Permission grading-based periodical trial data security management and control system and method

Abstract

The invention discloses a journal manuscript data security management and control system and method based on authority grading, and relates to the technical field of journal manuscripts. The management and control system comprises functional modules which run in sequence, and the functional modules comprise a permission modeling module, a manuscript grading module, a manuscript review linkage module, a data access feedback module and a service interface module. According to the technical key points, a safe, efficient and compliant periodical trial data management and control system is constructed through permission grading and data grading double-wheel driving, the core of the system is that a minimum permission principle, a dynamic adjustment principle and a traceability principle are integrated into executable technical modules, and feedback continuous optimization is performed; the operation of the whole system is not only suitable for academic periodicals, but also can be popularized to high-sensitivity scenes such as scientific research data management or medical paper review, and an intelligent development direction is provided for the construction of a periodical manuscript review data system.

Description

Technical Field

[0001] This invention relates to the field of journal peer review technology, specifically to a journal peer review data security management system and method based on access control levels. Background Technology

[0002] Journal peer review is the process of evaluating and providing feedback on academic papers. Reviewers primarily focus on the rationality of the research design, methods, and data, as well as the accuracy of the results analysis and citations. The basic review process is as follows: Initial reading: Reviewers first quickly skim the title, abstract, and conclusion of the paper to understand the research topic and its importance; Full text reading: Next, reviewers read the entire paper, focusing on the research objectives, methods, results, and conclusions to ensure a comprehensive understanding of the content; Structural evaluation: Evaluating the rationality of the paper's structure and organization, and whether it conforms to academic writing standards; Methods and data review: Reviewing the rationality and effectiveness of the research methods, and checking whether the data collection and analysis methods are appropriate; Results analysis: Evaluating the accuracy and sufficiency of the results section, ensuring that the discussion section is closely related to the results, and interpreting and explaining the results; Citation check: Checking whether the citations and references are appropriate, comprehensive, and consistent with the content of the paper.

[0003] However, current journal peer review systems generally suffer from problems such as loose access control and lack of data grading. For example, a university editorial office uses a general submission system, allowing reviewers to directly download complete PDFs containing the authors' affiliations, resulting in competitors gaining early access to unpublished work. Another example is that after a special issue concludes, temporarily invited overseas reviewers can still access other highly sensitive manuscripts within the system, as their access is not automatically revoked. Alternatively, a reviewer account might download multiple AI-generated manuscripts in bulk late at night without any system alerts, revealing a lack of behavioral monitoring. Access allocation relies on manual operation, easily leading to the misassignment of highly sensitive clinical research manuscripts to unqualified reviewers. The lack of an automatic verification mechanism based on sensitivity and role permission hinders the development and optimization of journal peer review data systems to some extent. Summary of the Invention

[0004] To achieve the above objectives, the present invention provides the following technical solution: A journal peer review data security management system based on access control levels, comprising: The permission modeling module defines a set of user roles R = {r1, r2, ..., rn}, assigns a minimum necessary permission set P(ri) ⊆ A to each user role, where A is the set of atomic operation permissions of the system, and constructs a permission determination function. The manuscript classification module constructs a rule engine to classify the sensitivity level of each manuscript and determine the sensitivity level as any one of L1, L2, L3 and L4. The peer review linkage module implements a permission linkage mechanism during the target manuscript review stage. It matches the target manuscript with the reviewers based on the sensitivity level of the target manuscript, hides the labeling information on the review interface, and provides a structured form simultaneously. The data access feedback module records all data access behaviors to generate audit logs and triggers abnormal behavior detection and dynamic permission adjustment actions in real time. The service interface module supports integration with service interfaces for target manuscripts of different sensitivity levels.

[0005] Furthermore, n represents the total number of user roles, i represents the ID of the corresponding user role, and i = 1, 2, ..., n.

[0006] Furthermore, the set of atomic operation permissions includes at least: viewing the full text of the manuscript, revising peer review comments, and exporting statistical data.

