. The final right allocation scheme is generated by combining the results of the forward search and the backward search.

[0117] Based on device information authentication, improve access security The invention uses SHA-256 hash algorithm to calculate the unique identifier of the device, and compares it with the enterprise registered device database. This mechanism can effectively prevent device forgery attacks, and ensure that only trusted devices can access the system. In addition, the invention supports dynamic trust adjustment based on access history. If the device identifier does not match but the IP address is consistent, secondary authentication is triggered to further enhance security. Dynamic permission management based on IP and network environment, optimize access control strategy The invention combines GeoIP API to parse IP address information, judges whether the visitor belongs to the enterprise internal network, and analyzes the Wi-Fi connection type based on BSSID (Basic Service Set Identifier). This mechanism can realize hierarchical access permission management based on network environment, for example: in the enterprise internal network environment, users can get complete permissions, while in public Wi-Fi, the system will limit the access permission of sensitive data. This dynamic adjustment strategy effectively reduces the risk of data leakage caused by external network access. Build environment database to provide accurate security score The invention builds an access environment database through device, IP, network environment and other information, and uses a weighted calculation method to calculate the security score of user access. Different security scores correspond to different access permission levels (high security, medium security, low security). This way overcomes the limitations of traditional static permission management, making permission configuration more flexible and adaptable to different environmental changes. Based on RBAC and access behavior analysis, realize fine-grained permission allocation. The invention uses RBAC rules to allocate basic permissions according to user roles, and dynamically adjusts permissions based on access behavior data (such as access time, access frequency, access location). To improve analysis accuracy, the invention introduces a time window function and an exponential weighted moving average method to calculate access frequency score and environment adaptability factor, so that the system can accurately judge the user's access mode and adjust the permissions accordingly, avoiding false positives or allowing potential risks. Optimize role permission allocation to improve the intelligence of permission management In terms of role permission management, the invention uses non-dominated sorting genetic algorithm (NSGA-II) to optimize the trade-off between role coverage and role change cost. Specifically, the system first calculates the role matching degree S and the role coverage Z, and uses Pareto frontier optimization to select the optimal role configuration scheme. This way avoids the single and inefficient permission allocation, while ensuring system security, it improves the flexibility of access management. Reinforcement learning optimizes permission management to improve security protection capability The invention further introduces Q-learning reinforcement learning to optimize and evaluate different role permission schemes. Q-learning calculates the access reward value Q(s,a) to find the optimal solution among multiple role permission allocation schemes. This mechanism can continuously adjust access permissions as user behavior changes, improve the adaptive ability of the system, and actively adjust the permission strategy when detecting abnormal access patterns to prevent malicious attacks or data breaches.

[0118] Further, the abnormality detection and security policy execution according to the real-time access data comprises collecting user access behavior data using log analysis, storing access logs using a time series database, establishing a user's historical access time series using time series analysis, calculating a user historical access average value and an access frequency standard deviation, and calculating an access behavior deviation M according to a current access frequency minus the historical access average value divided by the access frequency standard deviation;

[0119] setting an abnormal threshold N and , and N If M is less than the threshold N, it is low risk, an abnormal log is recorded, and the user continues to access. If M is less than the threshold and greater than the threshold N, it is medium risk, manual review is performed. If M is greater than or equal to the threshold , it is high risk, access rights are limited, the account is frozen, and the administrator is notified.

[0120] Based on log analysis, time series database, time series analysis, and access behavior deviation calculation, an efficient real-time abnormality detection and security policy execution mechanism is realized. Compared with the traditional static access control method, the method can dynamically analyze user access patterns, improve the accuracy of abnormality detection, and optimize access security through hierarchical security policies. The time series database is used to store access logs, ensuring the efficiency and scalability of data storage. Combined with time series analysis, a user access history model is constructed to improve the adaptability and intelligence level of the system. Finally, by calculating the access behavior deviation M, the system can effectively quantify the degree of user access behavior abnormality, and based on the set threshold N and T, different levels of security response measures are taken to reduce false blocking, improve the accuracy and security of access management, and thus enhance the overall security and stability of the human resource data sharing system.

