Desensitization data evaluation method and device, electronic equipment and storage medium

Abstract

The invention relates to the technical field of information security, in particular to a desensitized data evaluation method and device, electronic equipment and a storage medium. The method comprises the steps that firstly, data desensitization demand information is acquired, and the data desensitization demand information can comprise a desensitization user identifier, an application scene applying desensitization data and a desensitization demand corresponding to the application scene; secondly, determining a corresponding data desensitization algorithm according to the data desensitization demand information, and generating a desensitization evaluation standard according to a preset evaluation index; wherein each application scene has a corresponding data desensitization algorithm. Finally, the desensitization effect can be evaluated according to the desensitization evaluation standard to generate a data desensitization result. In the process of evaluating the desensitization data, the evaluation standard of the desensitization data is determined in combination with the application scene of the sensitive information, and the scene and application of desensitization data evaluation are improved.

Description

Technical Field

[0001] The present invention relates to the field of information security technology, and in particular to a method, device, electronic device and storage medium for evaluating desensitized data. Background Art

[0002] With the development of communication technology, people are more and more concerned about information security. In order to protect the information security of users, various industries have continuously proposed personalized data desensitization solutions according to their specific needs. Among them, data desensitization refers to the deformation of sensitive information through specific rules or algorithms to protect personal privacy and the security of commercial sensitive data. Common scenarios for the application of sensitive information can be application development, testing and outsourcing environments. In order to ensure the safe use of real data sets in these non-production environments, it is necessary to desensitize the data used in these environments. For example, sensitive information includes: identity card number, mobile phone number, card number, customer number and other personal information.

[0003] In recent years, with the release of a series of laws and industry standards, the links and scenarios of personal information de-identification processing have been clarified. Furthermore, various types of data desensitization products have appeared on the market. These products can meet common desensitization scenarios by integrating multiple desensitization tools. In this case, it is still necessary to effectively evaluate the quality of the desensitized data obtained and the degree of protection of the desensitized data for user sensitive information.

[0004] In some implementation scenarios, privacy inference attacks are carried out on desensitized data, and inference attacks are carried out on desensitized data in combination with multiple sensitive attributes to achieve evaluation of desensitized data. However, this solution only focuses on the evaluation of the desensitization algorithm, and does not effectively evaluate the actual data results after desensitization. In other implementation scenarios, the real data in the production environment is "moved and simulated and replaced" before being sent to downstream links for access and reading and writing, and the desensitized data is isolated from the production environment. In actual applications, this method cannot process a variety of structured files, which limits the analysis and evaluation of various forms of desensitized data. Moreover, these implementation methods cannot evaluate the adaptability of different data application scenarios. Summary of the invention

[0005] The present invention provides a method, device, electronic device and storage medium for evaluating desensitized data, so that in the process of evaluating desensitized data, its evaluation criteria are determined in combination with the application scenarios of sensitive information, thereby improving the scenarios and applications of desensitized data evaluation.

[0006] In order to achieve the above technical objectives, the present invention adopts the following technical solutions:

[0007] In the first aspect, an embodiment of the present invention provides a method for evaluating desensitized data, the method comprising: first obtaining data desensitization demand information, the data desensitization demand information may include a desensitized user identifier, an application scenario for applying desensitized data, and a desensitization demand corresponding to the application scenario. Secondly, a corresponding data desensitization algorithm is determined according to the data desensitization demand information, and a desensitization evaluation standard is generated according to a preset evaluation indicator. Finally, the desensitization effect can be evaluated according to the desensitization evaluation standard to generate a data desensitization result.

[0008] Specifically, the combination of user identification and application scenario is determined according to the evaluation requirements, that is, whether the data used in the application scenario meets the requirements of desensitized data and whether it affects the application scenario. This evaluation method can provide comprehensive evaluation results and ensure the effectiveness and security of desensitization processing, promoting the balance between data privacy protection and data availability in the application scenario.

[0009] In combination with the first aspect, in a possible implementation, in a single desensitized user single application scenario, a corresponding data desensitization algorithm is determined according to the data desensitization requirement, and a desensitization evaluation standard is generated according to a preset evaluation indicator. This implementation may specifically include: determining a target user identifier based on data desensitization requirement information, and the target user identifier is used to determine the target user. It can be understood that the target user is the privacy data of a specific user, and the desensitization effect is evaluated. Sample data and historical data are obtained, wherein the historical data includes historical evaluation data and historical desensitized data. The corresponding data desensitization algorithm is determined according to the data desensitization requirement information. Furthermore, the sample data is desensitized according to the desensitization algorithm to generate sample data. A desensitization evaluation standard is generated based on the historical evaluation data and the sample data.

[0010] Among them, the evaluation method provided by the embodiment of the present invention is a desensitization evaluation based on application scenarios, and an appropriate desensitization algorithm can be adaptively selected for different application scenarios. This implementation method can reliably evaluate different sensitive information application scenarios, so that the evaluation results meet the privacy protection requirements and can take into account the actual availability of the data.

[0011] It should be noted that the degree of desensitization of sensitive information and the privacy protection requirements of sensitive information are different in different data application scenarios. Therefore, matching different desensitization algorithms and setting different evaluation standards for different application scenarios can maximize the protection of sensitive information while ensuring the reliability of the application scenarios of sensitive information.

[0012] In combination with the first aspect, in another possible implementation, in a single desensitized user multi-application scenario, the corresponding data desensitization algorithm is determined according to the data desensitization requirement, and the desensitization evaluation standard is generated according to the preset evaluation index. This implementation may specifically include: determining the target user identifier according to the data sensitivity requirement information, and the target user identifier is used to determine the target user. From multiple application scenarios, historical desensitization information and original data related to the target user identifier are obtained. The historical desensitization information is processed using a preset algorithm to obtain the processed desensitized data. A data mapping relationship is established between the original data and the processed desensitized data to generate a desensitization evaluation standard. It can be understood that for a single target user, in multiple sensitive information application scenarios, the desensitization of sensitive information may be different. For example, the target user identifier is a user identity identifier (Identity, ID). In the first application scenario, it is sufficient to mark the user ID as a legal identity; in the second application scenario, the user ID needs to be fully marked to provide the user with complete services. It can be understood that in these two application scenarios, the desensitization requirements for the user ID are different. Therefore, different algorithms are determined for different application scenarios to adapt to different application scenarios, and then different evaluation algorithms are adopted to evaluate the degree of desensitization in different application scenarios. This can protect the privacy of user data while ensuring the security of privacy data.

