Privacy data desensitization method based on big data algorithm

Abstract

The application discloses a privacy data desensitization method based on a big data algorithm, and relates to the technical field of data desensitization.The application analyzes a plurality of access log records of a plurality of pre-access personnel, determines all dynamic desensitization departments of different static features based on the access quantity and access time of the business personnel of different business departments to the static features during the analysis process, adopts a dynamic desensitization mode to query and desensitize the feature data of the static features that are not frequently accessed by different departments, and deletes the desensitized data that is queried and desensitized by the dynamic desensitization mode from a storage management module.Through the mode, on the one hand, the unreasonable increase of the storage space occupancy caused by the processing mode of one department and one desensitized copy is avoided, and on the other hand, when the original data is updated or the desensitization rules are adjusted, it is not necessary to synchronously update all copies, so that the problems such as data redundancy and version confusion are avoided.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of data desensitization, and particularly to a privacy data desensitization method based on a big data algorithm. BACKGROUND

[0002] With the rapid development of digital technology, the scale of enterprise data grows exponentially, which contains a large amount of sensitive information related to user privacy and business secrets, such as personal identification numbers, bank account information, transaction records, enterprise customer data, etc. To meet the current requirements for data privacy protection and support the data use needs of internal business departments (such as development testing, data analysis, business auditing, etc.), privacy data desensitization technology has become a core means of data security governance.

[0003] Static data, as an important part of enterprise data assets, usually refers to data that does not change frequently within a certain period and can be stored and reused for a long time, such as historical transaction records, user basic information archives, and archived business contracts. For the desensitization of static data, existing technologies usually use pre-defined desensitization rules (such as masking, replacement, encryption, and generalization) to process the original static data offline, generate a copy of the desensitized static data, and store it in a non-production environment (such as a test database or an analysis database) for access and use by business personnel in different business departments.

[0004] However, different business departments within an enterprise have significant differences in function and authority, and their use of static data also varies. For example, the sales department needs to maintain customer information based on customer names and contact information, so the desensitization level needs to retain some information availability. The finance department needs to rely on transaction amounts and payment accounts to check accounts, so the desensitization rules need to adapt to the accuracy requirements of financial data. The administrative department only needs to count the number of users and other macro information, so the desensitization process can significantly simplify the data content. To meet these different needs, existing technologies usually generate multiple copies of data with different desensitization levels for the same batch of original static data based on the function and authority of different business departments, and store them separately.

[0005] This one-department-one-desensitization-copy processing method will result in a significant increase in storage space usage, especially when the scale of static data reaches PB, the excessive copy data will significantly increase the storage cost and data management difficulty of the enterprise. At the same time, the consistency of each desensitized copy cannot be guaranteed. If the original data is updated or the desensitization rules are adjusted, all copies need to be updated synchronously, which may cause data redundancy, version confusion, and other problems, affecting the accuracy and efficiency of data use.

[0006] To solve the above problems, the present application provides a solution. SUMMARY

[0007] The present application aims to provide a privacy data desensitization method based on big data algorithm.

[0008] The present application provides a privacy data desensitization method based on big data algorithm, comprising the following steps:

[0009] Step one: the data access module obtains the search information of the current pre-visitor and the business department of the post, generates an access request for the current pre-visitor according to the search information, and transmits it to the access response module;

[0010] Step two: the access response module extracts the search information and the business department after receiving the access request, parses the search information to determine the query subject and the query condition, converts it into an executable query statement, and transmits the query statement, the query subject and the business department to the storage management module;

[0011] Step three: the storage management module stores attribute data of several types of static attributes, and each type of static attribute contains all feature data of several static features. After receiving the transmitted query statement, query subject and business department, it determines whether the feature whitelist of the business department is stored and whether the query subject is contained in the feature whitelist according to the query subject and the business department. If it is contained, the feedback result data of the query statement is transmitted to the access response module according to the preset dynamic query rule;

[0012] Step four: the access response module displays the feedback result data of the query statement after receiving it, and generates an access log record based on the time when the access request is received, the feedback result data and the business department of the current pre-visitor, and transmits it to the log analysis module for storage;

[0013] Step five: the log analysis module analyzes the stored access log records when the number of access log records reaches a fixed amount, selects several business departments from all business departments of the target enterprise as dynamic desensitization departments for any one static feature, and transmits them to the storage management module;

[0014] Step six: the storage management module generates a feature whitelist for these dynamic desensitization departments after receiving all the dynamic desensitization departments of all static features, and the feature whitelist contains several static features.

