Data abstraction classification processing method and apparatus, electronic device, and storage medium

Abstract

The application relates to the technical field of data processing, and provides a data abstract classification processing method and device, electronic equipment and a storage medium, wherein the method comprises the following steps: acquiring a data block, performing attribute analysis on the data block to obtain attribute values of various attributes in the data block; generating unique feature attributes of the data block based on the attribute value of each attribute in the data block and the composition information of the unique feature attributes of various categories; the composition information is composed of unique attributes or composed of a primitive attribute combination, the primitive attribute combination comprises a plurality of primitive attributes or comprises a plurality of primitive attributes and the relationship between the primitive attributes; determining the category to which the data block belongs based on the generation result of the unique feature attribute generation; and processing the data block based on the category to which the data block belongs. The method, device, electronic equipment and storage medium provided by the application ensure the reliability of abstract classification for the data block and the reliability and accuracy of data processing.

Description

Technical Field

[0001] This invention relates to the field of data processing technology, and in particular to a data abstraction and classification processing method, apparatus, electronic device, and storage medium. Background Technology

[0002] With the rapid development of information technology, the amount of data is growing explosively, and the complexity of data is constantly increasing.

[0003] Traditional data classification methods mainly classify data by judging whether there are specific contents or fields in the data, or by judging the contents associated with specific contents or fields associated with specific fields.

[0004] The above-mentioned solutions rely entirely on specific information, lack flexibility, and face many challenges when dealing with increasingly complex data. They are difficult to guarantee effective data classification and are prone to information leakage and data damage during data processing due to classification errors.

[0005] Therefore, improving the accuracy and reliability of classification-based data processing methods has become an urgent problem to be solved. Summary of the Invention

[0006] This invention provides a data abstraction and classification processing method, apparatus, electronic device, and storage medium to address the shortcomings of existing fixed data classification methods that cannot cope with complex and ever-changing data content, thus affecting the reliability of classification-based processing.

[0007] This invention provides a data abstraction and classification processing method, comprising: Obtain a data block, perform attribute parsing on the data block, and obtain the attribute values ​​of each attribute in the data block; Based on the composition information of each unique feature attribute of each category, and the attribute value of each attribute in the data block, a unique feature attribute is generated for the data block; the composition information is based on unique attributes or based on a combination of original attributes, the combination of original attributes includes multiple original attributes, or includes multiple original attributes and the relationship between the multiple original attributes. Based on the generation result generated by the unique feature attribute, the category to which the data block belongs is determined; The data block is processed based on the category to which it belongs.

[0008] According to a data abstraction and classification processing method provided by the present invention, the constitutive information of each unique feature attribute of each category is represented as a first knowledge graph. The first knowledge graph uses categories and attributes as nodes. The attributes connected to the categories include unique attributes and / or primitive attributes. Any unique feature attribute of the category is the unique attribute, or is a combination of primitive attributes. The combination of primitive attributes is determined based on multiple primitive attributes connected to the category, or based on the relationship between multiple primitive attributes connected to the category and multiple primitive attributes connected to the category.

[0009] According to a data abstraction and classification processing method provided by the present invention, the step of generating unique feature attributes for the data block based on the composition information of each unique feature attribute of each category and the attribute values ​​of each attribute in the data block includes: The attribute values ​​of each attribute in the data block are filled into the first knowledge graph to obtain the second knowledge graph; Based on the attribute nodes filled with attribute values ​​in the second knowledge graph and the node connection relationships in the second knowledge graph, unique feature attributes are generated for the data block.

[0010] According to a data abstraction and classification processing method provided by the present invention, the method generates unique feature attributes for the data block based on the composition information of each unique feature attribute of each category and the attribute values ​​of each attribute in the data block, and then further includes: Obtain the newly added data block of the data block, perform attribute parsing on the newly added data block, and obtain the attribute values ​​of each attribute in the newly added data block; The attribute values ​​of each attribute in the newly added data block are filled into the second knowledge graph to obtain the third knowledge graph; Based on the attribute nodes filled with attribute values ​​in the third knowledge graph and the node connection relationships in the third knowledge graph, a unique feature attribute is generated for the newly added data block.

[0011] According to a data abstraction and classification processing method provided by the present invention, the acquisition step of the first knowledge graph includes: Obtain the target category and the attributes under the target category; From the attributes, determine the unique attributes and / or original attributes of the target category; The original attributes are combined to obtain the original attribute combination, which serves as the unique feature attribute of the target category; The first knowledge graph is constructed based on the unique attributes of the target category and / or the combination of the original attributes.

[0012] According to a data abstraction and classification processing method provided by the present invention, the step of processing the data block based on the category to which the data block belongs includes: Based on the category to which the data block belongs and the data processing environment corresponding to the data block, the level of the category to which the data block belongs is determined; The data block is processed based on the category to which it belongs, and / or based on the level of the category to which it belongs; The data processing environment includes at least one of the following: the scenario, time, location, business system, application system, and industry in which the data block is processed.

[0013] According to a data abstraction and classification processing method provided by the present invention, the processing of the data block based on the category to which the data block belongs, and / or based on the level of the category to which the data block belongs, includes: The data block is processed based on the data processing environment corresponding to the data block, as well as the category to which the data block belongs and / or the level of the category to which the data block belongs.

[0014] According to a data abstraction and classification processing method provided by the present invention, the processing of the data block based on the category to which the data block belongs, and / or based on the level of the category to which the data block belongs, includes: Based on the category to which the data block belongs and / or the level of the category to which the data block belongs, or based on the data processing environment corresponding to the data block, and the category to which the data block belongs and / or the level of the category to which the data block belongs, a risk prediction is made on the data block to obtain the risk prediction result of the data block; Based on the risk assessment results of the data block, the data block is processed.