[0013] In combination with the first aspect, in another possible implementation, when multiple desensitized users are in multiple application scenarios, a corresponding data desensitization algorithm is implemented according to the data desensitization requirements, and a desensitization evaluation standard is generated according to preset evaluation indicators. This implementation may specifically include: determining the identity of the target user, and using a fuzzy matching algorithm to obtain the user information to be mined from the data desensitization requirement information. Among them, the user information to be mined includes at least two user identifiers. Obtain historical desensitized information and original data related to the user information to be mined, wherein the historical desensitized information is the historical desensitized data of each user to be mined (i.e., each user identifier) ​​in the user information to be mined in multiple application scenarios. Further, a preset algorithm is used to process the historical desensitized information to determine the user group characteristics of the user information to be mined, and the historical sensitive information is reduced according to the user group characteristics to obtain the target desensitized information. Determine the corresponding target original data according to the identity of the target user, and establish a data mapping relationship between the target original data and the target desensitized information to generate a desensitization evaluation standard.

[0014] Among them, determining the data desensitization information of the target user in multiple application scenarios can achieve a comprehensive analysis of the desensitized data, rather than relying on the data of a single user. This method can effectively integrate the desensitized data of all users, thereby reducing the potential risk of user privacy leakage in the desensitization algorithm and effectively and comprehensively protecting user privacy information.

[0015] In combination with the first aspect, in another possible implementation, a desensitization evaluation is performed based on a desensitization evaluation standard.

[0016] Specifically, for a single desensitized user in a single application scenario, data bias assessment, information loss assessment, reversibility assessment, and single desensitization effect feedback can be performed. For a single desensitized user in multiple application scenarios, data bias assessment and information loss assessment can be performed. For multiple desensitized users in multiple application scenarios, data bias assessment and information loss assessment can be performed.

[0017] In combination with the first aspect, in another possible implementation, in data deviation evaluation, the data modality may include: one or more of text data, table data, image data, audio data, video data, and trajectory graph data, etc. It is understandable that different data may be evaluated using different evaluation indicators.

[0018] Information loss evaluation can include: information entropy evaluation, mutual information evaluation and data correlation evaluation, etc.

[0019] Reversibility evaluation refers to the possibility of recovering sensitive information from desensitized data after being processed by the desensitization algorithm. It can effectively ensure the security of the desensitization process and the effectiveness of privacy protection.

[0020] In combination with the first aspect, in another possible implementation, the desensitization requirement information may include: user identification, various types of sensitive data information application scenarios, etc. Sensitive data information application scenarios may be desensitization requirement environments extracted from daily life. For example, 52 typical application scenarios are covered, including telephone / cable TV access, telecommunications service use, hotel check-in, training registration, and security monitoring.

[0021] In addition, the user identifier can be a user ID, a user email address, a user name, etc. Engineering personnel can preset these application scenarios and sensitive information into the evaluation system in the form of application templates. Each template can also provide multiple personal sensitive information fields and sensitive information types to ensure that the privacy input template can cover various privacy-related data to meet the privacy protection needs of different identities in different application scenarios. In other implementations, operators can also add, modify or delete sensitive information fields. Operators can also define desensitizing scenario templates based on specific application scenarios. In this way, the evaluation method can be more flexibly applied in different application scenarios.

[0022] In conjunction with the first aspect, in another possible implementation, before evaluating the desensitized data, a reasonable desensitization requirement can be set. That is, before the evaluation, the operator can set information such as the desensitization algorithm and the evaluation scenario according to the evaluation requirements.

[0023] It is understandable that in different application scenarios, different desensitization algorithms and privacy protection strategies are set for the same desensitized fields. Adaptive settings for the management of desensitized fields and desensitization algorithms for desensitized fields can provide users with a more flexible privacy protection method.

[0024] In combination with the first aspect, in another possible implementation, the method may also include: obtaining updated desensitization evaluation standards; comparing the updated desensitization evaluation standards and the desensitization effect, if the comparison result indicates that the desensitization effect meets the updated desensitization evaluation standards, then outputting the desensitization effect; if the comparison result indicates that the desensitization effect does not meet the updated desensitization evaluation standards, then adjusting the desensitization parameters and performing data desensitization operations.

[0025] It is understandable that in a single desensitized user and single application scenario, the evaluation standard can be iterated for a single evaluation result. For example, based on the multi-dimensional evaluation results of new files uploaded by users, the system will iteratively update the preliminary evaluation standards. In some implementations, the evaluation standards can be iterated in a way that the standard value of the existing standard accounts for 70% and the value of this evaluation result accounts for 30%. In addition, the evaluation system can also dynamically adjust the standard value using a weighted average method by calculating the deviation between the evaluation result and the existing standard.

[0026] In combination with the first aspect, in another possible implementation, the desensitization parameters can also be adjusted. For example, the data disturbance intensity can be increased, the encryption level can be improved, the randomness of the desensitization algorithm can be enhanced, etc., to further enhance the effect of privacy protection. If, after feedback analysis, it is found that the current desensitization algorithm cannot achieve the expected privacy protection goal, the user should consider switching to a desensitization technology that better meets the needs of the application scenario and re-execute the desensitization process.

[0027] It is understandable that using this feedback mechanism to update the evaluation results can optimize the desensitization process, so that the evaluation system can maintain the availability of desensitized data as much as possible while meeting the requirements of privacy data protection, and also improve the effectiveness and flexibility of the evaluation system.

[0028] In a second aspect, the present invention also provides a desensitized data evaluation device, including: an acquisition module, an evaluation module and a display module.

[0029] Among them, the acquisition module is used to obtain data desensitization demand information, which may include desensitization user identification, application scenarios for applying desensitized data, and desensitization requirements corresponding to the application scenarios. The evaluation module is used to determine the corresponding data desensitization algorithm according to the data desensitization demand information, and generate a desensitization evaluation standard according to the preset evaluation indicators. The display module is used to evaluate the desensitization effect according to the desensitization evaluation standard to generate a data desensitization result.

[0030] In conjunction with the second aspect, in a possible implementation, the evaluation module can also be used to determine a target user identifier based on data desensitization requirement information, and the target user identifier is used to determine a target user. It can be understood that the target user is to evaluate the desensitization effect of the privacy data of a specific user. Obtain sample data and historical data. Determine the corresponding data desensitization algorithm based on the data desensitization requirement information. Further, perform data desensitization on the sample data according to the desensitization algorithm to generate sample data. Generate a desensitization evaluation standard based on the historical evaluation data and the sample data.