[0015] Further, the log analysis module displays all the dynamic desensitization departments of all static features selected to the management personnel, and the management personnel compiles a second desensitization script for all static features based on all dynamic desensitization departments and inputs it into the storage management module for storage.

[0016] Further, step three, the query obtains the feedback result data of the query statement Dynamic query rule as follows:

[0017] The execution of the query statement obtains several characteristic data, and then for each characteristic data obtained, the query subject executes the second desensitization script based on the business department to obtain several desensitization data as the feedback result data of the query statement.

[0018] Further, the first desensitization script based on all business departments of the target enterprise is also stored in the storage management module.

[0019] Further, for any one characteristic data of any one static feature contained in the static data, when it is stored into the storage management module, the first desensitization script based on all business departments of the target enterprise is executed to obtain the desensitization data of the characteristic data based on all business departments and store it.

[0020] Further, step six, for any one characteristic data of any one static feature, based on all dynamic desensitization departments of the static feature, delete the characteristic data stored in the storage management module based on all desensitization data of all dynamic desensitization departments.

[0021] Further, step five, the analysis step is as follows:

[0022] S11: All static features contained in the static data of the target enterprise which have not been executed desensitization operation are marked as A1, A2,..., Aa, a≥1, all business departments of the target enterprise are marked as B1, B2,..., Bb, b≥1;

[0023] S12: Extract the total number of log record data whose corresponding static feature is A1 and business department is B1 from the log analysis module, and mark the total number as the department demand amount C1 of business department B1 based on static feature A1, and similarly, obtain the department demand amount C2, C3,..., Cb of business department B2, B3,..., Bb based on static feature A1;

[0024] S13: Compare the department demand amount C1 with the preset demand amount threshold P1, if the department demand amount C1 is greater than or equal to P1, select business department B1 as the reference department of static feature A1, otherwise do not do anything, similarly, compare department demand amount C2, C3,..., Cb with P1, and obtain all reference departments of static feature A1 based on the comparison result, and mark them as D1, D2,..., Dd, 1≤d≤b.

[0025] S14: Extract all the log record data corresponding to the feature data corresponding to the desensitization data from the log analysis module, and the static feature is A1 and the reference reference department is D1, and the extracted all log record data is marked as E1, E2,..., Ee, e>=1, in the order of storing the log analysis module from the beginning to the end; Extract the time F1, F2,..., Fe contained in the log record data E1, E2,..., Ee in turn;

[0026] S15: According to the time F1, F2,..., Fe, determine the frequency measurement index F1 of the reference reference department D1 based on the static feature A1;

[0027] S16: Determine whether the reference reference department D1 is a dynamic desensitization department of the static feature A1:

[0028] Compare the frequency measurement index F1 with the preset frequency reference threshold P2, if the frequency measurement index F1 is greater than or equal to P2, select the reference reference department D1 as the reference desensitization department of the static feature A1, otherwise do not do anything;

[0029] S17: According to S14 to S16, determine whether the reference reference department D2, D3,..., Dd is the reference desensitization department of the static feature A1 in turn, and obtain all the reference desensitization departments of the static feature A1 based on the determination result, and obtain all the business departments of the target enterprise except the all reference desensitization departments, and mark the business department as the dynamic desensitization department of the static feature A1;

[0030] S18: According to S11 to S17, obtain all the dynamic desensitization departments of the static feature A2, A3,..., Aa.

[0031] Compared with the prior art, the following beneficial effects are achieved:

[0032] The application analyzes a plurality of access logs of a plurality of pre-access personnel, and determines all dynamic desensitization departments of different static features based on the number of accesses of business personnel of different business departments to each static feature and the access time during the analysis process. The feature data of the static feature that is not frequently accessed by different departments is queried and desensitized in a dynamic desensitization manner, and the desensitized data queried and desensitized in a dynamic desensitization manner is deleted from the storage management module. In this way, on the one hand, the unreasonable increase of storage space caused by the processing method of one department and one desensitization copy is avoided, and on the other hand, when the original data is updated or the desensitization rule is adjusted, it is not necessary to update all the copies synchronously, thereby avoiding the problems of data redundancy, version confusion and the like;