[0015] The present invention also provides a data abstraction and classification processing apparatus, comprising: A data acquisition unit is used to acquire data blocks, perform attribute parsing on the data blocks, and obtain the attribute values ​​of each attribute in the data blocks; A unique feature attribute generation unit is used to generate unique feature attributes for the data block based on the composition information of each unique feature attribute of each category and the attribute value of each attribute in the data block; the composition information is based on unique attributes or based on a combination of original attributes, the combination of original attributes includes multiple original attributes, or includes multiple original attributes and the relationship between the multiple original attributes. The category labeling unit is used to determine the category to which the data block belongs based on the generation result generated by the unique feature attribute; A data processing unit is used to process the data block based on the category to which the data block belongs.

[0016] The present invention also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the data abstraction and classification processing method as described above.

[0017] The present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the data abstraction and classification processing method as described above.

[0018] The present invention also provides a computer program product, including a computer program that, when executed by a processor, implements the data abstraction and classification processing method as described above.

[0019] The data abstraction and classification processing method, apparatus, electronic device, and storage medium provided by this invention identify each category with a unique characteristic attribute and set the composition information of the unique characteristic attribute based on the unique attribute or the combination of original attributes. Thus, based on the attribute values ​​of each attribute in the data block and the composition information of the unique characteristic attributes of each category, the unique characteristic attribute of the data block can be generated, thereby determining the category to which the data block belongs. In this process, the attribute analysis and unique characteristic attribute generation based on atomic granularity can cope with the classification requirements of various complex and variable data blocks, ensuring the reliability of the abstract classification of data blocks, thereby ensuring the reliability and accuracy of data processing based on abstract classification. Attached Figure Description

[0020] To more clearly illustrate the technical solutions in this invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0021] Figure 1 This is a flowchart illustrating the data abstraction and classification processing method provided by the present invention; Figure 2 This is one of the schematic diagrams of the first knowledge graph provided by the present invention; Figure 3 This is the second schematic diagram of the first knowledge graph provided by the present invention; Figure 4 This is a flowchart illustrating the method for acquiring the first knowledge graph provided by the present invention; Figure 5 This is a schematic diagram of the data abstraction and classification processing device provided by the present invention; Figure 6This is a schematic diagram of the structure of the electronic device provided by the present invention. Detailed Implementation

[0022] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0023] Traditional data classification methods primarily categorize data by determining the presence of specific content or fields, or by classifying data based on associations with specific content or fields. These methods rely entirely on specific information, lack flexibility, and face numerous challenges when processing increasingly complex data. They struggle to guarantee effective data classification and are highly susceptible to information leakage and data corruption due to classification errors.

[0024] To address the aforementioned problems, embodiments of the present invention provide a data abstraction and classification processing method. Figure 1 This is a flowchart illustrating the data abstraction and classification processing method provided by the present invention, as shown below. Figure 1 As shown, the method includes: Step 110: Obtain the data block, perform attribute parsing on the data block, and obtain the attribute values ​​of each attribute in the data block.

[0025] Specifically, a data block refers to the specific data to be processed. A data block can consist of data types with different modalities and characteristics. For example, a data block can be displayed in modalities such as text, images, and videos, or in structured and unstructured data formats. Furthermore, the size of a data block can vary, and this embodiment of the invention does not impose a specific limitation on it. Here, data blocks include, but are not limited to, any of the following: XML, HTML, graph databases, relational databases, TXT text, formatted documents (ofd, pdf, etc.), streaming documents (doc, docx, xls, xlsx, ppt, pptx, etc.), images (jpg, png, psd, bmp, etc.), audio (mp3, aac, ogg, etc.), and video (mp4, avi, flv, wmv, mov, etc.).

[0026] After obtaining the data block, you can perform attribute parsing on the data block to obtain the attribute values ​​of various attributes contained in the data block.

[0027] Here, attribute parsing of the data block can be achieved through scanning, recognition, or analysis. Recognition methods include, but are not limited to, manual recognition, rule matching, natural language processing, machine learning, and deep learning. The attributes in the data block are identified using these methods, and their corresponding attribute values ​​are obtained. For example, the ID card attribute in the data block can be identified using regular expressions to obtain the ID card number as the attribute value, or the address attribute in the data block can be identified using entity recognition technology to obtain the address information as the attribute value.

[0028] It is understandable that for a given data block, one or more attributes can be parsed, and the attribute values ​​of one or more attributes can be obtained within that data block. Furthermore, this data block may contain the same or different content at different times and in different scenarios, and the same or different attributes and their values ​​can be obtained through attribute parsing.

[0029] Step 120: Based on the composition information of each unique feature attribute of each category and the attribute value of each attribute in the data block, generate unique feature attributes for the data block; the composition information is based on unique attributes or based on a combination of original attributes, the combination of original attributes includes multiple original attributes, or includes multiple original attributes and the relationship between the multiple original attributes.

[0030] Specifically, after obtaining the attribute values ​​of each attribute in the data block, unique feature attributes can be generated for the data block based on the composition information of each unique feature attribute of each category and the attribute values ​​of each attribute in the data block.

[0031] Here, a category is an instantiated object of a class. A class is an abstract concept, and a class should possess at least one of the following unique attributes: a combination of original attributes, or a unique characteristic attribute to identify the class. Without violating the definition of a class, categories can have different granularities; for example, levels can be used as fine-grained categories. The number and granularity of categories can be set by the user.

[0032] Furthermore, a category is a compressed quantity that facilitates data management; the number of categories is far greater than the number of classes. There can be a concrete abstract mapping between categories and scenarios, and the mapping relationship between categories and scenarios can be the same or different in different fields. For example, students can correspond to grade sheets, leave slips, research achievement registration forms, etc. These forms can be understood as a category, and the form with specific values ​​corresponding to a particular student is a digital representation of the student as the subject.

[0033] Here, each category is specific and unambiguous. Categories may include, but are not limited to, at least one of the following: identity information, personal health information, personal resume information, financial information, asset information, location information, mobile communication information, instant messaging / network communication information, personal internet browsing history / network trace information, personal interpersonal relationship information, order information, consumption behavior information, voucher information, tax declaration records, device information, and personal preferences.