[0031] In combination with the second aspect, in another possible implementation, the evaluation module can also be used to determine the target user identifier based on the data sensitivity requirement information, and the target user identifier is used to determine the target user. From multiple application scenarios, historical desensitized information and original data related to the target user identifier are obtained. The historical desensitized information is processed using a preset algorithm to obtain processed desensitized data. A data mapping relationship is established between the original data and the processed desensitized data to generate a desensitization evaluation standard.

[0032] In combination with the second aspect, in another possible implementation, the identifier of the target user is determined, and a fuzzy matching algorithm is used to obtain the user information to be mined from the data desensitization demand information. Wherein, the user information to be mined includes at least two user identifiers. Historical desensitized information and original data related to the user information to be mined are obtained, wherein the historical desensitized information is the historical desensitized data of each user to be mined (i.e., each user identifier) ​​in the user information to be mined in multiple application scenarios. Furthermore, a preset algorithm is used to process the historical desensitized information to determine the user group characteristics of the user information to be mined, and the historical sensitive information is narrowed down according to the user group characteristics to obtain the target desensitized information. The corresponding target original data is determined according to the identifier of the target user, and a data mapping relationship is established between the target original data and the target desensitized information to generate a desensitization evaluation standard.

[0033] In combination with the second aspect, in another possible implementation, the evaluation module can also be used to perform a desensitization evaluation based on the desensitization evaluation criteria.

[0034] Specifically, for a single desensitized user in a single application scenario, data bias assessment, information loss assessment, reversibility assessment, and single desensitization effect feedback can be performed. For a single desensitized user in multiple application scenarios, data bias assessment and information loss assessment can be performed. For multiple desensitized users in multiple application scenarios, data bias assessment and information loss assessment can be performed.

[0035] In combination with the second aspect, in another possible implementation, the evaluation module can also be used to obtain updated desensitization evaluation standards; compare the updated desensitization evaluation standards and the desensitization effect, and if the comparison result indicates that the desensitization effect meets the updated desensitization evaluation standards, output the desensitization effect; if the comparison result indicates that the desensitization effect does not meet the updated desensitization evaluation standards, adjust the desensitization parameters and perform data desensitization operations.

[0036] On the third aspect, the embodiment of the present application also provides a desensitized data evaluation device, including a desensitization demand management module, a single desensitization effect evaluation module, a desensitization effect evaluation module based on data mining, a desensitization system effect evaluation module and a result display module.

[0037] Among them, the desensitization requirement management module is used to obtain data desensitization requirement information, which may include desensitized user identification, application scenarios for applying desensitized data, and desensitization requirements corresponding to the application scenarios.

[0038] The single desensitization effect evaluation module is used to determine the target user ID according to the data sensitivity requirement information, and the target user ID is used to determine the target user. From multiple application scenarios, historical desensitization information and original data related to the target user ID are obtained. The historical desensitization information is processed using a preset algorithm to obtain processed desensitized data. A data mapping relationship is established between the original data and the processed desensitized data to generate a desensitization evaluation standard.

[0039] The desensitization effect evaluation module based on data mining is used to determine the target user ID according to the data sensitivity requirement information, and the target user ID is used to determine the target user. From multiple application scenarios, historical desensitization information and original data related to the target user ID are obtained. The historical desensitization information is processed using a preset algorithm to obtain processed desensitized data. A data mapping relationship is established between the original data and the processed desensitized data to generate a desensitization evaluation standard.

[0040] The desensitization system effect evaluation module is used to determine the identifier of the target user, and use a fuzzy matching algorithm to obtain the user information to be mined from the data desensitization demand information. Among them, the user information to be mined includes at least two user identifiers. Obtain historical desensitized information and original data related to the user information to be mined, wherein the historical desensitized information is the historical desensitized data of each user to be mined (i.e., each user identifier) ​​in the user information to be mined in multiple application scenarios. Furthermore, a preset algorithm is used to process the historical desensitized information to determine the user group characteristics of the user information to be mined, and the historical sensitive information is narrowed down according to the user group characteristics to obtain the target desensitized information. Determine the corresponding target original data according to the identifier of the target user, and establish a data mapping relationship between the target original data and the target desensitized information to generate a desensitization evaluation standard.

[0041] The result display module is used to display the desensitization result comparison and evaluation effect display.

[0042] In a fourth aspect, an embodiment of the present invention further provides an electronic device, comprising a processor and a memory for storing instructions executable by the processor, wherein the processor is configured to execute the instructions to implement the method in the first aspect and any possible implementation manner thereof.

[0043] In a fifth aspect, an embodiment of the present invention further provides a server, comprising a processor and a memory for storing instructions executable by the processor, wherein the processor is configured to execute the instructions to implement the method in the first aspect and any possible implementation manner thereof.

[0044] In a sixth aspect, an embodiment of the present invention further provides a computer-readable storage medium, on which computer instructions are stored. When the computer instructions are executed on an electronic device, the method in the above-mentioned first aspect and any possible implementation manner thereof is implemented.

[0045] It can be understood that the beneficial effects that can be achieved by the device in the second aspect and any possible implementation manner provided above, the equipment in the third aspect and any possible implementation manner, the electronic device in the fourth aspect, the server in the fifth aspect, and the computer-readable storage medium in the sixth aspect can be referred to as the beneficial effects in the first aspect and any possible design manner, and will not be repeated here. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] Figure 1 It is a schematic diagram of the structure of a desensitization evaluation system provided by an embodiment of the present invention;

[0047] Figure 2 This is a flow chart of a desensitized data evaluation method provided by an embodiment of the present invention;

[0048] Figure 3 is a flow chart of another method for evaluating desensitized data provided by an embodiment of the present invention;

[0049] Figure 4 is a flow chart of another method for evaluating desensitized data provided by an embodiment of the present invention;

[0050] Figure 5 This is a data deviation evaluation flow chart provided by an embodiment of the present invention;

[0051] Figure 6 is a flow chart of another method for evaluating desensitized data provided by an embodiment of the present invention;

[0052] Figure 7 is a flow chart of another method for evaluating desensitized data provided by an embodiment of the present invention;

[0053] Figure 8 It is a structural diagram of a server provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0054] In order to make the purpose, technical solutions and advantages of the present invention clearer, the present invention is further described in detail below in conjunction with the embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention. In the description of the embodiments of the present invention, unless otherwise specified, the meaning of "multiple" mentioned in the present invention is two or more.

[0055] The desensitization evaluation method provided by the embodiment of the present invention will be specifically described below in conjunction with the accompanying drawings. First, a system to which the desensitization evaluation method provided by the embodiment of the present invention is applied will be specifically described.