[0033] The application deletes the desensitized data obtained by querying desensitization in a dynamic desensitization manner from the storage management module, retains the storage of the desensitized data of all reference desensitization departments based on different static characteristics, and enables the access request of different business departments to efficiently retrieve the corresponding desensitized data or generate the desensitized data in real time through dynamic desensitization based on actual needs, which not only ensures the efficiency and stability of high-frequency desensitized data access, but also meets the access needs of dynamic desensitization departments, and further strengthens the fine management of data storage, realizes the optimal configuration of storage resources and flexible adaptation of data access under the premise of ensuring data security compliance. BRIEF DESCRIPTION OF DRAWINGS

[0034] Figure 1 The method flowchart of the application. DETAILED DESCRIPTION

[0035] The technical solutions in the embodiments of the application will be clearly and completely described below with reference to the drawings in the embodiments of the application. Obviously, the described embodiments are only part of the embodiments of the application, rather than all the embodiments of the application. Based on the embodiments in the application, all other embodiments obtained by those skilled in the art without creative labor fall within the scope of protection of the application.

[0036] Please refer to Figure 1 The application provides a privacy data desensitization method based on a big data algorithm, which is executed based on a privacy data desensitization system based on a big data algorithm, and the system comprises a data access module, a storage management module, an access response module and a log analysis module.

[0037] The data access module is used for business personnel in a plurality of business departments in a target enterprise to access static data after desensitization, and the static data after desensitization refers to static data after execution of a first desensitization script or a second desensitization script.

[0038] The static data contains attribute data of various static attributes formed by the target enterprise in the operation process, which are not dynamically changed with real-time business operation and have relative stability, including but not limited to enterprise basic information, personnel basic information, customer and supplier basic information, product and service basic data, asset basic information and historical business archive information.

[0039] The attribute data of any type of static attribute contains all feature data of a plurality of static features.

[0040] It should be noted that any feature data of any static feature has a corresponding relationship with the static feature.

[0041] A number of static features associated with enterprise basic information include but are not limited to enterprise name, unified social credit code, registered address, legal representative, date of establishment, registered capital, scope of business, etc.; a number of static features associated with personnel basic information include but are not limited to employee name, ID number, date of employment, department, job title, education information, etc.; a number of static features associated with customer and supplier basic information include but are not limited to, such as customer name, supplier name, unified social credit code, contact address, contact number, cooperation start date, etc.; a number of static features associated with product and service basic data include but are not limited to product name, model, specification, classification, benchmark unit price, service item name, service content, service standard, etc.; a number of static features associated with asset basic information include but are not limited to fixed asset number, name, specification and model, purchase date, original value, storage location, intangible assets (including patents, trademarks) registration number, right holder, validity period, etc.; a number of static features associated with historical business file information include but are not limited to completed contract text, historical transaction records, project closing report, and archived financial statements, etc.

[0042] Due to the preset differences in the specific business functions assumed by different business departments within the target enterprise, the business personnel belonging to each business department have different levels of detail of data content based on the functional authority of their corresponding department, which is configured to have corresponding differences; wherein the difference in the level of detail of the data content is that for the same static feature, the desensitization granularity of the feature data that can be viewed by the business personnel of different business departments is different:

[0043] For example: the sales department needs to directly interface with customers and follow up on orders, so it can view the complete mobile phone number 13812341234 of the customer;

[0044] The human resources department needs to verify the authenticity of the employee's cooperative customers, and only needs to view the complete information of the customer company name, but the mobile phone number of the customer contact may be processed as 138***1234;

[0045] It should be noted here that accessing desensitized static data requires the business personnel to be in a logged-in state, i.e., the business personnel has completed identity authentication and can know the business department they are in charge of;

[0046] After the current pre-access personnel enter the search information, the data access module obtains the search information and the business department of the current pre-access personnel, generates an access request for the current pre-access personnel according to the obtained search information and the business department, and transmits the access request to the access response module; the search information includes a number of search keywords;

[0047] The storage management module is used for storing and optimizing the static data before and after the target enterprise is desensitized. The storage management module stores the static data. The storage management module also stores all static features contained in the static data based on the first desensitization script of all business departments of the target enterprise. The first desensitization script is used to desensitize the feature data corresponding to the static feature. The first desensitization script specifies the specific rules, scope, method and constraint conditions for executing desensitization on the feature data, so as to ensure that the desensitization behavior meets the business requirements and safety standards, and specifically includes desensitization objects, granularity levels, rule types, etc. The granularity level specifies different desensitization depths for the feature data (such as specifying “hide middle 4 digits”, “hide last 4 digits”, “full hide” and the like for mobile phone desensitization), and the rule type specifies a specific desensitization algorithm or method, including but not limited to: mask replacement, character replacement, intervalization, randomization, truncation, etc.