[0121] S4, based on the stored data, the job seeker and the enterprise are bidirectionally matched, a matching ranking table is generated for real-time pushing, and the data is stored in a database;

[0122] Specifically, based on the stored data, the job seeker and the enterprise are bidirectionally matched, and a matching ranking table is generated for real-time pushing, comprising:

[0123] Skill keywords in the resume of the job seeker and skill keywords in the job position description of the enterprise are extracted using TF-IDF, and a matching degree of the skills of the job seeker and the job position requirements is calculated using cosine similarity;

[0124] The experience years of the job position requirements and the actual experience years of the job seeker are extracted for rule matching to obtain an experience matching degree;

[0125] The rule matching is performed based on the expected salary of the job seeker and the salary provided by the enterprise, and a salary matching degree is obtained;

[0126] The distribution probability of the job seeker in the past post is calculated based on the Bayesian classification, and the post matching degree is calculated according to the category of the current post;

[0127] The enterprise recruitment priority is calculated based on the decision tree and the vacancy time of the current post and the average recruitment time of the industry. If the vacancy time exceeds the average time, it means that the post has a higher degree of urgency;

[0128] The job seeker adaptation degree is obtained by weighted calculation of the skill matching degree, experience matching degree and salary expectation matching degree of the job seeker, and the enterprise adaptation degree is obtained by weighted calculation of the post demand matching degree and the enterprise recruitment priority;

[0129] The final matching score is obtained by mean calculation of the job seeker adaptation degree and the enterprise adaptation degree. The job seekers are sorted in descending order according to the matching score, a matching list is generated and pushed to the enterprise HR.

[0130] The skill matching degree is calculated by TF-IDF+cosine similarity, the experience and salary adaptation degree is calculated by rule matching, the post adaptability is predicted by Bayesian classification, and the recruitment priority is calculated by decision tree, so that the accurate two-way matching of the job seeker and the enterprise is realized, and the intelligent sorting and pushing mechanism is used to optimize the recruitment process. Compared with the traditional keyword matching method, the matching strategy of the application is more accurate and intelligent, which can effectively improve the recruitment conversion rate, reduce the recruitment cost of the enterprise, improve the job efficiency of the job seeker, and build a more scientific and reasonable talent recruitment ecological system. The skill matching degree, the experience matching degree, the salary matching degree, the post adaptability and the enterprise recruitment priority are selected as the core matching factors, which are mainly based on the key decision points in the actual recruitment process, so as to ensure the accurate matching of talents and posts and improve the success rate and matching efficiency of recruitment. The skill matching degree is a core standard for measuring whether the job seeker has the professional ability required by the post, and the TF-IDF keyword extraction+cosine similarity calculation can accurately evaluate the skill matching degree and avoid the limitations of traditional keyword matching. The experience matching degree ensures that the candidate has the work experience required by the post, reduces the screening errors caused by insufficient experience in the recruitment process, and the salary matching degree is a key factor affecting the success rate of recruitment. The rule matching of the expected salary of the job seeker and the salary provided by the enterprise can improve the real feasibility of matching and reduce the offer rejection rate caused by the difference in expected salary. The post adaptability is based on Bayesian classification analysis of the career background of the job seeker to predict the adaptability of the target post, optimize the career development path and improve the long-term stability. The enterprise recruitment priority is based on decision tree calculation of the post vacancy time and the average recruitment time of the industry, so that the post with high recruitment urgency can be matched with high adaptability candidates in priority, thereby optimizing the talent acquisition efficiency of the enterprise. In summary, these matching degree indexes comprehensively cover the core elements of the matching of the job seeker and the post, can build a more scientific and reasonable talent recruitment model, and improve the accuracy, stability and overall efficiency of recruitment.

[0131] Further, the collected real-time data, permission adjustment data and talent recommendation data are stored in the database, the database stores the data with a time stamp, and the stored data is uploaded to the cloud for backup.

[0132] The embodiment also provides a data sharing processing system applied to human resources, which comprises:

[0133] A data acquisition module is configured to acquire multi-source data and perform data cleaning and standardization processing.

[0134] A permission management module is configured to allocate initial permissions of users and perform data compliance management by using zero-knowledge proof and Fabric evidence.

[0135] A permission adjustment module is configured to automatically adjust access permissions according to user access environments and detect abnormal behaviors.

[0136] The matching module is responsible for the two-way matching of the job seeker and the enterprise, and generates a matching ranking table.

[0137] The data storage module is used for storing matching data, permission adjustment data, and sharing logs, and supports cloud backup.