[0056] Please refer to Figure 1 A schematic diagram of the structure of a desensitization evaluation system provided by an embodiment of the present invention is shown in FIG. Figure 1 As shown, the desensitization evaluation system includes an external sensitive information desensitization part 10 and a desensitization effect evaluation part 20. Among them, the desensitization effect evaluation part 20 includes a desensitization demand management module 21, a single desensitization effect evaluation module 22, a desensitization effect evaluation module based on data mining 23, a desensitization system effect evaluation module 24 and a result display module 25.

[0057] The desensitized demand management module 21 includes a demand management module 21 - 1 and an application scenario management module 21 - 2 .

[0058] The single desensitization effect evaluation module 22 includes a data deviation evaluation module 22-1, an information loss evaluation module 22-2, a reversibility evaluation module 22-3 and a single desensitization evaluation feedback module 22-4.

[0059] The desensitization effect evaluation module 23 based on data mining includes a data bias evaluation module 23-1 and an information loss evaluation module 23-2.

[0060] The desensitizing system effect evaluation module 24 includes a data bias evaluation module 24-1 and an information loss evaluation module 24-2.

[0061] The result display module 25 includes a desensitization result comparison module 25-1 and an evaluation effect display module 25-2.

[0062] In some embodiments, the method provided by the embodiments of the present invention can cooperate with the external sensitive information desensitization part 10. For example, the external sensitive information desensitization part 10 can transmit the desensitization source data and the desensitization result to the desensitization effect evaluation part 20, and can also receive operations from the user to adjust the desensitization parameters. The external sensitive information desensitization part 10 can also receive the single desensitization effect evaluation feedback from the desensitization effect evaluation part 20.

[0063] The desensitization requirement management module 21 can obtain the desensitization requirement information input by the user.

[0064] Specifically, the requirement management module 21-1 presets a variety of desensitization algorithms and the corresponding desensitization requirements for each desensitization algorithm, and can match the appropriate desensitization algorithm and privacy protection strategy according to the application scenario of the sensitive data. And it also presets the desensitization algorithms for the desensitization fields in different application scenarios and the corresponding privacy protection levels of the desensitization algorithms.

[0065] The application scenario management module 21-2 presets the desensitized user identification and the application scenarios for applying the desensitized data. Among them, the application scenario management module 21-2 sets a desensitization requirement template, and in response to the user's input, operations such as adding, modifying, and deleting sensitive information fields can be performed on the desensitization requirement template. Exemplarily, the application scenario management module 21-2 presets 52 typical application scenarios including covering telephone / cable TV network access, telecommunications service use, hotel check-in, training registration, security monitoring, etc., and the main privacy-sensitive fields for each application scenario set. For example, the desensitization requirement template provides no less than 15 personal sensitive information fields and their types, ensuring that the desensitization requirement template can cover various privacy-related data to meet the privacy protection requirements in different identities and different application scenarios.

[0066] Exemplarily, the embodiments of the present invention list 52 typical application scenarios. Please refer to Table 1 below for examples of 52 typical application scenarios.

[0067] Table 1 52 Typical Application Scenarios Table

[0068]

[0069]

[0070]

[0071]

[0072]

[0073] The single desensitization effect evaluation module 22 can be used to perform desensitization evaluation in a single subject and single application scenario.

[0074] Specifically, the data deviation evaluation module 22 - 1 can be used to measure the degree of distortion and deviation of the desensitized data after being processed by the desensitization algorithm.

[0075] The information loss assessment module 22-2 is used to measure the impact of the privacy information loss in the desensitized data on the data availability.

[0076] The reversibility evaluation module 22-3 is used to measure the possibility of restoring the privacy information in the desensitized data, ensuring the security of the desensitization process and the effectiveness of privacy protection.

[0077] The single desensitization evaluation feedback module 22 - 4 is used to determine whether the desensitization operation has achieved the expected privacy protection goals and data availability requirements by evaluating the effect of the desensitization operation.

[0078] The desensitization effect evaluation module 23 based on data mining can be used to perform desensitization evaluation in single-subject multi-application scenarios.

[0079] The desensitization system effect evaluation module 24 can be used to perform desensitization evaluation in multi-subject and multi-application scenarios.

[0080] The result display module 25 is used to display the desensitization evaluation results. The desensitization evaluation results include desensitization result comparison and desensitization effect data display.

[0081] The desensitization result comparison module 25-1 is used for desensitization result comparison and is mainly responsible for comprehensive comparison and analysis of original data and desensitized data, and providing a comparison table of original data and desensitized data.

[0082] The evaluation effect display module 25-2 is used to present the evaluation results generated in the evaluation phase in a visual manner, helping users to understand the desensitization effect more intuitively.

[0083] It should be noted that the above-mentioned single desensitization effect evaluation module 22, the data mining-based desensitization effect evaluation module 23 and the desensitization system effect evaluation module 24 all include a data bias evaluation module and an information loss evaluation module. Although they belong to different modules, their own functions remain unchanged, and the module can adaptively adjust the specific execution data.

[0084] Those skilled in the art can understand that the structural diagram of the desensitization evaluation system shown in the embodiment of the present invention does not constitute a limitation on the structure of the desensitization evaluation system. In practical applications, the desensitization evaluation system may also include more or fewer modules shown in the diagram, or may combine certain modules, or may rearrange the modules. Through the description of the above implementation mode, those skilled in the art can clearly understand that for the convenience and simplicity of description, only the division of the above-mentioned functional modules is used as an example. In practical applications, the above-mentioned functions can be assigned to different functional modules as needed, that is, the internal structure of the device can be divided into different functional modules to complete all or part of the functions described above.

[0085] Please refer to Figure 2 , is a flow chart of a method for evaluating desensitized data provided by an embodiment of the present invention. Figure 2 As shown, the method includes steps 201 to 203.

[0086] It should be noted that, for the sake of ease of understanding, the desensitized data evaluation method is combined with Figure 1 The desensitization evaluation system shown is specifically described.

[0087] Step 201: Obtain data desensitization requirement information.

[0088] The data desensitization requirement information may include a desensitized user identifier, an application scenario for applying the desensitized data, and a desensitization requirement corresponding to the application scenario.

[0089] Specifically, since the embodiment of the present invention takes into account the number of target users and the application scenarios of sensitive information when it is applied, the data desensitization requirement information is obtained to determine the target users and application scenario information. In this way, the evaluation of desensitization technology in multiple application scenarios and multiple sensitive data can be effectively solved.