[0048] It should be noted that for any one feature data of any one static feature, the first desensitization script of the static feature is executed only once when the feature data is stored into the storage management module.

[0049] It should be noted that for any one feature data of any one static feature contained in the static data, the first desensitization script based on all business departments of the target enterprise is executed when the feature data is stored into the storage management module to obtain desensitized data of the feature data based on the all business departments and store the desensitized data.

[0050] It should be noted that any one desensitized data has a corresponding relationship with the feature data obtained by executing the preset first desensitization script to obtain the desensitized data.

[0051] In this application, the storage mode of the static data before and after desensitization is to store the data in a database: the data is organized in a structured form and saved in a database system, which is convenient to maintain, but requires a tool to view;

[0052] The access response module is used for responding to the access request of the business personnel of all business departments in the target enterprise. After the access response module receives a transmitted access request of a current pre-access personnel, the access response module extracts the search information and the business department contained in the access request, analyzes the search information, determines the query subject and the query condition, and converts them into an executable query statement. The query statement, the query subject and the business department are transmitted to the storage management module. In this application, the query subject is represented as a static feature.

[0053] The storage management module receives the transmitted query statement, the query subject and the business department, and first determines whether the feature whitelist of the business department is stored and whether the query subject is contained in the feature whitelist according to the query subject and the business department. If the query subject is contained, the feedback result data of the query statement is obtained according to the preset dynamic query rule, and the dynamic query rule is as follows:

[0054] The query statement is executed to obtain a plurality of feature data, and then the second desensitization script of the business department based on the query subject is executed for each piece of feature data to obtain a plurality of desensitization data, which is transmitted to the access response module as the feedback result data of the query statement.

[0055] If the feature whitelist of the business department is not stored or the query subject is not contained in the feature whitelist of the business department, a plurality of desensitization data is obtained by executing the query statement, and all desensitization data based on the business department is extracted from the plurality of desensitization data. The all desensitization data is transmitted to the access response module as the feedback result data of the query statement.

[0056] The access response module receives the transmitted feedback result data of the query statement, and displays it to the current pre-access personnel for viewing. During the viewing process, an access log record is generated based on the time when the access request is received, the feedback result data and the business department of the current pre-access personnel, and the access log record is transmitted to the log analysis module for storage.

[0057] The log analysis module is used to analyze all the access log records stored therein when the number of the access log records stored therein reaches a fixed amount. The value of the fixed amount is selected according to the following principles: avoiding one-sided analysis conclusion caused by insufficient data (such as too small sample size to reflect the real rule), avoiding analysis delay or resource waste caused by excessive data accumulation, and ensuring that the analysis result is representative and practical.

[0058] The analysis steps are as follows:

[0059] S11: All static features contained in the static data of the target enterprise without desensitization operation are marked as A1, A2,..., Aa, respectively, a≥1, and all business departments of the target enterprise are marked as B1, B2,..., Bb, respectively, b≥1.

[0060] S12: Extract the total number of log record data of the business department B1 based on the static feature A1 from the log analysis module, and mark the total number as the department demand C1 of the business department B1 based on the static feature A1. Similarly, the department demand C2, C3,..., Cb of the business department B2, B3,..., Bb based on the static feature A1 is obtained in turn.

[0061] S13: Compare the department demand C1 with the preset demand threshold P1. If the department demand C1 is greater than or equal to P1, select the business department B1 as the reference department of the static feature A1. Otherwise, do not perform any processing. Similarly, compare the department demand C2, C3,..., Cb with P1 in turn, and obtain all reference departments of the static feature A1 based on the comparison result, and mark them as D1, D2,..., Dd respectively, 1≤d≤b. The value of P1 is selected based on the time length consumed for storing a fixed amount of access log records in the log analysis module.

