CRM-GIS-oriented multi-dimensional right dynamic adaptation method and system

Abstract

The invention relates to the technical field of permission adaptation, in particular to a CRM-GIS-oriented multi-protection permission dynamic adaptation method and system. The method comprises the following steps: firstly, defining in a CRM-GIS platform and acquiring a multi-dimensional authority attribute related to an access request in real time; based on a preset permission policy rule in the platform, processing an access request by using a policy decision engine; then, based on the current spatio-temporal information and behavior pattern information of the user, an artificial intelligence risk assessment module is utilized to analyze the real-time risk level of the current operation of the user; thirdly, dynamically judging and adjusting a final permission adaptation decision aiming at the access request by integrating the multi-dimensional permission attribute, the permission strategy rule and the risk level; according to the final permission adaptation decision, the CRM-GIS platform compulsorily executes corresponding access control operation, and limits or allows access to data and functions in the platform; according to the invention, the reliability of multi-dimensional right dynamic adaptation can be improved.

Description

Technical Field

[0001] The present invention relates to the technical field of permission adaptation, and particularly to a multi-dimensional permission dynamic adaptation method and system for CRM-GIS. Background Technique

[0002] Currently, the permission management in CRM-GIS systems mostly relies on the traditional role-based access control (RBAC) model. This model usually pre-defines user roles and corresponding static permission sets, such as salespersons, sales managers, system administrators, etc. Although this method can achieve access control to a certain extent, its limitations are as follows when facing increasingly complex business scenarios and dynamic usage environments: 1) Coarse permission granularity and insufficient flexibility: It is difficult for traditional RBAC to achieve fine-grained permission control for specific data objects (such as customers in a specific area, orders of a specific value) or in a specific context (such as a specific time, the specific geographical location of the user, a specific business process node); 2) Lack of context awareness ability: Existing permission systems often ignore the dynamic context information during user operations, such as the user's current physical location, access time, device security status, the urgency of the current task, etc., which results in the inability to dynamically adjust permissions according to real-time risks or business requirements; 3) Complex cross-organization collaborative permission management: In scenarios involving collaboration among upstream and downstream industrial chains of enterprises (such as suppliers, distributors), it is difficult to balance data sharing and permission control; how to provide appropriate, temporary, task-related access permissions for external partners while ensuring data security is an urgent problem to be solved; Therefore, a multi-dimensional permission dynamic adaptation method and system for CRM-GIS are proposed. Summary of the Invention

[0003] The purpose of the present invention is to provide a multi-dimensional permission dynamic adaptation method and system for CRM-GIS.

[0004] To achieve the above purpose, the present invention provides the following technical solutions: A multi-dimensional permission dynamic adaptation method for CRM-GIS, including: Defining and obtaining in real time multi-dimensional permission attributes related to an access request within the CRM-GIS platform; Processing the access request by using a policy decision engine according to preset permission policy rules stored in the platform to evaluate the initial access permission; Using the artificial intelligence risk assessment module configured within the platform, based on the user's current spatio-temporal information and behavior pattern information, analyze the real-time risk level of the user's current operation, and output the risk level as a decision adjustment factor to the policy decision engine; The policy decision engine comprehensively considers multi-dimensional permission attributes, permission policy rules, and the risk level output by the artificial intelligence risk assessment module to dynamically determine and adjust the final permission adaptation decision for the access request; According to the final permission adaptation decision, the CRM-GIS platform enforces the corresponding access control operations to restrict or allow access to the data and functions within the platform.

[0005] Furthermore, the multi-dimensional permission attributes include: identity and environmental attributes related to the subject initiating the access, object data attributes related to the access object, operation attributes of the requested execution, and context environment attributes characterizing the current operation context; among them, the context environment attributes include the user's current spatio-temporal information and the behavior pattern information of the user's interaction with the CRM-GIS platform.

[0006] Furthermore, the process of using the artificial intelligence risk assessment module configured within the platform to analyze the real-time risk level of the user's current operation based on the user's current spatio-temporal information and behavior pattern information includes: Real-time collect spatio-temporal information related to the current user operation and the behavior pattern information of the user's interaction with the CRM-GIS platform; Process the spatio-temporal information and the behavior pattern information into a structured input feature vector; Input the input feature vector into a pre-trained artificial intelligence model for processing. The artificial intelligence model infers the risk index of the current operation by analyzing the deviation degree of the input feature vector from the preset normal spatio-temporal range and the established user historical behavior baseline; Based on the inference result of the artificial intelligence model, generate a quantified real-time risk level and provide it to the policy decision engine as a decision adjustment factor to assist in dynamically adjusting the final permission adaptation decision.