[0090] In some implementations, multiple application scenarios and sensitive data are preset in the desensitization evaluation system. Before conducting desensitized data evaluation, operators can set up demand templates to meet the differences in data requirements for different privacy protection algorithms and application scenarios. Different desensitization algorithms and privacy protection strategies are matched according to application scenarios, and desensitization algorithms and their corresponding privacy protection levels for common desensitized fields in various scenarios are managed. Users can flexibly add, delete, and modify the desensitization algorithms and their privacy protection levels in the templates.

[0091] Step 202: Determine the corresponding data desensitization algorithm according to the data desensitization requirement information, and generate a desensitization evaluation standard according to the preset evaluation indicators.

[0092] It is worth mentioning that the evaluation method will be adaptively adjusted due to the different number of target users. Therefore, even for one target user, the algorithm rules of the privacy field are different in different application scenarios. In order to adaptively adjust the data desensitization algorithm, different evaluation methods can be used for different evaluation needs.

[0093] Step 203: Evaluate the desensitization effect according to the desensitization evaluation standard to generate a data desensitization result.

[0094] In some implementations, the data desensitization results include desensitization result comparison and evaluation effect display. Specifically, the desensitization result comparison is mainly responsible for the comprehensive comparison and analysis of the original data and the desensitized data, and provides a comparison table of the original data and the desensitized data. The evaluation effect display presents the evaluation results generated in the evaluation phase in a visual way to help users more intuitively understand the desensitization effect.

[0095] When the desensitization evaluation system displays the desensitization results for comparison, the operator can view the specific changes in each field row by row and column by column. In some implementations, a query function can also be added to the desensitization evaluation system to intuitively compare the desensitization results of sensitive fields that the operator is more concerned about.

[0096] In other implementations, interactive functions are provided to operators during the display of desensitization evaluation results, providing query functions, allowing users to enter sensitive field names, quickly locate and view desensitization results of related fields, and setting highlighting for each field for easy viewing by operators.

[0097] In other implementations, unstructured data comparison generates thumbnails and playback functions for unstructured data such as images, videos, and audio. For trajectory graphics, map trajectory demonstration is provided to show the intuitive changes before and after data processing.

[0098] In other implementations, an export function is provided to allow users to export the comparison results into Excel or PDF format for easy archiving and sharing.

[0099] In addition, for the desensitization of unstructured data, such as images, videos, trajectory graphics, etc., the system displays the data differences before and after desensitization through thumbnails, playback functions, map trajectory demonstrations, etc., thereby helping operators to intuitively feel the effect of desensitization of privacy information.

[0100] In addition, a variety of data display methods are used, including charts, tables, trajectory graphics, etc., to dynamically display the differences between the data before and after desensitization. According to different data types and comparison dimensions, visual charts such as line charts, bar charts, and scatter plots are automatically generated to display changes in data statistics, distribution characteristics, and correlation. Users can use these visualization methods to clearly observe the deviations of data before and after desensitization, such as the change curve of information entropy, the change trend of mutual information, and the change in the shape of data distribution.

[0101] In other implementations, the information amount change analysis displays the information entropy, mutual information, and data distribution shape change curves to quantify the degree of interdependence between the data before and after desensitization.

[0102] In other implementations, data distribution display uses visualization methods to show the distribution changes of data before and after desensitization, including distribution shape, quantiles, and cumulative distribution functions, so that users can analyze the impact of desensitization operations on data characteristics.

[0103] First, in conjunction with the desensitization test system, this article introduces the process of setting up the demand template by operators during the desensitization demand management phase. Figure 3 , is a flow chart of another method for evaluating desensitized data provided by an embodiment of the present invention. Figure 3 The flowchart shown includes an operator 31 , an interaction module 32 , a processing module 33 and a database 34 .

[0104] It should be noted that the interaction module 32, the processing module 33 and the database 34 are modular divisions of the desensitization evaluation system for the convenience of introducing the evaluation method of desensitized data, which is only exemplary. In other descriptions, the interaction module 32 can also be called the front end, and the processing module 33 can be called the back end.

[0105] like Figure 3 As shown, the desensitized data evaluation method includes steps 301 to 313.

[0106] The interaction module 32 receives the specific application scenario and data flow scenario input by the operator 31. 301: Specific application scenario and data flow scenario.

[0107] The interaction module 31 sends a request for obtaining database requirement template data to the processing module 33. 302: Request to obtain requirement template data.

[0108] The processing module 33 sends a demand template data query request to the database 34. 303: Query the demand template data.

[0109] The database 34 responds to the requirement template data query request and responds: 304: Return the requirement template data.

[0110] The processing module 33 returns the demand template data to the interaction module 32. 305: Return the demand template data.

[0111] 306: The interaction module 32 displays a demand template editing interface.

[0112] 307: The operator 31 provides a scene-adaptive desensitization control set in the editing interface.

[0113] 307: The interaction module 32 determines whether the operator inputs the desensitization control settings. If yes, execute 308: the system will use these user-defined settings; if no, execute 309: the system will automatically use the default values.

[0114] 310: Verify the format of the desensitization control set to determine whether the format of the desensitization control set meets the predetermined standard. If it meets the standard, execute 311 to persistently store the response data. 312: Store the data; if it does not meet the standard, execute 313: throw a format error exception and terminate the execution of the program.

[0115] Among them, data persistence means that the verified data will be persistently stored in the MySQL (a database name) database in the Structured Query Language (SQL). The system stores the response data in a safe and reliable manner according to the pre-negotiated format for subsequent data evaluation and analysis.

[0116] After the requirement template management is completed, the system starts the evaluation operation. The following will specifically illustrate the evaluation method of the corresponding desensitized data in combination with the requirements of the desensitization test.

[0117] In the first implementation, a method of desensitizing data is set in an application scenario for a single desensitized user and a single application scenario.

[0118] Please refer to Figure 4 , is a flow chart of another method for evaluating desensitized data provided by an embodiment of the present invention. Figure 4 The flowchart shown includes an operator 31, an interaction module 32, a processing module 33, a database 34, and an external sensitive information desensitization system 35. Figure 4 As shown, the desensitized data evaluation method includes steps 401 to 414.

[0119] 401: Determine whether the data comes from an external system.

[0120] If yes, execute 402: send desensitized data, desensitization results and desensitization control set data. 403: parse and process the received data.

[0121] If not, execute 404: upload the desensitized original data and desensitization results. 405: submit user data. 406: extract the desensitization control set data set set by the user.

[0122] 407: Verify the consistency of the masked data format before and after masking.

[0123] 408: Data preparation completed.