[0007] Furthermore, when constructing the permission determination function, an attribute-based access control model, namely the ABAC model, is adopted: access is allowed only when the user role is authorized to perform the operation and its security permission level is not lower than the sensitivity level; otherwise, access is not allowed.

[0008] Furthermore, the process of matching reviewers with the sensitivity level of the target manuscript is as follows: Based on the sensitivity level of the target manuscript, all reviewers with a sensitivity level ≤ the user's security permission level are selected. If it is L3, only reviewers with L3 or L4 permission are allowed to participate. Using a pre-trained academic domain embedding model, vector similarity is calculated between the target manuscript abstract and the reviewers' historical publication records. Top-N candidates with matching scores exceeding a preset standard threshold are selected and recommended. The reputation dataset of each candidate is retrieved, and a weighted calculation method is used to obtain the comprehensive reputation score of each candidate. Reviewers with a comprehensive reputation score exceeding 0.7 are selected first. If only one reviewer is needed, the reviewer with the highest comprehensive reputation score is selected as the matching target.

[0009] Furthermore, the reputation dataset includes at least: peer review timeliness, depth of comments, and percentage of violations.

[0010] Furthermore, the statement that the labeling information is hidden on the review interface means that after the reviewer logs into the system, the manuscript seen on the review interface is the version with the labeling information hidden; the labeling information includes at least the author's name, affiliation, coordinates, and ID number.

[0011] Furthermore, the process of abnormal behavior detection and dynamic permission adjustment is as follows: scan the frequency of access to manuscripts corresponding to L3 / L4 sensitivity levels according to a preset cycle, filter users whose frequency exceeds the set value, and issue a prompt for review; when any reviewer downloads more than the set number of L3 / L4 sensitivity level manuscripts in batches outside of working hours, a freeze warning instruction is issued; after the L3 / L4 sensitivity level manuscripts have passed the review stage, a permission revocation instruction is triggered.

[0012] Furthermore, the system supports connecting to service interfaces for target manuscripts with different sensitivity levels to achieve tiered filtering output. Specifically, when the sensitivity level is L1, the supported service interface is an open RESTful API that does not require authentication; when the sensitivity level is L2, the supported service interface requires API Key + user role verification; and when the sensitivity level is L3 / L4, the service interface is not supported.

[0013] A method for managing journal peer review data based on access control levels, which includes: Define a set of user roles R = {r1, r2, ..., rn}, assign a minimum necessary set of permissions P(ri) ⊆ A to each user role, where A is the set of atomic operation permissions of the system, and construct a permission determination function; Build a rule engine to classify the sensitivity level of each article and determine the sensitivity level as any one of L1, L2, L3 and L4; During the target manuscript review stage, an access control mechanism is implemented to match the target manuscript with the reviewers based on its sensitivity level. The labeling information is hidden on the review interface, and a structured form is provided simultaneously. Record all data access behaviors to generate audit logs and trigger abnormal behavior detection and dynamic permission adjustment actions in real time; Supports integration with service interfaces for target manuscripts of different sensitivity levels.

[0014] This invention provides a journal peer review data security management system and method based on access control levels, which has the following beneficial effects: (1) This solution achieves a refined permission linkage mechanism through different stages from allocation to display to submission, thus realizing security in advance. It ensures that some highly sensitive manuscripts only flow to compliant reviewers, and that reviewers only know the necessary information. It also prevents the introduction of risks from non-standard content. The operation of the overall mechanism not only meets the rigor requirements of academic publishing, but also meets the compliance requirements of data security, providing solid support for building a credible, efficient and transparent review system. (2) The operation of this solution, which combines abnormal behavior detection and dynamic permission adjustment, constitutes a dynamic, closed-loop and intelligent permission governance system. It can periodically review and solve the problem of permission redundancy, achieve static optimization, and at the same time, abnormal alarms can deal with sudden risks and achieve dynamic interception. Finally, the time limit of permissions is restricted through permission revoke, and life cycle management is achieved. The operation of the overall solution not only ensures the flexibility and professionalism required for academic review, but also strictly adheres to the bottom line of data security. It meets the requirements and has high practical feasibility and institutional sustainability. (3) This solution constructs a secure, efficient and compliant journal review data management system through a dual-drive approach of hierarchical access and hierarchical data. Its core lies in implementing the principles of least privilege, dynamic adjustment and traceability into executable technical modules and continuously optimizing them through feedback. This makes the operation of the overall system not only applicable to academic journals, but also applicable to high-sensitivity scenarios such as scientific research data management or medical paper review, providing an intelligent development direction for the construction of journal review data systems. Attached Figure Description