[0007] Furthermore, the dynamic optimization of the artificial intelligence risk assessment module and the preset permission policy rules is realized through a distributed machine learning framework that protects data privacy; the distributed machine learning framework allows multiple participants to collaboratively train and update the global risk assessment model and permission policy optimization model on the premise of protecting the privacy of their respective original data.

[0008] Further, the distributed machine learning framework is a federated learning framework. The process of using the federated learning framework to collaboratively optimize the artificial intelligence risk assessment module and the preset permission policy rules includes: The central coordination server distributes the initial global risk assessment model structure and the global permission policy meta-model to multiple clients; Each of the clients uses the original privacy data stored locally and related to the CRM-GIS platform operation to locally train the received model, generating local model update parameters, and the original privacy data does not leave the local; The central coordination server collects and aggregates the local model update parameters from multiple clients to iteratively update the global risk assessment model and the global permission policy meta-model, so as to make access control decisions without directly accessing the original data of each client.

[0009] Further, the process of the policy decision engine dynamically determining and adjusting the final permission adaptation decision for the access request by integrating multi-dimensional permission attributes, permission policy rules, and the risk level output by the artificial intelligence risk assessment module includes: Integrate the multi-dimensional permission attributes included in the access request, the optimized permission policy rules, and the real-time risk level output by the artificial intelligence risk assessment module; Based on the predefined decision logic, adjust the initial access permission according to the real-time risk level and the optimized permission policy rules; According to the output of the decision logic, determine the specific access control execution instruction and generate the final permission adaptation decision including the access control execution instruction.

[0010] A multi-dimensional permission dynamic adaptation system for CRM-GIS includes: a multi-dimensional permission attribute acquisition unit for defining and real-time acquiring multi-dimensional permission attributes related to an access request within the CRM-GIS platform; an initial access permission analysis unit for processing the access request using a policy decision engine based on the preset permission policy rules stored in the CRM-GIS platform to evaluate the initial access permission; a risk level analysis unit for analyzing the real-time risk level of the user's current operation using the artificial intelligence risk assessment module configured within the CRM-GIS platform based on the user's current spatio-temporal information and behavior pattern information; a permission adaptation management unit for dynamically determining and adjusting the final permission adaptation decision for the access request by integrating multi-dimensional permission attributes, permission policy rules, and the risk level output by the artificial intelligence risk assessment module; an access control unit for, according to the final permission adaptation decision, enforcing corresponding access control operations by the CRM-GIS platform to restrict or allow access to the data and functions within the CRM-GIS platform.

[0011] Compared with the prior art, the beneficial effects of the present invention are as follows: 1. Through the federated learning framework, the AI risk assessment model and the permission policy meta-model can aggregate the experiences of multiple clients for optimization, so as to better adapt to the risk patterns and business requirements unique to different clients, improving the overall applicability and accuracy of the model; the federated learning framework enables users and departments to jointly build a more powerful security protection system without sacrificing data sovereignty, which helps to break data silos and achieve a higher level of security intelligent collaboration.

[0012] 2. By collecting the spatio-temporal information and behavior patterns of users in real time and analyzing the deviation between the behavior pattern and the historical baseline, the operational risk can be dynamically evaluated for the current specific situation, which enables the permission system to no longer rely solely on static rules, but to identify those operations that seem compliant but are actually abnormal or high-risk, improving the accuracy and initiative of security protection and enhancing the ability to identify potential internal threats and account theft risks.