[0062] S14: Extract the log record data of the reference department D1 based on the static feature A1 from the log analysis module, and mark all the extracted log record data as E1, E2,..., Ee in turn according to the order of storage into the log analysis module, e≥1.

[0063] Extract the time F1, F2,..., Fe contained in the log record data E1, E2,..., Ee in turn.

[0064] S15: Determine the frequency measurement index F1 of the reference department D1 based on the static feature A1 according to the time F1, F2,..., Fe, and the specific content is as follows:

[0065] Calculate the time difference G1, G2,..., Ge-1 of the time F1 and F2, F2 and F3,..., Fe-1 and Fe using the formula Gg=Fg+1-Fg, and the value of g is 1, 2,..., e-1 in turn.

[0066] Perform data processing on the time difference G1, G2,..., Ge-1 using the discrete point filtering algorithm, and calculate the average value of all time difference values after data processing. Mark the average value as the frequency measurement index F1 of the reference department D1 based on the static feature A1.

[0067] In this application, the discrete point filtering algorithm is H-score filtering algorithm.

[0068] S16: Determine whether the reference department D1 is the reference desensitization department of the static feature A1, and the determination content is as follows:

[0069] The frequency measurement index F1 and the preset frequency reference threshold P2 are compared in size, and if the frequency measurement index F1 is greater than or equal to P2, the reference reference department D1 is selected as the reference desensitization department of the static feature A1, otherwise no processing is performed;

[0070] S17: Determine whether the reference reference department D2, D3,..., Dd is the reference desensitization department of the static feature A1 in turn according to S14 to S16, obtain all reference desensitization departments of the static feature A1 based on the determination result, and obtain all business departments of the target enterprise except the all reference desensitization departments, and mark the business departments as dynamic desensitization departments of the static feature A1;

[0071] S18: Obtain all dynamic desensitization departments of the static features A2, A3,..., Aa in turn according to S11 to S18;

[0072] The log analysis module displays all dynamic desensitization departments of the static features A1, A2,..., Aa obtained by analysis to the administrator, and transmits them to the storage management module;

[0073] The administrator writes a second desensitization script for all dynamic desensitization departments of the static features A1, A2,..., Aa, and the second desensitization script clearly specifies the specific rules, range, method and constraint conditions for performing desensitization on feature data, to ensure that the desensitization behavior meets the business requirements and security standards, including desensitization objects, granularity levels, rule types, etc.

[0074] After receiving the transmitted all dynamic desensitization departments of the static features A1, A2,..., Aa, the storage management module first generates a feature whitelist for these dynamic desensitization departments, and the feature whitelist of any one of the dynamic desensitization departments contains the corresponding static feature;

[0075] For example, if all dynamic desensitization departments of the static features A1 and A3 contain the dynamic desensitization department Z1, the feature whitelist of the dynamic desensitization department Z1 contains the static features A1 and A3.

[0076] For any one feature data of any one of the static features A1, A2,..., Aa, based on all dynamic desensitization departments of the static feature, the feature data stored in the storage management module is deleted based on all desensitization data of the all dynamic desensitization departments.

[0077] Some data in the above formula are dimensionless numerical calculations, and the contents not described in detail in the specification all belong to the prior art known to those skilled in the art.

[0078] The above examples are only used to illustrate the technical method of the present application but not limit the present application. Although the present application is described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical method of the present application can be modified or equivalently replaced without departing from the spirit and scope of the technical method of the present application.