[0034] The unique characteristic attribute of a category can be a subset of the unique characteristic attributes of a class, the number of which is fixed and the attribute value is uncertain, that is, each category has a unique characteristic attribute that is both precise and unambiguous.

[0035] For any given category, that category can correspond to one or more unique feature attributes. A unique feature attribute is an attribute of that category, and it can serve as an identifier for that category, allowing the category to be uniquely identified.

[0036] For any unique feature attribute of any category, the constituent information of that unique feature attribute is used to reflect its specific content. In this embodiment of the invention, the unique feature attribute can be represented in two ways: it can be a unique attribute or a combination of original attributes. Correspondingly, the constituent information of the unique feature attribute can be composed of unique attributes or a combination of original attributes.

[0037] A unique attribute is a single, atomic attribute within a category, and is semantically unambiguous. Unique attributes are independent of unique features within other categories. As a special type of attribute, a unique attribute can constitute a unique feature as a single attribute. A unique attribute is a single tuple (including untuples), and a single unique attribute can be represented as... , For example, a single unique attribute can be represented as a triple <document type, attribute type, attribute value>, or a single unique attribute can be represented as a untuple <ID number>.

[0038] Primitive attribute combinations include multiple independent primitive attributes, or multiple related primitive attributes, as well as relationships between these related primitive attributes. Primitive attribute combinations are semantically unambiguous and independent of unique attribute features in other categories. Primitive attribute combinations can constitute unique feature attributes in the form of attribute combinations.

[0039] In this context, a primitive attribute is a single, atomic attribute within a category, semantically unambiguous. Unlike unique attributes, a single primitive attribute lacks the ability to represent a unique characteristic and cannot serve as a standalone category identifier; however, primitive attributes across different categories may overlap. A primitive attribute can be a single tuple (including a untuple), and a single primitive attribute can be represented as... , .

[0040] Within a category, relationships can exist between multiple primitive attributes. These relationships can be understood as ordered combinations of primitive attributes. Through these relationships, multiple primitive attributes that do not possess unique characteristics can be combined to form the unique characteristic attribute of the category. Here, the relationships between primitive attributes can be represented as multiple untuples or tuples, specifically: [< < ] in, .

[0041] In this embodiment of the invention, for each category, the unique characteristic attribute of each category can be predetermined, that is, the constitutive information used to constitute each unique characteristic attribute can also be predetermined.

[0042] Therefore, after determining the attribute values ​​of each attribute in the data block, the unique feature attributes of the data block can be generated based on the pre-determined composition information of each unique feature attribute of each category. That is, it can be determined which attribute or several attributes among the attributes contained in the data block can constitute a unique feature attribute.

[0043] Here, in the process of generating the unique feature attribute of the data block, the attributes contained in the data block can be matched with the composition information of the unique feature attributes of each category. Specifically, the attributes contained in the data block can be matched with the unique feature attributes of each category consisting of a single unique attribute, and the combination of the attributes contained in the data block can be matched with the unique feature attributes of each category consisting of multiple original attributes. The unique feature attribute that is successfully matched is taken as the unique feature attribute of the data block, which is the generation result of the unique feature attribute generation for the data block.

[0044] It is understood that the generated result here may not contain any unique feature attributes, that is, there are no unique feature attributes of any category in the data block; or, the generated result here may contain only one unique feature attribute; or, the generated result here may include multiple unique feature attributes, and the embodiments of the present invention do not specifically limit this.

[0045] Step 130: Based on the generation result generated by the unique feature attribute, determine the category to which the data block belongs.

[0046] Specifically, after obtaining the generation result of unique feature attributes for the data block, the category to which the data block belongs can be determined based on the unique feature attributes in the generation result and the pre-determined unique feature attributes of each category.

[0047] It is understood that if the generated result does not contain any unique feature attributes, it can be determined that the data block does not belong to any category; if the generated result contains only one unique feature attribute, it can be determined that the data block belongs to the category identified by that unique feature attribute, and in this case, the data block belongs to only one category; if the generated result includes multiple unique feature attributes, considering that multiple unique feature attributes can identify the same category or different categories, the data block can belong to one or more categories. This embodiment of the invention does not make specific limitations in this regard.

[0048] It is understandable that once the category to which the data block belongs is obtained, the abstract classification of the data block is completed.

[0049] In addition, there is a situation where the data blocks are pre-labeled with categories. In this case, by executing steps 110-130, the category determined based on the unique feature attribute of the data block can be obtained. If the category determined based on the unique feature attribute is consistent with the pre-labeled category, then the pre-labeled category can be determined to be correct.

[0050] Step 140: Process the data block based on the category to which it belongs.

[0051] Specifically, after determining the category to which the data block belongs, targeted data processing can be performed on the data block based on its category. Here, the processing methods for the data block can include de-identification, de-sensitization, deletion, filtering, and control, etc. In specific processing, at least one of the above processing methods can be performed, or other processing methods can be performed. This embodiment of the invention does not specifically limit this.

[0052] The method provided in this invention identifies each category with a unique characteristic attribute and sets the composition information of the unique characteristic attribute based on the unique attribute or the combination of original attributes. Thus, based on the attribute values ​​of each attribute in the data block and the composition information of the unique characteristic attributes of each category, the unique characteristic attribute of the data block can be generated, thereby determining the category to which the data block belongs. In this process, the attribute analysis and unique characteristic attribute generation based on atomic granularity can cope with the classification requirements of various complex and variable data blocks, ensuring the reliability of abstract classification of data blocks, thereby ensuring the reliability and accuracy of data processing based on abstract classification.

[0053] Based on the above embodiments, the identity information category may include at least one of the following categories: personal digital identity category, personal biometric identity category, personal online identity category, institutional digital identity category, and institutional online identity category.

[0054] Specifically, for the personal digital identity category, the unique attributes of the personal digital identity category may include a combination of unique attributes and / or intrinsic attributes.