[0013] 3. By integrating multi-dimensional permission attributes, permission policy rules, and the risk levels output by the artificial intelligence risk assessment module, various dynamic adjustment measures can be taken according to specific situations; this differential processing method can tighten permissions to ensure security in high-risk situations and ensure the smoothness of business processes in low-risk situations, achieving a better balance between security and efficiency, thereby improving the flexibility of the permission system and its adaptability to complex business scenarios. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] Figure 1 is a schematic flow chart of a multi-dimensional permission dynamic adaptation method for CRM-GIS according to the present invention; Figure 2 is a schematic flow chart of model parameter aggregation and global model iterative update according to the present invention; Figure 3 is a schematic structural diagram of a multi-dimensional permission dynamic adaptation system for CRM-GIS according to the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0015] The following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0016] Please refer to Figures 1 to 3 , the present invention provides a multi-dimensional permission dynamic adaptation method and system for CRM-GIS, and the technical solutions are as follows:

[0017] Example 1: To achieve efficient permission management for the CRM-GIS platform, an enterprise uses a multi-dimensional permission dynamic adaptation method for CRM-GIS proposed by the present invention. The process of this method is schematically shown as Figure 1 follows, specifically including: Within the CRM-GIS platform, define and obtain in real time multi-dimensional permission attributes related to access requests; Furthermore, the multi-dimensional permission attributes include: identity and environment attributes related to the subject initiating the access, object data attributes related to the access object, operation attributes for the requested execution, and context environment attributes representing the current operation context; among them, the context environment attributes include the user's current spatio-temporal information and the behavior pattern information of the user's interaction with the CRM-GIS platform; Furthermore, the identity attributes related to the subject include: the user's unique identifier (UserID), the affiliated department, the position role (such as salesperson, regional manager, administrator, etc.), and the credential status used when the user logs in; the environment attributes related to the subject include the compliance status of the access device and the network environment; the object data attributes include CRM data attributes and GIS data attributes; the operation attributes include CRM data operations and GIS data operations; the user's current spatio-temporal information includes the specific time when the access occurs and the user's geographical location; the behavior pattern information includes the historical behavior baseline and the current session behavior; Furthermore, the compliance status of the access device includes: the type of the access source device, the operating system version, whether the device has been registered and trusted in the system, etc.; the network environment includes the network area where the user accesses and the geographical location of the IP address, etc.; Furthermore, the CRM data attributes include the sensitivity identifier of the data record, the order amount range, the customer attribution, etc.; the CRM data operations represent the creation, reading, updating, deletion, export, etc. of data such as customer records, orders, contacts, etc.; the GIS data attributes include the geographical layer access control list, the confidentiality level of GIS data, etc.; the GIS data operations can represent the invocation of spatial analysis functions, map visualization retrieval, surrounding circle search, etc.; Furthermore, the historical behavior baseline is a behavior portrait established by analyzing the user's operation logs (such as login frequency, frequently used functions, data access scope, average session duration, frequently used devices and IPs, etc.) over a period of time in the past, and is used for subsequent calculation of the behavior deviation degree; the current session behavior represents the operation sequence, the amount of data accessed, the operation frequency, etc. of the user in the current session.

[0018] By introducing multi-dimensional permission attributes, it not only conforms to the complex processing environment of the CRM-GIS platform, but also provides a multi-dimensional perspective and data basis for subsequent permission evaluation, thereby improving the reliability of multi-dimensional permission dynamic adaptation.

[0019] According to the preset permission policy rules stored in the platform, use the policy decision engine to process access requests to evaluate the initial access permissions; Furthermore, the preset permission policy rules stored in the platform are defined as an access control policy language based on multi-dimensional attributes, which can be expressed as "IF (condition) THEN (Effect: Permit / Deny) [BECAUSE (Obligations / Advice)]", and it is not unique; among them, the condition is a logical expression of multi-dimensional attributes, used to judge whether the multi-dimensional attribute value is a set value; Furthermore, the policy decision engine, that is, the Policy Decision Point (PDP), is responsible for evaluating access requests and making permission decisions (allow or deny). Based on the predefined permission policy rules, the collected multi-dimensional attributes and policy evaluation, it decides whether to authorize the access request; PDP adopts a rule-based evaluation algorithm to match the attributes with the rules in the policy. If the match is successful, the authorization decision of the rule is applied.