Claims

1. A privacy-sensitive data de-identification method based on big data algorithms, characterized in that: Includes the following steps: Step 1: The data access module obtains the search information of the current pre-accessing person and the business department where the person works, and generates an access request for the current pre-accessing person based on this information, and transmits it to the access response module. Step 2: After receiving the access request, the access response module extracts the retrieval information and business department, parses the retrieval information to determine the query subject and query conditions, converts it into an executable query statement, and transmits the query statement, query subject, and business department to the storage management module. Step 3: The storage management module stores attribute data of several types of static attributes. The attribute data of any type of static attribute contains all the feature data of several static features. After receiving the transmitted query statement, query subject and business department, it determines whether the feature whitelist of the business department is stored according to the query subject and business department, and whether the feature whitelist contains the query subject. If it does, the feedback result data of the query statement is obtained according to the preset dynamic query rules and transmitted to the access response module. Step 4: After receiving the feedback result data of the query statement, the access response module displays it and generates an access log record based on the time when the access request was received, the feedback result data, and the business department of the current pre-accessing person. The log log record is then transmitted to the log analysis module for storage. Step 5: After the number of storage access log records reaches a fixed amount, the log analysis module analyzes them. For any static feature, it selects several business departments from all business departments of the target enterprise as the dynamically anonymized departments for the static feature, and transmits them to the storage management module. The analysis steps are as follows: S11: Label all static features contained in the static data of the target company that have not undergone de-identification as A1, A2, ..., Aa, where a≥1; label all business departments of the target company as B1, B2, ..., Bb, where b≥1; S12: Extract the total number of log records with static feature A1 and business department B1 corresponding to all de-identified data from the log analysis module. Define this total number as the departmental demand C1 for business department B1 based on static feature A1. Similarly, obtain the departmental demand C2, C3, ..., Cb for business departments B2, B3, ..., Bb based on static feature A1. S13: Compare the department demand C1 with the preset demand threshold P1. If the department demand C1 is greater than or equal to P1, select business department B1 as the benchmark reference department for static feature A1. Otherwise, do not process it. Similarly, compare the department demand C2, C3, ..., Cb with P1 in turn. Based on the comparison results, obtain all benchmark reference departments for static feature A1 and re-label them as D1, D2, ..., Dd, where 1≤d≤b. S14: Extract all log record data with static feature A1 and reference department D1 corresponding to the feature data of all de-identified data from the log analysis module. Mark all extracted log record data as E1, E2, ..., Ee, e≥1, according to the order of storage of the log analysis module. Sequentially extract the time points F1, F2, ..., Fe contained in the log record data E1, E2, ..., Ee. S15: Determine the reference department D1 based on the frequency measurement index F1 of static feature A1 according to time F1, F2, ..., Fe; S16: Determine whether the benchmark reference section D1 is a dynamic desensitization section of static feature A1: The frequency measurement index F1 is compared with the preset frequency benchmark threshold P2. If the frequency measurement index F1 is greater than or equal to P2, the benchmark reference department D1 is selected as the benchmark desensitization department of static feature A1; otherwise, no processing is performed. S17: Determine whether the benchmark reference departments D2, D3, ..., Dd are benchmark desensitized departments of static feature A1 in sequence according to S14 to S16. Based on the determination results, obtain all benchmark desensitized departments of static feature A1. Obtain all business departments other than the benchmark desensitized departments from all business departments of the target enterprise, and mark the business departments as dynamic desensitized departments of static feature A1. S18: Obtain all dynamic desensitization parts of static features A2, A3, ..., Aa in sequence from S11 to S17; Step Six: After receiving all the dynamic desensitization departments of all the transmitted static features, the storage management module generates a feature whitelist for all the dynamic desensitization departments. The feature whitelist contains several static features.

2. The privacy data anonymization method based on big data algorithms according to claim 1, characterized in that, The log analysis module displays all dynamically de-identified departments for all selected static features to the administrator. The administrator then writes a second de-identification script for all static features based on all dynamically de-identified departments and inputs it into the storage management module for storage.

3. The privacy data anonymization method based on big data algorithms according to claim 2, characterized in that, Step 3: The dynamic query rules for obtaining the feedback results of the query statement are as follows: The query statement is executed to obtain several pieces of feature data. Then, for each piece of feature data, the query subject executes the second de-identification script based on the business department to obtain several pieces of de-identified data, which are used as the feedback result data of the query statement.

4. The privacy data desensitization method based on big data algorithms according to claim 1, characterized in that, The storage management module also stores the first de-identification script based on all static features contained in the static data and all business departments of the target enterprise.

5. The privacy data anonymization method based on big data algorithms according to claim 4, characterized in that, For any single feature data of any static feature contained in the static data, when it is stored in the storage management module, the first de-identification script based on all business departments of the target enterprise is executed to obtain the de-identified data of the feature data based on all business departments and then stored.

6. The privacy data desensitization method based on big data algorithms according to claim 1, characterized in that, Step 6: For any single feature data of any static feature, based on all dynamic desensitization departments of the static feature, delete the feature data stored in the storage management module based on all desensitized data of all dynamic desensitization departments.

Images