[0055] The unique attribute can be represented as <document type, document number>, where document type is the unique attribute type and document number is the attribute value. <document type, document number> can specifically include, but is not limited to, the following: (1) <ID card, ID number>: consists of 18 digits, 6 regional code + 8 birth date + 3 personal code + 1 check code.

[0056] (2) <Passport, Passport Number>: Consists of English letters and numbers.

[0057] (3) <Driver's license, driver's license number>: consistent with the ID card number.

[0058] (4) <Social Security Card, Social Security Card Number>: consistent with the ID card number.

[0059] (5) <Passport, Reader Code>: consisting of “P” + Passport Type Code + Country Code + Name + Separator + Passport Number + Passport Number Verification Code + Country Code + Birth Date + Birth Date Verification Code + Gender + Passport Validity Period + Passport Validity Period Verification Code + Personal Code + Personal Code Verification Code + Verification Code.

[0060] (6) <Residence Permit, Residence Permit Number>: consistent with the ID card number.

[0061] The primitive attributes used to form combinations of primitive attributes may include, but are not limited to, the following: (1) <Work permit, work permit number>: can be composed of letters and numbers.

[0062] (2) <Document type, document validity period>: The time interval consisting of year, month and day, specifically including <ID card, document validity period>, <passport, document validity period>, <driver's license, document validity period>, <social security card, document validity period>, <residence permit, document validity period>, etc.

[0063] (3) <Document type, issuing authority>: including <ID card, issuing authority>, <passport, issuing authority>, <driver's license, issuing authority>, <social security card, issuing authority>, <residence permit, issuing authority>.

[0064] (4) <Document type, name>: Consists of surname and given name, including <ID card, name>, <passport, name>, <driver's license, name>, <social security card, name>, <residence permit, name>.

[0065] (5) <Document type, date of birth>: Consists of year, month and day, including <ID card, date of birth>, <passport, date of birth>, <driver's license, date of birth>, <social security card, date of birth>, <residence permit, date of birth>.

[0066] (6) <Document type, gender>: The attribute value is divided into two types: "male" and "female", including <ID card, gender>, <passport, gender>, <driver's license, gender>, <social security card, gender>, and <residence permit, gender>.

[0067] (7) <Document type, ethnicity>: Attribute values ​​include "Han ethnicity", including <ID card, ethnicity>, <social security card, ethnicity>, <residence permit, ethnicity>.

[0068] (8) <Document type, nationality>: Attribute values ​​include "China", including <passport, nationality>.

[0069] (9) Age: The attribute value is in the range of 0-150.

[0070] Based on the fundamental properties, the following combinations of fundamental properties can be formed, but are not limited to: (1) [<ID card, address>, <ID card, name>, <ID card, date of birth>, <ID card, validity period>] (2) [<passport, passport type>, <passport, gender>, <passport, name>, <passport, date of issue>, <passport, place of issue>, <passport, place of birth>, <passport, date of birth>] (3) [<Driver's license, name>, <Driver's license, address>, <Driver's license, date of birth>, <Driver's license, expiry date>, <Driver's license, permitted vehicle types>] (4) [<Social Security Card, Cooperative Bank>, <Social Security Card, Social Security Card Number>, <Social Security Card, Bank Card Number>, <Social Security Card, Issuance Date>, <Social Security Card, Issuing Authority>] (5) [<Residence Permit, Name>, <Residence Permit, Date of Birth>, <Residence Permit, Place of Household Registration>, <Residence Permit, Current Address>, <Residence Permit, Validity Period>] For a personal biometrics category, the unique and distinctive attributes of a personal biometrics category may include a combination of unique attributes and / or original attributes.

[0071] The unique attribute can be represented as <type, content>, and can include, but is not limited to, the following unique attributes: (1) <Fingerprint, texture and inflection point set>: Composed of feature points, a certain fingerprint can be determined through feature point image recognition.

[0072] (2) <Human face, facial contour and inflection point set>: The face can be recognized by feature points.

[0073] (3) <Gene, gene fragment set>: DNA sequence.

[0074] The primitive attributes for forming the primitive attribute combination may include but are not limited to the following content: (1) <Voiceprint, voiceprint feature>: The voice can be recognized by the voiceprint feature.

[0075] (2) <Palmprint, feature point>: The palmprint can be recognized by the feature points.

[0076] (3) <Iris, feature point>: The iris can be recognized through feature extraction and matching.

[0077] (4) <Blood type, blood type value>: The attribute value may be "A", "B", "O", "AB", etc.

[0078] Based on the primitive attributes, the following primitive attribute combinations can be formed, including but not limited to: (1) [<Voiceprint, voiceprint feature>, <Palmprint, feature point>, <Iris, feature point>] (2) [<Palmprint, feature point>, <Iris, feature point>, <Blood type, blood type value>] For the personal network identity category, the unique special attributes of the personal network identity category may include unique attributes and / or primitive attribute combinations.

[0079] The unique attributes can be expressed as <Type, Content>, and specifically may include but are not limited to the following unique attributes: (1) <Email, email address>: Composed of letters and numbers + "@" + "email service address".

[0080] (2) <Personal subject account, number>: Including <WeChat, WeChat number>, <Weibo, Weibo account>, <QQ, QQ account>, <Alipay, Alipay account>, <JD, JD account>, etc.

[0081] The primitive attributes for forming the primitive attribute combination may include but are not limited to the following content: (1) <Email, email address book> (2) <Personal subject account, contact account list> (3) <IPv4, address> (4) <IPv6, address> Based on the primitive attributes, the following primitive attribute combinations can be formed, but are not limited to: (1)[<Email, email address book>, <personal subject account, contact account list>, <IPv4, address>] (2)[<Email, email address book>, <personal subject account, contact account list>, <IPv6, address>] For the institutional digital identity category, the unique special attributes of the institutional digital identity category may include unique attributes and / or primitive attribute combinations.