[0020] Use the artificial intelligence risk assessment module configured in the platform to analyze the real-time risk level of the user's current operation based on the user's current spatio-temporal information and behavior pattern information, and output the risk level as a decision adjustment factor to the policy decision engine; Furthermore, the process of using the artificial intelligence risk assessment module configured in the platform to analyze the real-time risk level of the user's current operation based on the user's current spatio-temporal information and behavior pattern information includes: Real-time collect the spatio-temporal information related to the current user operation and the behavior pattern information of the user's interaction with the CRM-GIS platform; Process the spatio-temporal information and behavior pattern information into a structured input feature vector; Input the input feature vector into a pre-trained artificial intelligence model for processing. The artificial intelligence model infers the risk index of the current operation by analyzing the deviation degree of the input feature vector from the preset normal spatio-temporal range and the established user historical behavior baseline; Based on the inference result of the artificial intelligence model, generate a quantified real-time risk level and provide it to the policy decision engine as a decision adjustment factor to assist in dynamically adjusting the final permission adaptation decision; Furthermore, process and transform the collected spatio-temporal information and behavior pattern information into a structured input feature vector, that is, spatio-temporal features and behavior features, so that the AI model can understand and process them; spatio-temporal features are the distance difference and time interval calculated based on the current time and location information and the time and location information of the previous operation; behavior features can be expressed as the similarity between the current operation sequence and the historical operation sequence, and the comparison results of the current operation frequency and data access volume with the historical average level; Furthermore, the pre-trained artificial intelligence model adopts the network structure of an autoencoder; among them, the encoder uses an LSTM network, and the decoder uses a residual network; it is trained with historical data (spatio-temporal data and behavioral data) that does not contain abnormal data, and the mean squared error is selected as the loss function to measure the difference between the original input and the reconstructed data output by the decoder; then, combined with the current input feature vector and the trained model, the reconstruction error (mean squared error) is calculated; the error value is converted into a risk score in the range of [0, 1] using min-max normalization; according to the preset threshold and the risk score, it is divided into different risk levels; Furthermore, in order to reduce false alarms in risk assessment and improve the user experience, a user feedback mechanism is introduced into the model, that is, the user feedback, together with the spatio-temporal features, behavioral features, and risk scores of related operations, is stored as labeled data and input into the model in real time. Incremental learning (such as online gradient descent) is used to update the parameters of the autoencoder and adjust the risk assessment of similar operations.

[0021] By collecting the spatio-temporal information and behavioral patterns of users in real time and using the artificial intelligence model to analyze the deviation between the behavioral pattern and the historical baseline, it is possible to dynamically evaluate the operational risk for the current specific situation, no longer relying solely on static rules, and being able to identify potential risky operations, thereby improving the reliability of multi-dimensional permission dynamic adaptation.

[0022] Furthermore, the dynamic optimization of the artificial intelligence risk assessment module and the preset permission policy rules is realized through a distributed machine learning framework that protects data privacy; the distributed machine learning framework allows multiple participants to jointly train and update the global risk assessment model and permission policy optimization model on the premise of protecting the privacy of their respective original data; By using the distributed machine learning framework, differences between users and enterprise departments, as well as issues such as the original data privacy of customers, can be considered during the process of risk assessment and access permission assessment, ensuring the reliability of multi-dimensional permission dynamic adaptation.

[0023] Furthermore, the distributed machine learning framework is a federated learning framework. The process of using the federated learning framework to jointly optimize the artificial intelligence risk assessment module and the preset permission policy rules includes: The central coordination server distributes the initial global risk assessment model structure and the global permission policy meta-model to multiple clients; Each client uses the original privacy data stored locally and related to the CRM-GIS platform operations to locally train the received model, generating local model update parameters, and the original privacy data does not leave the local area; The central coordination server collects and aggregates the local model update parameters from multiple clients to iteratively update the global risk assessment model and the global permission policy meta-model, so as to make access control decisions without directly accessing the original data of each client. Furthermore, the permission policy meta-model adopts a reinforcement learning model, and an agent is used to adjust the conditional parameters in the permission policy rules; the state is the current permission policy configuration, the action is to modify or generate policy rules, and the environment gives rewards according to the performance (security and business efficiency) of the adjusted permission policy on real historical data; security is represented by the number of abnormal data occurrences within a period of time, and business efficiency is represented by the business processing time. Furthermore, each client uses local data to supervise and train the local risk assessment model to minimize the loss function of local risk prediction, and at the same time iteratively trains the local permission policy meta-model (DQN network) until the end. Furthermore, the process of model parameter aggregation and global model iterative update is as Figure 2 shown, specifically: each client sends the generated local model update parameters to the central coordination server; after collecting the update parameters from multiple clients, the federated averaging aggregation algorithm is used for parameter aggregation; among them, different weights can be assigned to the update parameters of different clients according to factors such as the size of the client data volume during the parameter aggregation process; the aggregated model parameters are applied to the current global risk assessment model and the global permission policy meta-model to form an optimized global model; the optimization process will be iteratively carried out for multiple rounds until the performance of the global model reaches the preset target. Furthermore, in order to prevent the central server from reverse-inferring the original data through parameter updates, the differential privacy algorithm (such as DP-SGD) is applied during the local training process of the client, and noise (such as Gaussian noise) is added to the local model update parameters to further protect user privacy.