[0124] 409: Data bias assessment.

[0125] In some implementations, the data before and after desensitization can be processed as follows: Figure 5 As shown in the process. Figure 5 As shown, it is a data deviation evaluation flow chart provided in an embodiment of the present invention.

[0126] Among them, the system will conduct a comprehensive bias analysis on the data before and after desensitization. Exemplarily, the data analysis methods include: determining data modality analysis and statistical bias analysis. Determine the data modality and divide the data modality into text, table, graphic, video, audio and trajectory graphic data. According to different data modalities, select appropriate evaluation indicators for bias evaluation. Statistical bias analysis, compare the statistical indicators of the original data and the desensitized data.

[0127] For text data, the desensitized source data is considered as character text A, and the desensitized data is considered as character text B. The following similarity indicators are calculated:

[0128] Jaccard distance: By comparing the common and different elements in the text, through the formula: Calculate and get the quantitative text similarity.

[0129] Levenshtein distance: Calculates the edit distance between characters, that is, the minimum number of operations required to transform one string into another.

[0130] Cosine similarity: Using the vector space model, calculate Reflects the similarity between texts.

[0131] Hamming distance: Calculates the number of different character positions between two strings of equal length To assess the differences between the texts.

[0132] For table data, extract the corresponding data columns before and after desensitization, and regard each column as a data set. For the data set before desensitization X and the data set after desensitization Y, n represents the number of data points, x represents the number of data points, and i represents the i-th data point in the X data set, y i Represents the i-th data point in the Y data set, and the following indicators can be calculated:

[0133] Mean Square Error (MSE): Measures the mean squared difference between data sets.

[0134] Mean Absolute Error (MAE): Reflects the mean absolute deviation between data sets.

[0135] Euclidean distance: Quantify the geometric distance between two sets of data points.

[0136] KL divergence: Evaluate the difference between the original data distribution and the masked data distribution. i ) represents the data point x in the P data set i The probability, P(y i ) represents the data point y in the P data set i The probability of , etc.

[0137] If the column data is numeric data, it can be regarded as a data set containing n data points, and the value of each data point is represented by x. 1 ,x 2 ,...,x means that more statistical values ​​can be calculated:

[0138] Mean:

[0139] Median: When n is an odd number,

[0140] When n is an even number,

[0141] variance: Population variance: Among them, s 2 represents the sample variance, σ 2 represents the population variance, represents the average, x i represents the value of each data point in the data set, μ represents the overall mean, and n represents the total number of data points.

[0142] Standard Deviation: Population standard deviation: Among them, s represents the sample standard deviation, σ represents the population standard deviation, represents the average, x i represents the value of each data point in the data set, μ represents the overall mean, and n represents the total number of data points.

[0143] Maximum value: Max=max(x 1 ,x 2 ,…,x n ).

[0144] Minimum value: Min = min(x i ,x i ,…,x n ).

[0145] Image data, also consider the images before and after desensitization as two data sets X and Y, and calculate the following indicators:

[0146] Peak signal-to-noise ratio: Here, MAX represents the maximum possible value of the signal (usually 255 for 8-bit images), and MSE represents the mean square error, which is the average difference between the two data sets.

[0147] Structural similarity index: SSIM(x,y) = [l(x,y)*c(x,y)*s(x,y)]^α, where x and y represent two images, l(x,y) represents brightness similarity, c(x,y) represents contrast similarity, s(x,y) represents structural similarity, and α is a parameter (usually 1).

[0148] Video data, video data can be regarded as the superposition of multiple frames of images. Frame extraction can be performed first. The peak signal-to-noise ratio and structural similarity index can also be calculated to evaluate the desensitization effect of the video data.

[0149] For audio data, we first need to extract features from the original audio and the desensitized audio, and represent the feature vectors as A and B respectively.

[0150] Cosine similarity: Using the vector space model, calculate

[0151] Euclidean distance:

[0152] Trajectory graphics, first you need to extract the trajectory as coordinate points (x, y), you will get two sets of trajectory data, the original trajectory is {(x 1 ,y 1 ),(x 2 ,y 2 ),…,(x n ,y n )}, the desensitization trajectory is {(x′ 1 ,y′ 1 ),(x′ 2 ,y′ 2 ),…,(x′ n ,y′ n )}. The following metrics are calculated:

[0153] Euclidean distance: The average Euclidean distance of all points can be calculated:

[0154] Path length difference: Calculate the path length of the original trajectory and the desensitized trajectory, The differences between the original trajectory length and the desensitized trajectory length were then compared.

[0155] Distribution bias analysis compares the distribution differences between the original data and the anonymized data. By calculating the distribution shape, quantile, cumulative distribution function (CDF), etc., the system will generate a specific distribution bias assessment report based on the mode and characteristics of the data.

[0156] Overall data bias analysis integrates the above analysis results, calculates the overall bias score, and comprehensively considers the bias of each mode. The score can be calculated based on the weighted average method, and the specific formula is:

[0157]

[0158] Among them, w i is the weight of each mode, D i is the deviation index of each mode.

[0159] 410: Information loss assessment. The system evaluates the possible information loss caused by the desensitization process by analyzing the change in the amount of information between the original data and the desensitized data.

[0160] Information entropy analysis: The system calculates the information entropy of the original data and the desensitized data, compares their differences, and evaluates the loss of information. For data set X, the information entropy calculation formula is: Among them, H(X) represents the information entropy of the random variable X, n is the number of values ​​of the random variable, P(x i ) is a random variable with value x i probability.

[0161] The random variable takes the value x i probability.

[0162] After that, compare the information entropy difference between the original data set X and the desensitized data set Y:

[0163] ΔH=H(X)-H(Y)

[0164] If ΔH>0, it indicates information loss, otherwise it indicates information gain.

[0165] Mutual information analysis: the system calculates the mutual information between the original data and the desensitized data, and quantitatively evaluates the mutual dependence between the data before and after desensitization. Calculate the mutual information between the original data X and the desensitized data Y:

[0166]

[0167] Through this analysis, it is possible to determine the impact of the desensitization process on the internal structure and correlation of the data.

[0168] Data correlation analysis: The system uses the correlation coefficient metric to evaluate the correlation changes between the data before and after desensitization and calculates the correlation coefficient:

[0169]

[0170] Where Cov(X,Y) is the covariance of X and Y, σ X and σ Y is the standard deviation of X and Y.