[0082] Wherein, the unique attribute can be expressed as <type, number>, and the type herein can be the organization attribute type. Specifically, the following unique attributes may be included but are not limited to: (1)<Organization Code Certificate, organization code>: for example, <Organization, number>, which is the unique identification number assigned to the institution upon registration or establishment.

[0083] The primitive attributes used to form the primitive attribute combination may include but are not limited to the following contents: (1)<Organization Code Certificate, institution name> (2)<Organization Code Certificate, institution type> (3)<Organization Code Certificate, institution registration location> (4)<Organization Code Certificate, industry classification> (5)<Organization Code Certificate, institution registration time> (6)<Organization Code Certificate, institution legal person or responsible person> (7)<Organization Code Certificate, institution office address> Based on the primitive attributes, the following primitive attribute combinations can be formed, but are not limited to: (1)[<Organization Code Certificate, institution name>, <Organization Code Certificate, institution registration time>, <Organization Code Certificate, institution legal person or responsible person>] (2)[<Organization Code Certificate, institution name>, <Organization Code Certificate, institution registration location>, <Organization Code Certificate, institution legal person or responsible person>] Based on any of the above embodiments, the constituent information of each unique feature attribute of each category is represented as a first knowledge graph. The first knowledge graph takes categories and attributes as nodes, and the attributes connected to the categories include unique attributes and / or primitive attributes. Any unique feature attribute of the category is the unique attribute, or the primitive attribute combination, and the primitive attribute combination is determined based on multiple primitive attributes connected to the category, or based on multiple primitive attributes connected to the category and the relationship between the multiple primitive attributes connected to the category.

[0084] Specifically, the composition information of each unique feature attribute of each category can be represented in the form of a knowledge graph. Here, the knowledge graph used to identify the composition information of the unique feature attributes of a category is denoted as the first knowledge graph.

[0085] In the first knowledge graph, both categories and attributes are nodes. The edges connecting categories and attributes reflect that the attribute is an attribute under that category, and the edges connecting attributes reflect the relationships between attributes.

[0086] Figure 2 This is one of the schematic diagrams of the first knowledge graph provided by the present invention. Figure 2 The provided first knowledge graph is a general example of first knowledge graphs for all categories, namely... Figure 2 The provided first knowledge graph can be instantiated as a construction example of the first knowledge graph for each category. Figure 2 In this context, a class is an abstraction of a category. A class node can connect to multiple attribute nodes, specifically nodes with N primitive attributes and M unique attributes. It's understood that N and M can both be any integer greater than or equal to 0. The N primitive attributes can form at least one combination of primitive attributes, and each combination of primitive attributes constitutes a unique characteristic attribute. For example, in... Figure 2 In this context, primitive attributes 1 and 2 constitute unique feature attribute 1, primitive attributes 2, 3, and 4 constitute unique feature attribute 2, and primitive attributes 4 and N constitute unique feature attribute 3. Additionally, M unique attributes can be considered as M unique feature attributes, for example, in... Figure 2 In this context, unique attribute 1 is equivalent to unique feature attribute 4, unique attribute 2 is equivalent to unique feature attribute 5, and so on. Unique attribute M is equivalent to unique feature attribute P, where P is the sum of the number of original attribute combinations and unique attributes. Furthermore, Figure 2 Each unique feature attribute in the class is a category identifier for that class; that is, the number of unique feature attributes is the same as the number of category identifiers.

[0087] Figure 3 This is a second schematic diagram of the first knowledge graph provided by the present invention, such as... Figure 3As shown, for the case of the Personal Identity Information class, the Personal Identity Information class can include basic attributes such as employer, employee ID, IP address, name, and address. Among them, the combination of employer and employee ID serves as the unique characteristic attribute 1 of the Personal Identity Information class and can also serve as the identifier of the Personal Digital Identity category; the combination of employee ID, IP address, and name serves as the unique characteristic attribute 2 of the Personal Identity Information class and can also serve as the identifier of the Personal Online Identity category; the combination of name and address serves as the unique characteristic attribute 3 of the Personal Identity Information class and can also serve as the identifier of the Personal Digital Identity category. In addition, the Personal Identity Information class can also include unique attributes such as ID card number, fingerprint, and email address. Each unique attribute serves as the unique characteristic attribute of the Personal Identity Information class, where ID card number can serve as the identifier of the Personal Digital Identity category, fingerprint can serve as the identifier of the Personal Biometric Identity category, and email address can serve as the identifier of the Personal Online Identity category.

[0088] Based on any of the above embodiments, step 120 includes: The attribute values ​​of each attribute in the data block are filled into the first knowledge graph to obtain the second knowledge graph; Based on the attribute nodes filled with attribute values ​​in the second knowledge graph and the node connection relationships in the second knowledge graph, unique feature attributes are generated for the data block.

[0089] Specifically, after obtaining the attribute values ​​of each attribute in the data block, each attribute in the data block can be matched with the attributes that serve as nodes in the first knowledge graph. After a successful match, the attribute values ​​are filled into the corresponding attribute nodes in the first knowledge graph. Here, the first knowledge graph filled with the attribute values ​​of each attribute in the data block is denoted as the second knowledge graph.

[0090] It should be noted that the number of second knowledge graphs obtained based on the data blocks can be one or more. For example, if there are personal information of two people in the data block, the attribute values ​​of each attribute of each person's personal information can be filled into the first knowledge graph, thereby obtaining the second knowledge graph for each person.

[0091] Understandably, the first knowledge graph is used to reflect the composition information of each unique feature attribute under the category, while the second knowledge graph, while reflecting the composition information of each unique feature attribute under the category, also covers each attribute contained in the data block, as well as the attribute value of each attribute contained in the data block.

[0092] After obtaining the second knowledge graph, the attribute nodes filled with attribute values ​​in the second knowledge graph can be aggregated upwards based on the node connection relationship in the second knowledge graph, thereby generating unique feature attributes for the data block.