[0138] The embodiment also provides a computer device suitable for the data sharing processing method applied to human resources, which comprises a memory and a processor; the memory is used for storing computer executable instructions, and the processor is used for executing the computer executable instructions to realize the data sharing processing method applied to human resources as proposed in the above embodiment.

[0139] The computer device can be a terminal, and the computer device comprises a processor, a memory, a communication interface, a display screen and an input device connected through a system bus. The processor of the computer device is used for providing computing and control capabilities. The memory of the computer device comprises a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operating system and the computer program in the non-volatile storage medium. The communication interface of the computer device is used for wired or wireless communication with external terminals. The wireless communication can be achieved through WIFI, an operator network, NFC (Near Field Communication) or other technologies. The display screen of the computer device can be a liquid crystal display screen or an electronic ink display screen. The input device of the computer device can be a touch layer overlaid on the display screen, or a key, trackball or touchpad arranged on the shell of the computer device, or an external keyboard, touchpad or mouse, etc.

[0140] The embodiment also provides a storage medium having a computer program stored thereon, the program being executed by a processor to realize the data sharing processing method applied to human resources as proposed in the above embodiment. The storage medium can be realized by any type of volatile or non-volatile storage device or a combination thereof, such as a static random access memory (SRAM), an electrically erasable programmable read-only memory (EEPROM), an erasable programmable read-only memory (EPROM), a programmable read-only memory (PROM), a read-only memory (ROM), a magnetic storage, a flash memory, a magnetic disk or an optical disk.

[0141] It should be noted that the above examples are only used to illustrate the technical solutions of the present application but not to limit the present application. Although the present application has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present application can be modified or equivalently replaced without departing from the spirit and scope of the present application, and all modifications and equivalents should be included in the scope of the claims of the present application.

Claims

1. A data sharing and processing method applied to human resources, characterized in that: The application relates to a data sharing and matching system and method. After collecting and preprocessing multi-source data, the data is classified and stored, including a public storage area and an authorized access area; The multi-source data includes recruitment data, employee data, salary and welfare data and performance data; After storing the data, role and access permission management is performed, and data sharing compliance auditing is executed; Based on the access environment, the access permission is dynamically adjusted, the permission change record is stored in the Fabric alliance chain, the real-time access data is used for abnormality detection and security strategy execution; Based on the stored data, the job seeker and enterprise are bidirectionally matched, a matching and sorting table is generated and pushed in real time, and the data is stored in a database; The dynamic adjustment of the access permission based on the access environment comprises the following steps: When the visitor logs in the system, the device hardware information is automatically obtained; Based on the device hardware information, high-trust devices, medium-trust devices and low-trust devices, high-trust IP, medium-trust IP and low-trust IP and high-trust networks, medium-trust networks and low-trust networks are obtained; The collected device, IP and network information is integrated into an access environment database; Based on the trust degree marks of the device, IP and network information in the environment database, corresponding trust degree scores are set, and based on the trust degree scores, the final security score of the user access environment is calculated by using a weighting method; User identity information is obtained, and RBAC rules are used to allocate default access permissions to the user according to the role type of the user, and different permissions are mapped to values; User access behavior data is collected, access time scores, location scores and access frequency scores are calculated, and the calculated scores are normalized and then weighted to obtain an environment adaptability factor; Based on the permission, security score and environment adaptability factor, the user access permission score A is calculated by weighting, and the score is compared with a preset threshold value to obtain high-level permission, medium-level permission and low-level permission; All user sets U, role category sets R and permission sets P in an enterprise are defined; A role and permission mapping matrix is established based on the sets; The matching degree S of the role r on the permission p is calculated, the coverage degree Z of the role r on the permission set P is calculated, and the management cost c brought by the role adjustment is measured; The role permission change log is counted, and the trade-off between the role coverage degree and the role change cost is optimized by using a non-dominated sorting genetic algorithm to obtain a Pareto optimal solution set; The Pareto optimal solution set is traversed, the role r and the associated user u and permission p are obtained, the frequency of user access p is counted, the frequency is compared with a preset threshold value I, if the frequency is greater than or equal to the threshold value I, the user is a high-access-frequency user, and the permission is preferentially allocated, otherwise, the user is a low-access-frequency user, and the access is limited; The Q-learning access reward value is calculated to measure the optimization effect of different role-permission allocation schemes, a Q-learning update formula is defined, and a reward function E is defined; The calculated Q-learning update value is stored in a reinforcement learning optimization database, based on the Q-learning calculation result, the cross point with the highest Q value is selected as the target point of the genetic algorithm cross operation. From the intersection point calculated by Q-learning, forward search the best access right allocation scheme, from the tail of the role right allocation scheme, search the role allocation scheme with the highest matching degree, combine the results of forward search and reverse search, and generate the final right allocation scheme.