[0171] For overall information loss analysis, the system can calculate the overall information loss score by combining the scores of various indicators. The specific calculation method is as follows:

[0172] Information entropy loss score

[0173] Mutual Information Loss Score

[0174] Data association loss score S C , by comparing the changes in correlation coefficients before and after desensitization:

[0175]

[0176] Among them, r desensitizef and r original are the correlation coefficients of the original data and the anonymized data, respectively.

[0177] Calculate the overall information loss score:

[0178] S T =w H ·S H +w I ·S I +w C ·S C

[0179] Among them, w H 、w I 、w C is the weight of the corresponding score, which can be adjusted according to the specific application scenario and data characteristics to reflect the importance of different indicators. T The value range is usually between 0 and 1. T =0 means the information loss is minimal, and the desensitized data is almost the same as the original data. T =1 indicates the greatest information loss, and the desensitized data has lost almost all the original information.

[0180] 411: Reversibility evaluation. The system will evaluate the reversibility and parameter strength of the desensitization algorithm at this stage.

[0181] Desensitization algorithm reversibility assessment: the system will analyze whether the algorithm used in the desensitization operation is reversible. If the algorithm is irreversible desensitization, the system will evaluate the extent to which it meets the irreversible desensitization criteria.

[0182] Desensitization algorithm parameter strength evaluation: For different desensitization algorithms, the system will conduct encryption strength evaluation on key parameters. For encryption algorithms, the ratio of key length is calculated:

[0183]

[0184] Among them, L is the key length currently used, L min It is the minimum key length under the security standard.

[0185] Information restoration analysis: The system will try to restore the original private information through the desensitized data, analyze the accuracy of the restoration, and use the accuracy A to calculate: Where TP is the number of true positives and FP is the number of false positives.

[0186] 412: Determine whether there is user feedback. If so, execute 413: Return the desensitization evaluation results to indicate the deficiencies of the desensitization operation. If not, the external system feedback is executed, and 414: Feedback the desensitization evaluation results to help optimize the desensitization process.

[0187] In the second implementation, the method of desensitizing data is set in the application scenario of single desensitized user and multiple application scenarios.

[0188] Please refer to Figure 6 This is a flow chart of another method for evaluating desensitized data provided by an embodiment of the present invention. Figure 6 As shown, the desensitized data evaluation method includes steps 601 to 609.

[0189] It should be noted that when implementing the data desensitization evaluation in multiple application scenarios of a single desensitized user, historical data has been collected through sensitive information collection and data mining of a single user for a specific application scenario.

[0190] Step 601: Determine the identifier of the target user to be mined.

[0191] The target user's identifier includes information such as user ID, user email or name. This operation can ensure the accuracy and effectiveness of subsequent data extraction. The selection of identifiers should take into account the uniqueness and reliability of the data in order to accurately locate user data.

[0192] Step 602: For the structured data in the database, a query statement is constructed using the SQL language to obtain the desensitized data related to the target identifier.

[0193] Specifically, this query should cover single desensitization effect evaluation data in multiple scenarios to ensure that the user's historical desensitization data in different scenarios can be extracted.

[0194] Step 603: Determine whether the text document contains text data. If yes, execute step 604; if no, execute step 605.

[0195] Step 604: Use regular expressions to perform information recognition and extract text data.

[0196] Step 605: Obtain the desensitized historical data of the target user in multiple scenarios.

[0197] Step 606: For all the extracted historical desensitized data, potential user group characteristics are mined and the scope of the desensitized data is narrowed.

[0198] In some implementations, cluster analysis is performed on all the extracted historical desensitized data using data mining algorithms such as DBSCAN. By identifying dense areas in the data, potential user group characteristics are mined and the scope of desensitized data is narrowed.

[0199] Step 607: A mapping relationship is established, where the original information is regarded as data set A, and the corresponding desensitized data is regarded as data set B.

[0200] According to the selected identifier, the original data and the desensitized data after data mining clustering are mapped. The original information is regarded as data set A, and the corresponding desensitized data is regarded as data set B.

[0201] Step 608: Perform data bias assessment.

[0202] Step 609: Perform information loss evaluation.

[0203] Among them, the specific data bias evaluation and information loss evaluation can refer to the above-mentioned related descriptions and will not be elaborated here.

[0204] In the third implementation, the method of desensitizing data is set in the application scenario for multiple desensitized users and multiple application scenarios.

[0205] Please refer to Figure 7 This is a flow chart of another method for evaluating desensitized data provided by an embodiment of the present invention. Figure 7 As shown, the desensitized data evaluation method includes steps 701 to 710.

[0206] It should be noted that when implementing the data desensitization evaluation in multiple application scenarios for a single desensitized user, historical data has been collected for sensitive information collection and data mining for a single user in a specific application scenario. System historical data collection is different from the historical data collection for a single subject in multiple scenarios. Here, the desensitized information of all subjects in various scenarios is mined from the entire database. By applying data mining technology, the system will analyze and integrate the desensitized data of all users and identify potential target user privacy information.

[0207] Step 701: Determine the identifier of the target user to be mined.

[0208] Step 702: Expand the target user's identifier range to cover the fuzzy matched identifier.

[0209] Step 703: For the structured data in the database, a query statement is constructed using the SQL language to obtain the desensitized data related to the target identifier.

[0210] Among them, for the structured data in the database, the SQL language is used to construct a query statement to obtain the desensitized data related to the target identifier. This query should cover the desensitization effect evaluation data of multiple scenarios and multiple users to ensure that the historical desensitized data of the target user in all scenarios can be extracted.

[0211] Step 704: Determine whether the text document contains text data. If yes, execute step 705; if no, execute step 706.

[0212] Step 705: Use regular expressions to perform information recognition and extract text data.

[0213] Step 706: Obtain the target user's anonymized historical data in multiple scenarios.

[0214] Step 707: For all the extracted historical desensitized data, potential user group characteristics are mined and the scope of the desensitized data is narrowed.

[0215] In some implementations, cluster analysis is performed on all the extracted historical desensitized data using data mining algorithms such as DBSCAN. By identifying dense areas in the data, potential user group characteristics are mined and the scope of desensitized data is narrowed.

[0216] Step 708: A mapping relationship is established, where the original information is regarded as data set A, and the corresponding desensitized data is regarded as data set B.

[0217] According to the selected identifier, the original data and the desensitized data after data mining clustering are mapped. The original information is regarded as data set A, and the corresponding desensitized data is regarded as data set B.

[0218] Step 709: Perform data bias evaluation.

[0219] Step 710: Perform information loss evaluation.

[0220] Among them, the specific data bias evaluation and information loss evaluation can refer to the above-mentioned related descriptions and will not be elaborated here.