[0093] In the second knowledge graph, attribute nodes filled with attribute values ​​reflect the attributes contained in the data block. The node connections in the second knowledge graph indicate which primitive attribute combinations can constitute the unique feature attribute of a category, and which unique attributes can be directly used as the unique feature attribute of a category. Based on these two factors, it is possible to determine whether the primitive attributes contained in the data block can be combined to form a unique feature attribute, and whether the data block contains unique attributes that can be directly used as unique feature attributes, thereby realizing the generation of unique feature attributes.

[0094] The method provided in this invention, based on the structural characteristics of knowledge graphs, generates unique feature attributes for data blocks, which can cope with the classification requirements of various complex and variable data blocks, ensuring the reliability of abstract classification of data blocks, thereby guaranteeing the reliability and accuracy of data processing based on abstract classification.

[0095] Based on any of the above embodiments, in step 120, the generation of unique feature attributes for the data block based on the composition information of each unique feature attribute of each category and the attribute values ​​of each attribute in the data block, further includes: Obtain the newly added data block of the data block, perform attribute parsing on the newly added data block, and obtain the attribute values ​​of each attribute in the newly added data block; The attribute values ​​of each attribute in the newly added data block are filled into the second knowledge graph to obtain the third knowledge graph; Based on the attribute nodes filled with attribute values ​​in the third knowledge graph and the node connection relationships in the third knowledge graph, a unique feature attribute is generated for the newly added data block.

[0096] Specifically, after generating unique feature attributes based on the second knowledge graph, for cases where new content has been added to a data block, iteration can be performed on the second knowledge graph to achieve classification of the new content. In this embodiment of the invention, the new content of the data block is denoted as the new data block. It should be noted that the new data block here may be the same as or different from the aforementioned data block in terms of modality, characteristics, data type, size, etc., and this embodiment of the invention does not impose specific limitations on this.

[0097] After obtaining the newly added data block, attribute parsing can be performed on the new data block to obtain the attribute values ​​of each attribute in the new data block. Here, the method for parsing attributes on the new data block can refer to the method for parsing attributes on the data block in step 110, and will not be repeated here.

[0098] After obtaining the attribute values ​​of each attribute in the newly added data block, each attribute in the newly added data block can be matched with the attributes that serve as nodes in the second knowledge graph. Upon successful matching, the attribute values ​​are then populated into the corresponding attribute nodes in the second knowledge graph. Here, the second knowledge graph populated with the attribute values ​​of each attribute from the newly added data block is denoted as the third knowledge graph.

[0099] Understandably, compared to the second knowledge graph, the third knowledge graph not only reflects the composition information of each unique feature attribute under the category and the attributes and attribute values ​​contained in the data block, but also reflects the attributes and attribute values ​​contained in the newly added data block.

[0100] After obtaining the third knowledge graph, the attribute nodes filled with attribute values ​​in the third knowledge graph can be aggregated upwards based on the node connection relationships in the third knowledge graph, thereby generating unique feature attributes for newly added data blocks. It can be understood that the unique feature attributes generated for newly added data blocks here reflect the unique feature attributes corresponding to the data block and the newly added data block as a whole.

[0101] The method provided in this embodiment of the invention forms a third knowledge graph based on a second knowledge graph for newly added data blocks, and generates unique feature attributes for the newly added data blocks based on the third knowledge graph. This realizes iterative operation of data blocks and newly added data blocks in abstract classification, ensuring the comprehensiveness of abstract classification for newly added data blocks and avoiding the omission of information during abstract classification.

[0102] or, After obtaining the third knowledge graph, it is not necessary to aggregate upwards. That is, a document can be composed of information of different modalities and categories. For example, a personal resume can be composed of identity information, personal experience information, etc. The third knowledge graph here can serve as a structured representation of information in newly added data blocks.

[0103] Based on any of the above embodiments Figure 4 This is a flowchart illustrating the method for acquiring the first knowledge graph provided by the present invention, as shown below. Figure 4 As shown, the steps for acquiring the first knowledge graph include: Step 410: Obtain the target category and the attributes under the target category.

[0104] Step 420: Determine the unique attributes and / or intrinsic attributes of the target category from the attributes.

[0105] Here, the target category refers to the category corresponding to the first knowledge graph to be generated. The target category can be one category or multiple categories. After determining the target category, the attributes under that target category can be identified. These attributes can include unique attributes that can stand alone as identifiers of the target category, as well as primitive attributes that cannot stand alone as identifiers of the target type. Here, for the target category, the vertical columns under that target category can be obtained based on the category name of the target category, and at least one of the following: relevant laws, regulations, standards, systems, and various knowledge bases.

[0106] For example, N primitive attributes (N≥1) and / or M unique attributes (M≥1) can be determined from the attributes. Here, the values ​​of N and M can be the same or different for different target categories.

[0107] Step 430: Combine the original attributes to obtain the original attribute combination that serves as the unique feature attribute of the target category.

[0108] Step 440: Construct the first knowledge graph based on the unique attributes of the target category and / or the combination of the original attributes.

[0109] Specifically, for each target category, x sets of original attribute combinations are found from N original attributes as unique feature attributes, and M unique attributes are respectively used as M unique feature attributes. Thus, based on the unique feature attributes of each target category, the first knowledge graph can be constructed.

[0110] In this process, a first knowledge graph can be constructed separately for each target category. Alternatively, a knowledge graph can be constructed separately for each target category first, and then the common primitive attributes in each knowledge graph can be integrated to obtain a first knowledge graph that includes multiple target categories. This embodiment of the invention does not specifically limit this approach.

[0111] In practical applications, data processing solutions based on data abstraction and classification typically require classifying data blocks and then grading them according to their category. The data is then processed and used correctly based on this grade. For example, regarding data related to personal information, laws and regulations such as the Personal Information Protection Law, the Data Security Law, and the GDPR (General Data Protection Regulation) regulate various processing and uses of personal information. Correct use and processing depend on classifying and grading data blocks composed of multiple attributes. Only by grading these data blocks can they be processed and used correctly without causing information leakage.