[0024] Through the federated learning framework, the AI risk assessment model and the permission policy meta-model can gather the experiences of multiple clients for optimization, so as to better adapt to the unique risk patterns and business requirements of different clients, and improve the overall applicability and accuracy of the model; the federated learning framework enables each user and each department to jointly build a more powerful security protection system without sacrificing data sovereignty, which helps to break data islands and achieve a higher level of security intelligent collaboration.

[0025] The policy decision engine comprehensively considers multi-dimensional permission attributes, permission policy rules, and the risk level output by the artificial intelligence risk assessment module, and dynamically determines and adjusts the final permission adaptation decision for the access request. Furthermore, the process by which the policy decision engine dynamically determines and adjusts the final permission adaptation decision for an access request by integrating multi-dimensional permission attributes, permission policy rules, and the risk level output by the artificial intelligence risk assessment module includes: Integrate the multi-dimensional permission attributes included in the access request, the optimized permission policy rules, and the real-time risk level output by the artificial intelligence risk assessment module; Based on predefined decision logic, adjust the initial access permission according to the real-time risk level and the optimized permission policy rules; According to the output of the decision logic, determine the specific access control execution instruction and generate a final permission adaptation decision including the access control execution instruction; Furthermore, input the current permission policy configuration into the global permission policy meta-model optimized by the federated learning framework to obtain the optimized permission policy rules; Furthermore, evaluate and update the initial access permission according to the multi-dimensional permission attributes in the access request and the optimized permission policy rules, and the evaluation method is the same as above; at the same time, adjust the initial access permission according to the real-time risk level. For example, if the risk level exceeds a certain level, even if the initial permission is to allow access, it should be downgraded according to the degree of risk to require additional identity verification, restrict the data access scope, or directly reject access; Furthermore, define the decision logic by using thresholds and rules; adjust the permission level according to different risk levels, and at the same time use the threshold range of the risk score to refine the adjustment range of the permission level; The following gives the corresponding results of the permission levels, as shown in Table 1; set the risk score threshold for the low risk level to 0.3, the risk score threshold for the high risk level to 0.7, and the risk score threshold (0.5) to refine the adjustment range; Table 1. Corresponding Results of Permission Levels

[0026] Furthermore, in addition to allowing access and denying access, the access control execution instructions also include: requiring additional identity verification, restricting the data access scope, triggering security alerts, etc.; Furthermore, the final permission adaptation decision includes the access control execution instruction, the reason for the decision, the accompanying obligations (for example, the user must accept a security audit), and the timestamp of the decision, etc.; Furthermore, in addition to dynamically adjusting the real-time permission adaptation decision, analyze the user behavior pattern and the trend of the risk score by using a time series model (such as LSTM or ARIMA), so as to predict the risk level of future access requests and adjust the permission policy in advance, which can actively prevent potential risks and improve the system response speed.

[0027] By integrating multi-dimensional permission attributes, permission policy rules, and the risk level output by the artificial intelligence risk assessment module, various dynamic adjustment measures can be taken according to specific circumstances; this differential processing method can not only tighten permissions to ensure security in high-risk situations but also ensure the smoothness of business processes in low-risk situations, achieving a better balance between security and efficiency, thereby improving the flexibility of the permission system and its adaptability to complex business scenarios.

[0028] According to the final permission adaptation decision, the CRM-GIS platform enforces corresponding access control operations to restrict or allow access to data and functions within the platform.