[0221] The embodiment of the present invention further provides a computer storage medium, which includes computer instructions. When the computer instructions are executed on the electronic device, the electronic device performs the functions or steps in the method embodiment. For example, the computer readable storage medium can be a read-only memory (ROM), a random access memory (RAM), a compact disc (CD-ROM), a magnetic tape, a floppy disk, an optical data storage device, etc.

[0222] In the case of an integrated unit, Figure 8 A possible structural diagram of the server 800 involved in the above embodiment is shown. The server 800 may include: a processing module 801, a storage module 802 and a communication module 803. The processing module 801 is used to control and manage the actions of the server. The storage module 802 is used to save the program code and data of the server, such as the evaluation method of desensitized data, etc. Among them, the processing module 801 can be a processor or a controller. The communication module 803 can be a transceiver, an RF circuit or a communication interface, etc. The storage module 802 can be a memory.

[0223] Some other embodiments of the present invention provide a server, which may include: a memory and one or more processors. The memory is coupled to the processor. The memory is used to store computer program code, and the computer program code includes computer instructions. When the processor executes the computer instructions, the server may execute the various functions or steps executed by the server in the above method embodiment.

[0224] In the several embodiments provided by the present invention, it should be understood that the disclosed devices and methods can be implemented through other implementations. For example, the device embodiments described above are only schematic. For example, the division of the modules or units is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another device, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0225] In addition, the units described as separate components may or may not be physically separated, and the components shown as units may be one physical unit or multiple physical units, that is, they may be located in one place or distributed in multiple different places. Some or all of the units may be selected according to actual needs to achieve the purpose of the present embodiment.

[0226] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0227] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a readable storage medium. Based on this understanding, the technical solution of the embodiment of the present invention is essentially or the part that contributes to the prior art or all or part of the technical solution can be embodied in the form of a software product, which is stored in a storage medium and includes several instructions for a device (which can be a single-chip microcomputer, chip, etc.) or a processor (processor) to perform all or part of the steps of the method described in each embodiment of the present invention.

[0228] The above contents are only specific embodiments of the present invention, but the protection scope of the present invention is not limited thereto. Any changes or substitutions within the technical scope disclosed by the present invention should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.

Claims

1. A method for evaluating desensitized data, characterized in that: include: Obtaining data desensitization requirement information, the data desensitization requirement information including a desensitized user identifier, an application scenario for applying desensitized data, and a desensitization requirement corresponding to the application scenario; Determine the corresponding data desensitization algorithm according to the data desensitization requirement information, and generate a desensitization evaluation standard according to the preset evaluation indicators; The desensitization effect is evaluated according to the desensitization evaluation standard to generate the data desensitization result.

2. The method according to claim 1, characterized in that When the method is applied to a single desensitized user and a single application scenario, Determining the corresponding data desensitization algorithm according to the data desensitization requirement information, and generating a desensitization evaluation standard according to preset evaluation indicators, includes: Determine a target user identifier according to the data desensitization requirement information, where the target user identifier is used to determine the target user; Obtaining sample data and historical data, wherein the historical data includes historical evaluation data and historical anonymized data; Determine a corresponding data desensitization algorithm according to the data desensitization requirement information; Desensitizing the sample data according to the data desensitization algorithm to generate sample data; Generate a desensitization evaluation standard based on the historical evaluation data and the sample data.

3. The method according to claim 1, characterized in that When the method is applied to a single desensitized user with multiple applications, Determining the corresponding data desensitization algorithm according to the data desensitization requirement information, and generating a desensitization evaluation standard according to preset evaluation indicators, includes: Determine a target user identifier according to the data desensitization requirement information, where the target user identifier is used to determine the target user; Obtaining historical desensitized information and original data related to the target user identification from multiple application scenarios; Processing the historical desensitized information using a preset algorithm to obtain processed desensitized data; A data mapping relationship is established between the original data and the processed desensitized data to generate a desensitization evaluation standard.

4. The method according to claim 1, characterized in that When the method is applied to multiple desensitized users in multiple application scenarios, Determining the corresponding data desensitization algorithm according to the data desensitization requirement information, and generating a desensitization evaluation standard according to preset evaluation indicators, includes: Determine the target user's identifier, and use a fuzzy matching algorithm to obtain the user information to be mined from the data desensitization requirement information; wherein the user information to be mined includes at least two user identifiers; Acquire historical desensitized information and original data related to the user information to be mined, wherein the historical desensitized information is historical desensitized data of each user to be mined in the user information to be mined in multiple application scenarios; Using a preset algorithm to process the historical desensitized information to determine the user group characteristics of the user information to be mined; Narrowing the historical desensitized information according to the user group characteristics to obtain target desensitized information; Determining corresponding target original data according to the identifier of the target user; A data mapping relationship is established between the target original data and the target desensitized information to generate a desensitization evaluation standard.

5. The method according to any one of claims 1 to 4, characterized in that: The method further comprises: A desensitization evaluation is performed according to the desensitization evaluation standard; among them, data deviation evaluation and information loss evaluation are performed.

6. The method according to claim 2, characterized in that The method further comprises: Perform a reversibility test on the sample data and / or historical desensitized data to determine the possibility of restoring the private information.

7. The method according to any one of claims 1 to 4, characterized in that: The method further comprises: Obtain updated desensitization evaluation standards; Compare the updated desensitization evaluation standard and the desensitization effect. If the comparison result indicates that the desensitization effect meets the updated desensitization evaluation standard, output the desensitization effect; if the comparison result indicates that the desensitization effect does not meet the updated desensitization evaluation standard, adjust the desensitization parameters and perform the data desensitization operation.

8. A desensitized data evaluation device, characterized in that: include: an acquisition module, an evaluation module, and a display module; The acquisition module is used to acquire data desensitization requirement information, wherein the data desensitization requirement information includes a desensitized user identifier, an application scenario for applying desensitized data, and a desensitization requirement corresponding to the application scenario; The evaluation module is used to determine the corresponding data desensitization algorithm according to the data desensitization requirement information, and generate a desensitization evaluation standard according to the preset evaluation indicators; The display module is used to evaluate the desensitization effect according to the desensitization evaluation standard to generate a data desensitization result.

9. An electronic device, characterized in that: include: a processor and a memory for storing instructions executable by the processor; The processor is configured to execute the instructions to implement the method according to any one of claims 1 to 7.

10. A computer-readable storage medium having computer instructions stored thereon, characterized in that: When the computer instructions are executed on an electronic device, the method according to any one of claims 1 to 7 is implemented.

Images