[0112] In this process, the classification and grading of data blocks remains within a static and limited scope. That is, the classification and grading of a data block does not change dynamically; once a level is determined, it will not be changed unless there are updates to laws, regulations, or standards. For example, classifying and grading personal financial data based solely on a single content, field, or related fields and content is a static approach that cannot meet the needs of classifying and grading the same data element at different times and in different usage scenarios. This may lead to information leakage problems at inappropriate times and in inappropriate scenarios.

[0113] Based on any of the above embodiments, step 140 includes: Based on the category to which the data block belongs and the data processing environment corresponding to the data block, the level of the category to which the data block belongs is determined; The data block is processed based on the category to which it belongs, and / or based on the level of the category to which it belongs; The data processing environment includes at least one of the following: the scenario, time, location, business system, application system, and industry in which the data block is processed.

[0114] Specifically, after determining the category to which a data block belongs, it is necessary to determine the level of the data block's category based on the category and the corresponding data processing environment. That is, the data block is classified and graded based on the data processing environment. On this basis, the data block is then processed according to its category and / or the level of its category.

[0115] The data processing environment includes at least one of the following: the scenario, time, location, business system, application system, and industry in which the data block is processed. Here, the business system can be various types of business processing systems, such as ride-hailing platforms, shopping platforms, social platforms, and chat platforms; the application system can be a front-end application system or a back-end application system.

[0116] It is understandable that data blocks of the same category may correspond to the same or different levels in different data processing environments. The level here can reflect the security level or confidentiality level of the data block, etc. The data processing methods used for data blocks of different levels can also be different. For example, for data blocks at the level of not requiring confidentiality, the processing method can be direct transmission or public access; for data blocks at the level of being confidential to the public, the processing method can be anonymized transmission or encrypted access. This embodiment of the invention does not specifically limit this.

[0117] The method provided in this invention determines the level of a data block based on the data processing environment in which the data block is located, and processes the data block based on the level, so that the processing of the data block can adapt to different data processing environments, ensuring the compliance and reliability of data processing.

[0118] Based on any of the above embodiments, in step 140, processing the data block based on the category to which the data block belongs, and / or based on the level of the category to which the data block belongs, includes: The data block is processed based on the data processing environment corresponding to the data block, as well as the category to which the data block belongs and / or the level of the category to which the data block belongs.

[0119] Specifically, when processing data blocks based on their category and / or level, the data processing environment in which the data block resides can also be considered. Therefore, when processing data blocks, a data processing method corresponding to the data processing environment and the category and / or level to which the data block belongs can be selected to achieve the desired data processing.

[0120] By combining the data processing environment in which the data block resides with the category and / or level to which the data block belongs, it is possible to accurately select the data processing method. This ensures that the data processing method performed on the data block is not only consistent with the content contained in the data block itself, but also with the actual data processing scenario, thereby further guaranteeing the compliance and reliability of data processing.

[0121] For any of the above embodiments, step 140, processing the data block based on the category to which the data block belongs, and / or based on the level of the category to which the data block belongs, includes: Based on the category to which the data block belongs and / or the level of the category to which the data block belongs, or based on the data processing environment corresponding to the data block, and the category to which the data block belongs and / or the level of the category to which the data block belongs, a risk prediction is made on the data block to obtain the risk prediction result of the data block; Based on the risk assessment results of the data block, the data block is processed.

[0122] Specifically, after obtaining the category and / or level to which the data block belongs, or after obtaining the data processing environment in which the data block is located, as well as the category and / or level to which the data block belongs, risk assessment can be performed on the data block based on the above information.

[0123] The risk assessment here is used to estimate the potential information leakage risk during data processing of the data block. The risk assessment result can be either whether there is a risk of information leakage, or the probability that such a risk exists.

[0124] After obtaining the risk assessment results, appropriate data processing methods can be selected based on these results, thereby avoiding information leakage during data processing and ensuring information security.

[0125] Based on any of the above embodiments, the data abstraction and classification processing method may include the following steps: First, for each category, determine the attributes of each category, and from the attributes of each category, identify unique attributes and / or combinations of original attributes, and use the unique attributes and / or combinations of original attributes as the unique characteristic attributes of each category.

[0126] Secondly, a first knowledge graph can be built for each category based on its unique attributes and / or combinations of original attributes.

[0127] In actual data abstraction and classification processing, we can obtain the data blocks to be processed and perform attribute parsing on the data blocks to obtain the attribute values ​​of each attribute in the data blocks.

[0128] Next, the attributes in the data block can be matched with the attributes that serve as nodes in the first knowledge graph. After a successful match, the attribute values ​​are filled into the nodes of the corresponding attributes in the first knowledge graph, thus obtaining the second knowledge graph.

[0129] After obtaining the second knowledge graph, the attribute nodes filled with attribute values ​​in the second knowledge graph can be aggregated upwards to generate unique feature attributes for data blocks, and the category to which the data block belongs can be determined based on the generated structure.

[0130] After obtaining the category to which the data block belongs, the level of the data block category is determined based on the category and the corresponding data processing environment. Then, based on the category and / or level of the data block, the specific processing method is determined to realize the data processing for the data block.

[0131] The method provided in this invention identifies each category with a unique characteristic attribute and sets the composition information of unique characteristic attributes based on unique attributes or combinations of original attributes. This allows the generation of unique characteristic attributes for a data block based on the attribute values ​​of each attribute in the data block and the unique characteristic attributes of each category, thereby determining the category to which the data block belongs. In this process, the attribute analysis and unique characteristic attribute generation based on atomic granularity can cope with the classification requirements of various complex and variable data blocks, ensuring the reliability of abstract classification of data blocks, thereby guaranteeing the reliability and accuracy of data processing based on abstract classification.

[0132] Furthermore, the level of a data block is determined by considering its data processing environment, and the data block is processed based on that level. This allows the processing of data blocks to adapt to different data processing environments, ensuring the compliance and reliability of data processing.