[0029] This embodiment proposes a multi-dimensional permission dynamic adaptation method for CRM-GIS; first, define and obtain in real time multi-dimensional permission attributes related to access requests within the CRM-GIS platform; based on the preset permission policy rules within the platform, use the policy decision engine to process access requests; then, based on the user's current spatio-temporal information and behavior pattern information, use the artificial intelligence risk assessment module to analyze the real-time risk level of the user's current operation; next, integrate the multi-dimensional permission attributes, permission policy rules, and risk level to dynamically determine and adjust the final permission adaptation decision for the access request; according to the final permission adaptation decision, the CRM-GIS platform enforces corresponding access control operations to restrict or allow access to data and functions within the platform; the present invention can improve the reliability of multi-dimensional permission dynamic adaptation.

[0030] Embodiment Two: The present invention also proposes a multi-dimensional permission dynamic adaptation system for CRM-GIS, and the structure of the system is as Figure 3 shown, including: a multi-dimensional permission attribute acquisition unit, an initial access permission analysis unit, a risk level analysis unit, a permission adaptation management unit, and an access control unit; The multi-dimensional permission attribute acquisition unit is used to define and obtain in real time multi-dimensional permission attributes related to access requests within the CRM-GIS platform; The initial access permission analysis unit is used to evaluate the initial access permission by processing the access request using the policy decision engine according to the preset permission policy rules stored in the CRM-GIS platform; The risk level analysis unit is used to analyze the real-time risk level of the user's current operation based on the user's current spatio-temporal information and behavior pattern information by using the artificial intelligence risk assessment module configured in the CRM-GIS platform; Further, the process by which the risk level analysis unit analyzes the real-time risk level of the user's current operation by using the artificial intelligence risk assessment module configured in the CRM-GIS platform includes: Collect real-time spatio-temporal information related to the current user's operations and behavioral pattern information on the interaction between the user and the CRM-GIS platform; Process the spatio-temporal information and behavioral pattern information into a structured input feature vector; Input the input feature vector into a pre-trained artificial intelligence model for processing. The artificial intelligence model infers the risk index of the current operation by analyzing the deviation degree of the input feature vector from the preset normal spatio-temporal range and the established user historical behavior baseline; Based on the inference result of the artificial intelligence model, generate a quantified real-time risk level and provide it to the policy decision engine as a decision adjustment factor to assist in dynamically adjusting the final permission adaptation decision.

[0031] In order to adapt to the specific risk patterns and business requirements of different clients and improve the applicability and accuracy of the model, the federated learning framework is used to optimize the risk assessment model. The performance comparison results before and after optimization are given below, as shown in Table 2.

[0032] Table 2. Performance comparison results

[0033] The permission adaptation management unit is used to comprehensively consider multi-dimensional permission attributes, permission policy rules, and the risk level output by the artificial intelligence risk assessment module, and dynamically determine and adjust the final permission adaptation decision for the access request; Furthermore, the process of the permission adaptation management unit dynamically determining and adjusting the final permission adaptation decision for the access request includes: Integrate the multi-dimensional permission attributes included in the access request, the optimized permission policy rules, and the real-time risk level output by the artificial intelligence risk assessment module; Based on the predefined decision logic, adjust the initial access permission according to the real-time risk level and the optimized permission policy rules; According to the output of the decision logic, determine the specific access control execution instruction and generate the final permission adaptation decision including the access control execution instruction.

[0034] The access control unit is used to enforce the corresponding access control operations according to the final permission adaptation decision by the CRM-GIS platform, restricting or allowing access to the data and functions within the CRM-GIS platform.

[0035] Although the embodiments of the present invention have been shown and described, for those of ordinary skill in the art, it can be understood that various changes, modifications, substitutions, and variations can be made to these embodiments without departing from the principles and spirit of the present invention. The scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A multi-dimensional permission dynamic adaptation method for CRM-GIS, characterized in that, Including: Within the CRM-GIS platform, define and obtain in real time multi-dimensional permission attributes related to access requests; According to the preset permission policy rules stored in the platform, use the policy decision engine to process the access request to evaluate the initial access permission; Use the artificial intelligence risk assessment module configured in the platform to analyze the real-time risk level of the user's current operation based on the user's current spatio-temporal information and behavior pattern information, and output the risk level as a decision adjustment factor to the policy decision engine; The policy decision engine comprehensively considers the multi-dimensional permission attributes, permission policy rules, and the risk level output by the artificial intelligence risk assessment module, and dynamically determines and adjusts the final permission adaptation decision for the access request; According to the final permission adaptation decision, the CRM-GIS platform enforces corresponding access control operations to restrict access to data and functions within the platform.