[0015] Figure 1 This is a modular schematic diagram of the journal peer review data security management system based on access control in this invention. Detailed Implementation

[0016] 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 of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0017] Example 1: Please see Figure 1 This embodiment provides a journal peer review data security management system based on hierarchical access control. This system aims to build a highly secure, highly available, and compliant journal peer review data security management system or platform, and to achieve effective management of journal peer review data through a corresponding hierarchical access control mechanism and a full-process data security control strategy. The system comprises several sequentially operating functional modules, namely: The permission modeling module defines a set of user roles in the system, R={r1, r2, ..., rn}, assigns a minimum necessary permission set P(ri)⊆A to each user role, where A is the set of atomic operation permissions in the system, and adopts an attribute-based access control model to construct a permission determination function. Where n represents the total number of user roles, i represents the ID of the corresponding user role, and i = 1, 2, ..., n; The permission hierarchy is set as follows: When the user role is "Author," the corresponding atomic operation permissions are: only able to submit, view the status of their own manuscript, and receive notifications; when the user role is "Reviewer," the corresponding atomic operation permissions are: able to view assigned manuscripts and submit review comments, but not able to know the author's identity, i.e., double-blind mechanism, etc.; when the user role is "Editor," the corresponding atomic operation permissions are: able to assign reviewers, summarize comments, and initiate revision / acceptance processes, etc.; when the user role is "Administrator," the corresponding atomic operation permissions are: have global permissions such as system configuration, user management, and log auditing, but cannot view manuscript content by default, etc. By adopting the above role-permission matrix, the problem of traditional systems often mixing editor and administrator permissions, causing sensitive data to be easily exposed is solved. This solution strictly limits the operation boundaries, ensuring that editors cannot delete users and administrators cannot modify manuscripts, avoiding the phenomenon of unauthorized operations due to permission redundancy.

[0018] Therefore, the set of atomic operation permissions should include at least: viewing the full text of the manuscript, revising peer review comments, and exporting statistical data; Using the attribute-based access control model, i.e., the ABAC model, the constructed permission determination function is as follows: if (u∈Ra)∧Sensitivity(d)≤Clearance(u), Allow(u, a, d)=1; if otherwise, Allow(u,a,d)=0; Wherein, Allow represents the permitted operation permission, u, a, and d represent the current user role, the requested operation, and the target data object, respectively; Ra represents the set of user roles allowed to execute operation a; Sensitivity(d)∈{1, 2, 3, 4} represents the sensitivity level, namely L1, L2, L3, and L4; Clearance(u)∈{1, 2, 3, 4} represents the user's security permission level. It should be noted that access is only allowed when the user role is authorized to execute the operation and its security permission level is not lower than the sensitivity level; otherwise, access is not allowed. For example, L3 level manuscripts can only be viewed by editors with L3 or higher permissions, while ordinary reviewers at L2 level can only see the anonymized abstract.

[0019] The manuscript classification module constructs a rule engine to classify the sensitivity level of each manuscript. L1 corresponds to metadata of accepted manuscripts, including titles and keywords, and is classified as public. L2 corresponds to the full text of regular manuscripts, i.e., without sensitive information such as ID numbers or birthdates, and is classified as internal. L3 corresponds to anonymized data, including unpublished algorithms, journal concepts, and anonymized clinical data, and is classified as anonymized. L4 corresponds to sensitive data, including human genetic resources and identity information, and is classified as sensitive. The hierarchical criteria for the aforementioned rule engine are as follows: Content recognition: NLP models are used to scan for keywords such as clinical trials and coordinates; Author declaration: When submitting a manuscript, authors can select whether it contains sensitive content; Historical behavior: If a manuscript by the same author has been labeled L3 multiple times in the past, the new manuscript will default to L3; In addition, a dynamic downgrade mechanism can be introduced according to actual needs: After the target manuscript is accepted, the system will automatically trigger the desensitization process, downgrading the target manuscript corresponding to L3 / L4 to L2 for subsequent public retrieval.