2. The data sharing process application for human resources according to claim 1, characterized in that: The stored data is used for bidirectional matching of job seekers and enterprises, and a matching ranking table is generated for real-time pushing, including: Skill keywords in the resume of the job seeker and skill keywords in the job position description are extracted using TF-IDF, and the matching degree of the job seeker's skills and the job requirements is calculated using cosine similarity; The experience years required by the job position and the actual experience years of the job seeker are extracted for rule matching to obtain the experience matching degree; Based on the expected salary of the job seeker and the salary provided by the enterprise, rule matching is performed to obtain the salary matching degree; Based on Bayesian classification, the distribution probability of the job seeker in the past positions is calculated, and the job matching degree is calculated according to the category of the current position; Based on the decision tree, the enterprise recruitment priority is calculated based on the vacancy time of the current position and the average recruitment time of the industry; The job seeker's adaptation degree is calculated by weighting the skill matching degree, experience matching degree, and salary expectation matching degree, and the enterprise adaptation degree is calculated by weighting the job demand matching degree and the enterprise recruitment priority; The final matching score is calculated by averaging the job seeker's adaptation degree and the enterprise adaptation degree, the job seekers are ranked in descending order according to the matching score, a matching list is generated and pushed to the enterprise HR.

3. The data sharing process applied to human resources according to claim 2, characterized in that: The classified and shared storage of data includes: Index tree is used for hierarchical management of data, each piece of data is indexed according to data category, and Kubernetes+Hadoop HDFS is used for distributed storage and cold and hot data hierarchical storage.

4. The data sharing process applied to human resources according to claim 3, characterized in that: The role and access right management of the stored data, and the data sharing compliance audit include: According to the user's role, basic access rights are granted, the user's submitted data sharing request is analyzed, the data category is identified, and when sharing data, zero-knowledge proof is used, and data filtering is performed according to the sharing party; All data sharing requests are stored, Merkle Tree technology is used to calculate the hash of data sharing records, and Fabric blockchain is used to store the hash value of shared data.

5. The data sharing process for human resources according to claim 4, wherein: The real-time access data is used for anomaly detection and security policy execution, which includes collecting user access behavior data using log analysis, storing access logs using time series database, establishing user's historical access time series using time series analysis, calculating user's historical access average and access frequency standard deviation, and calculating access behavior deviation M; Set abnormal threshold N and And N If M is less than threshold N, it is low risk, record abnormal log, user continues to access, if M is less than threshold And greater than threshold N, it is medium risk, manual review, if M is greater than or equal to threshold It is high risk, limit access rights, freeze account, and notify administrator.

6. The data sharing process application for human resources according to claim 5, characterized in that: The collected real-time data, permission adjustment data, and talent recommendation data are stored in the database, the database is marked with a timestamp, and the stored data is uploaded to the cloud for backup.

7. A data sharing process system for human resources, based on any one of the data sharing process methods for human resources according to claims 1 to 6, characterized in that: It includes, Data acquisition module, used for collecting multi-source data and performing data cleaning and standardization processing; Permission management module, used for allocating initial permissions to users, using zero-knowledge proof and Fabric evidence for data compliance management; The permission adjustment module is configured to automatically adjust the access permission according to the user access environment and detect abnormal behaviors. The matching module is configured to be responsible for the bidirectional matching of the job seekers and the enterprises, and generate a matching and sorting table. The data storage module is configured to store matching data, permission adjustment data and sharing logs, and support cloud backup.

8. A computer device comprising a memory and a processor, the memory storing a computer program, characterized in that: The computer program is executed by the processor to implement the steps of the data sharing processing method for human resources according to any one of claims 1-6.

9. A computer readable storage medium having stored thereon a computer program, characterized in that: The computer program is executed by the processor to implement the steps of the data sharing processing method for human resources according to any one of claims 1-6.

Images