2. The multi-dimensional permission dynamic adaptation method for CRM-GIS according to claim 1, characterized in that The multi-dimensional permission attributes include: identity and environmental attributes related to the subject initiating the access, object data attributes related to the access object, operation attributes of the request execution, and context environment attributes representing the current operation context; among them, the context environment attributes include the user's current spatio-temporal information and the behavior pattern information of the user's interaction with the CRM-GIS platform.

3. A multi-dimensional permission dynamic adaptation method for CRM-GIS according to claim 1, characterized in that, The process of using the artificial intelligence risk assessment module configured in the platform to analyze the real-time risk level of the user's current operation based on the user's current spatio-temporal information and behavior pattern information includes: Real-time collect spatio-temporal information related to the current user operation and the behavior pattern information of the user's interaction with the CRM-GIS platform; Process the spatio-temporal information and the behavior pattern information into a structured input feature vector; Input the input feature vector into a pre-trained artificial intelligence model for processing. The artificial intelligence model infers the risk index of the current operation by analyzing the deviation degree of the input feature vector from the preset normal spatio-temporal range and the established user historical behavior baseline; Based on the inference result of the artificial intelligence model, generate a quantified real-time risk level and provide it to the policy decision engine as a decision adjustment factor to assist in dynamically adjusting the final permission adaptation decision.

4. A multi-dimensional permission dynamic adaptation method for CRM-GIS according to claim 1, characterized in that The dynamic optimization of the artificial intelligence risk assessment module and the preset permission policy rules is realized through a distributed machine learning framework that protects data privacy; the distributed machine learning framework allows multiple participants to jointly train and update the global risk assessment model and permission policy optimization model on the premise of protecting the privacy of their respective original data.

5. A multi-dimensional permission dynamic adaptation method for CRM-GIS according to claim 4, characterized in that, The distributed machine learning framework is a federated learning framework. The process of using the federated learning framework to jointly optimize the artificial intelligence risk assessment module and the preset permission policy rules includes: The central coordination server distributes the initial global risk assessment model structure and global permission policy meta-model to multiple clients; Each client uses the original private data stored locally related to the operation of the CRM-GIS platform to perform local training on the received model to generate local model update parameters, and the original private data does not leave the local area; The central coordination server collects and aggregates the local model update parameters from multiple clients to iteratively update the global risk assessment model and the global permission policy meta-model, so as to make access control decisions without directly accessing the original data of each client.

6. A multi-dimensional permission dynamic adaptation method for CRM-GIS according to claim 1, characterized in that The process in which the policy decision engine dynamically determines and adjusts the final permission adaptation decision for the access request by integrating multi-dimensional permission attributes, permission policy rules, and the risk level output by the artificial intelligence risk assessment module includes: Integrating the multi-dimensional permission attributes included in the access request, the optimized permission policy rules, and the real-time risk level output by the artificial intelligence risk assessment module; Based on the predefined decision logic, adjusting the initial access permission according to the real-time risk level and the optimized permission policy rules; According to the output of the decision logic, determining the specific access control execution instruction and generating the final permission adaptation decision including the access control execution instruction.

7. A multi-dimensional permission dynamic adaptation system for CRM-GIS, characterized in that, Including: A multi-dimensional permission attribute acquisition unit for defining and real-time acquiring multi-dimensional permission attributes related to the access request within the CRM-GIS platform; An initial access permission analysis unit for evaluating the initial access permission by using the policy decision engine to process the access request according to the preset permission policy rules stored in the CRM-GIS platform; A risk level analysis unit for analyzing the real-time risk level of the user's current operation based on the user's current spatio-temporal information and behavior pattern information by using the artificial intelligence risk assessment module configured in the CRM-GIS platform; A permission adaptation management unit for dynamically determining and adjusting the final permission adaptation decision for the access request by integrating multi-dimensional permission attributes, permission policy rules, and the risk level output by the artificial intelligence risk assessment module; An access control unit for enforcing the corresponding access control operation by the CRM-GIS platform according to the final permission adaptation decision to restrict or allow access to the data and functions within the CRM-GIS platform.

Images