[0020] The peer review linkage module implements a permission linkage mechanism during the target manuscript review stage. It matches the target manuscript with the reviewers based on the sensitivity level of the target manuscript, hides the labeling information on the review interface, and provides a structured form simultaneously. The specific process of matching reviewers with the target manuscript based on its sensitivity level is as follows: First, based on the sensitivity level (Sensitivity(d)) of the target manuscript, all reviewers with Sensitivity(d) ≤ user security permission level (Clearance(u)) are selected. If it is L3, only reviewers with L3 or L4 permission are allowed to participate. Among them, the target manuscript corresponding to L4 is processed by a pre-designated reviewer and requires secondary confirmation from the administrator. Second, using a pre-trained academic domain embedding model, vector similarity is calculated between the target manuscript abstract and the reviewer's historical publication records. Candidates with Top-N matching scores exceeding a preset standard threshold are selected and recommended. Finally, the reputation dataset of each candidate is retrieved, and a weighted calculation method is used to obtain the comprehensive reputation score of each candidate. Reviewers with a comprehensive reputation score higher than 0.7 are selected first. If only one reviewer is needed, the reviewer with the highest comprehensive reputation score is selected as the matching target.

[0021] It should be noted that the academic field embedding model in this embodiment can adopt the existing SciBERT model; for the reputation dataset, it includes at least: peer review timeliness rate, comment depth, and violation record ratio; among them, the peer review timeliness rate corresponds to a weight of 40%, the comment depth is scored by the editor and finally normalized to the range of 0 to 1, corresponding to a weight of 30%, and the violation record ratio corresponds to a weight of 30%. The specific weighting calculation will not be elaborated in detail here.

[0022] The annotation information is hidden on the review interface. After the reviewers log into the system, the manuscripts they see on the review interface are versions with the labeling information masked. The labeling information includes at least the author's name, affiliation, coordinates, and ID number. Its purpose is to ensure the completeness of the information required for the review while minimizing the possibility of reverse identity verification, which is especially suitable for double-blind review systems.

[0023] The basis or explanation for providing structured forms simultaneously is as follows: To prevent reviewers from bypassing security checks through unstructured methods, such as uploading watermarked Word files or PDFs with embedded scripts, a fully structured form is required for submitting comments. An example of this form is shown below: The form fields include: Innovation Evaluation (single choice): Outstanding / Average / Insufficient / No Innovation; Methodological Rigor (multiple choice): Reasonable Experimental Design / Sufficient Data / Appropriate Statistical Methods / Flaws Exist (specific items must be checked); Conclusion Reliability (text box, limited to 500 words); Revision Suggestions (text box, limited to 800 words); Acceptance Suggestions (single choice): Direct Acceptance / Minor Revision / Major Revision / Rejection; Prohibited Behaviors include: Uploading any attachments, including images, documents, and compressed files; Pasting excessively long content into text boxes (e.g., exceeding 1000 words will be considered excessive content); and Using special characters.

[0024] Specifically, structured forms facilitate subsequent automated analysis, such as generating peer review quality reports; prevent malicious code injection or steganography leaks; and guide reviewers to focus on evaluation rather than free interpretation. In summary, through a refined permission linkage mechanism implemented at different stages from allocation to display to submission, security is prioritized, ensuring that some highly sensitive manuscripts only flow to compliant reviewers, who only receive necessary information, and preventing the introduction of risks from non-standard content. The overall mechanism not only meets the rigorous requirements of academic publishing but also satisfies the compliance requirements of data security, providing solid support for building a trustworthy, efficient, and transparent review system.

[0025] The data access feedback module records all data access behaviors to generate audit logs and triggers abnormal behavior detection and dynamic permission adjustment actions in real time. The audit logs must be immutable, and the audit content must include at least: user role ID, IP address, timestamp, accessed data ID, sensitivity level, and operation type, such as viewing / downloading / exporting. The process of abnormal behavior detection and dynamic permission adjustment is as follows: The system automatically scans the frequency of access to manuscripts corresponding to L3 / L4 sensitivity levels according to a preset cycle, filters users whose access frequency exceeds the set value, and issues a prompt for review to remind the administrator to verify the necessity of their permissions. If any reviewer downloads more than the set number of L3 / L4 sensitivity level manuscripts in batches outside of working hours, a freeze warning is issued to automatically freeze their account and notify the administrator. After the L3 / L4 sensitivity level manuscripts have passed the review stage, a permission revocation command is triggered to automatically revoke the corresponding reviewer's permissions.

[0026] Non-working hours refer to the period from 0:00 to 6:00 AM or holidays; Specifically, the abnormal behavior detection and dynamic permission adjustment actions together constitute a dynamic, closed-loop, and intelligent permission governance system. It can periodically review and resolve permission redundancy issues to achieve static optimization. At the same time, abnormal alarms can deal with sudden risks and achieve dynamic interception. Finally, permission revoke can limit the timeliness of permissions and achieve lifecycle management. The operation of the overall solution not only ensures the flexibility and professionalism required for academic review, but also strictly adheres to the bottom line of data security. It meets the requirements and has high practical feasibility and institutional sustainability.

[0027] The service interface module supports connecting to service interfaces for target manuscripts with different sensitivity levels, enabling tiered filtering output. Specifically, when the sensitivity level is L1, the supported service interface is an open RESTful API, requiring no authentication. When the sensitivity level is L2, the supported service interface requires API Key + user role verification. When the sensitivity level is L3 / L4, no service interface is supported. In addition, for sensitivity levels L3 / L4, a result-only return mode can be selected, where external systems submit analysis requests, such as calculating the annual number of articles in a certain field. The system performs calculations within a secure sandbox and only returns the aggregated results, with the original data remaining within the field.

[0028] In summary, this solution constructs a secure, efficient, and compliant journal peer review data management system through a dual-drive approach of hierarchical access control and data classification. Its core lies in implementing the principles of least privilege, dynamic adjustment, and traceability into executable technical modules, and continuously optimizing them through feedback. This allows the overall system to be applied not only to academic journals but also to highly sensitive scenarios such as scientific research data management or medical paper review, providing an intelligent development direction for the construction of journal peer review data systems.

[0029] Example 2: Based on Example 1, this example also provides a method for security management of journal peer review data based on permission hierarchy. The method includes: defining a set of user roles R={r1, r2, ..., rn}, assigning a minimum necessary permission set P(ri)⊆A to each user role, where A is the set of atomic operation permissions of the system, and constructing a permission determination function; Build a rule engine to classify the sensitivity level of each article and determine the sensitivity level as any one of L1, L2, L3 and L4; During the target manuscript review stage, an access control mechanism is implemented to match the target manuscript with the reviewers based on its sensitivity level. The labeling information is hidden on the review interface, and a structured form is provided simultaneously. Record all data access behaviors to generate audit logs and trigger abnormal behavior detection and dynamic permission adjustment actions in real time; Supports integration with service interfaces for target manuscripts of different sensitivity levels.

[0030] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented in software, the above embodiments can be implemented, in whole or in part, as a computer program product. Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution.

[0031] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment, depending on actual needs.

[0032] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application.

Claims

1. A journal peer review data security management system based on access control, characterized in that: The system includes: The permission modeling module defines a set of user roles R = {r1, r2, ..., rn}, assigns a minimum necessary permission set P(ri) ⊆ A to each user role, where A is the set of atomic operation permissions of the system, and constructs a permission determination function. The manuscript classification module constructs a rule engine to classify the sensitivity level of each manuscript and determine the sensitivity level as any one of L1, L2, L3 and L4. The peer review linkage module implements a permission linkage mechanism during the target manuscript review stage. It matches the target manuscript with the reviewers based on the sensitivity level of the target manuscript, hides the labeling information on the review interface, and provides a structured form simultaneously. The data access feedback module records all data access behaviors to generate audit logs and triggers abnormal behavior detection and dynamic permission adjustment actions in real time. The service interface module supports integration with service interfaces for target manuscripts of different sensitivity levels.

2. The journal peer review data security management system based on access control according to claim 1, characterized in that, n represents the total number of user roles, i represents the ID of the corresponding user role, and i = 1, 2, ..., n.

3. The journal peer review data security management system based on access control according to claim 1, characterized in that, The set of atomic operation permissions should include at least: viewing the full text of the manuscript, modifying peer review comments, and exporting statistical data.

4. The journal peer review data security management system based on access control according to claim 1, characterized in that, When constructing the permission determination function, an attribute-based access control model, namely the ABAC model, is adopted: access is allowed only if the user role is authorized to perform the operation and its security permission level is not lower than the sensitivity level; otherwise, access is not allowed.

5. The journal peer review data security management system based on access control according to claim 1, characterized in that, The process of matching reviewers based on the sensitivity level of the target manuscript is as follows: Based on the sensitivity level of the target manuscript, all reviewers with a sensitivity level ≤ the user's security permission level are selected. If it is L3, only reviewers with L3 or L4 permissions are allowed to participate. Using a pre-trained academic domain embedding model, vector similarity is calculated between the target manuscript abstract and the reviewers' historical publication records. Top-N candidates with matching scores exceeding a preset standard threshold are selected and recommended. The reputation dataset of each candidate is retrieved, and a weighted calculation method is used to obtain the comprehensive reputation score of each candidate. Reviewers with a comprehensive reputation score higher than 0.7 are selected first. If only one reviewer is needed, the reviewer with the highest comprehensive reputation score is selected as the matching target.

6. The journal peer review data security management system based on access control according to claim 5, characterized in that, The reputation dataset should include at least: peer review timeliness, depth of comments, and percentage of violations.

7. The journal peer review data security management system based on access control according to claim 1, characterized in that, The statement that labeling information is hidden on the review interface means that the manuscript seen on the review interface after the reviewer logs into the system is the version with labeling information hidden; the labeling information includes at least the following: author's name, affiliation, coordinates, and ID number.

8. The journal peer review data security management system based on access control according to claim 1, characterized in that, The process of abnormal behavior detection and dynamic permission adjustment is as follows: scan the frequency of access to manuscripts corresponding to L3 / L4 sensitivity levels according to a preset cycle, filter users whose frequency exceeds the set value, and issue a prompt for review; when any reviewer downloads more than the set number of L3 / L4 sensitivity level manuscripts in batches outside of working hours, a freeze warning instruction is issued; after the L3 / L4 sensitivity level manuscripts have passed the review stage, a permission revocation instruction is triggered.

9. The journal peer review data security management system based on access control according to claim 1, characterized in that, To handle target manuscripts with different sensitivity levels, the system supports integration with service interfaces for tiered filtering output. Specifically, for sensitivity level L1, the supported service interface is an open RESTful API that requires no authentication. For sensitivity level L2, the supported service interface requires API Key and user role verification. For sensitivity levels L3 / L4, no service interface is supported.

10. A method for secure management of journal peer review data based on access control, characterized in that: The method includes: Define a set of user roles R = {r1, r2, ..., rn}, assign a minimum necessary set of permissions P(ri) ⊆ A to each user role, where A is the set of atomic operation permissions of the system, and construct a permission determination function; Build a rule engine to classify the sensitivity level of each article and determine the sensitivity level as any one of L1, L2, L3 and L4; During the target manuscript review stage, an access control mechanism is implemented to match the target manuscript with the reviewers based on its sensitivity level. The labeling information is hidden on the review interface, and a structured form is provided simultaneously. Record all data access behaviors to generate audit logs and trigger abnormal behavior detection and dynamic permission adjustment actions in real time; Supports integration with service interfaces for target manuscripts of different sensitivity levels.

Images