[0133] Based on any of the above embodiments Figure 5 This is a schematic diagram of the data abstraction and classification processing device provided by the present invention, as shown below. Figure 5 As shown, the device includes: The data acquisition unit 510 is used to acquire a data block, perform attribute parsing on the data block, and obtain the attribute values ​​of each attribute in the data block. The unique feature attribute generation unit 520 is used to generate unique feature attributes for the data block based on the composition information of each unique feature attribute of each category and the attribute value of each attribute in the data block; the composition information is based on unique attributes or based on a combination of original attributes, the combination of original attributes includes multiple original attributes, or includes multiple original attributes and the relationship between the multiple original attributes. Category labeling unit 530 is used to determine the category to which the data block belongs based on the generation result generated by the unique feature attribute; The data processing unit 540 is used to process the data block based on the category to which the data block belongs.

[0134] The apparatus provided in this invention identifies each category with a unique characteristic attribute and sets the composition information of unique characteristic attributes based on unique attributes or combinations of original attributes. This allows for the generation of unique characteristic attributes for each data block based on the attribute values ​​of each attribute in the data block and the unique characteristic attributes of each category, thereby determining the category to which the data block belongs. In this process, the atomic-level attribute analysis and unique characteristic attribute generation can handle the classification requirements of various complex and variable data blocks, ensuring the reliability of abstract classification of data blocks, and thus guaranteeing the reliability and accuracy of data processing based on abstract classification.

[0135] Based on any of the above embodiments, the constitutive information of each unique feature attribute of each category is represented as a first knowledge graph. The first knowledge graph uses categories and attributes as nodes. The attributes connected to the category include unique attributes and / or primitive attributes. Any unique feature attribute of the category is the unique attribute, or is a combination of primitive attributes. The combination of primitive attributes is determined based on multiple primitive attributes connected to the category, or based on the relationship between multiple primitive attributes connected to the category and multiple primitive attributes connected to the category.

[0136] Based on any of the above embodiments, the unique feature attribute generation unit is used for: The attribute values ​​of each attribute in the data block are filled into the first knowledge graph to obtain the second knowledge graph; Based on the attribute nodes filled with attribute values ​​in the second knowledge graph and the node connection relationships in the second knowledge graph, unique feature attributes are generated for the data block.

[0137] Based on any of the above embodiments, the unique feature attribute generation unit is further configured to: Obtain the newly added data block of the data block, perform attribute parsing on the newly added data block, and obtain the attribute values ​​of each attribute in the newly added data block; The attribute values ​​of each attribute in the newly added data block are filled into the second knowledge graph to obtain the third knowledge graph; Based on the attribute nodes filled with attribute values ​​in the third knowledge graph and the node connection relationships in the third knowledge graph, a unique feature attribute is generated for the newly added data block.

[0138] Based on any of the above embodiments, the unique feature attribute generation unit is further configured to: Obtain the target category and the attributes under the target category; From the attributes, determine the unique attributes and / or original attributes of the target category; The original attributes are combined to obtain the original attribute combination, which serves as the unique feature attribute of the target category; The first knowledge graph is constructed based on the unique attributes of the target category and / or the combination of the original attributes.

[0139] Based on any of the above embodiments, the data processing unit is used for: Based on the category to which the data block belongs and the data processing environment corresponding to the data block, the level of the category to which the data block belongs is determined; The data block is processed based on the category to which it belongs, and / or based on the level of the category to which it belongs; The data processing environment includes at least one of the following: the scenario, time, location, business system, application system, and industry in which the data block is processed.

[0140] Based on any of the above embodiments, the data processing unit is used for: The data block is processed based on the data processing environment corresponding to the data block, as well as the category to which the data block belongs and / or the level of the category to which the data block belongs.

[0141] Based on any of the above embodiments, the data processing unit is used for: Based on the category to which the data block belongs and / or the level of the category to which the data block belongs, or based on the data processing environment corresponding to the data block, and the category to which the data block belongs and / or the level of the category to which the data block belongs, a risk prediction is made on the data block to obtain the risk prediction result of the data block; Based on the risk assessment results of the data block, the data block is processed.

[0142] Figure 6 An example is a schematic diagram of the physical structure of an electronic device, such as... Figure 6 As shown, the electronic device may include: a processor 610, a communications interface 620, a memory 630, and a communication bus 640, wherein the processor 610, the communications interface 620, and the memory 630 communicate with each other via the communication bus 640. The processor 610 can call logical instructions in the memory 630 to execute a data abstraction and classification processing method, which includes: Obtain a data block, perform attribute parsing on the data block, and obtain the attribute values ​​of each attribute in the data block; Based on the composition information of each unique feature attribute of each category, and the attribute value of each attribute in the data block, a unique feature attribute is generated for the data block; the composition information is based on unique attributes or based on a combination of original attributes, the combination of original attributes includes multiple original attributes, or includes multiple original attributes and the relationship between the multiple original attributes. Based on the generation result generated by the unique feature attribute, the category to which the data block belongs is determined; The data block is processed based on the category to which it belongs.

[0143] Furthermore, the logical instructions in the aforementioned memory 630 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0144] On the other hand, the present invention also provides a computer program product, the computer program product comprising a computer program that can be stored on a non-transitory computer-readable storage medium, wherein when the computer program is executed by a processor, the computer is able to execute the data abstraction and classification processing method provided by the above methods, the method comprising: Obtain a data block, perform attribute parsing on the data block, and obtain the attribute values ​​of each attribute in the data block; Based on the composition information of each unique feature attribute of each category, and the attribute value of each attribute in the data block, a unique feature attribute is generated for the data block; the composition information is based on unique attributes or based on a combination of original attributes, the combination of original attributes includes multiple original attributes, or includes multiple original attributes and the relationship between the multiple original attributes. Based on the generation result generated by the unique feature attribute, the category to which the data block belongs is determined; The data block is processed based on the category to which it belongs.

[0145] In another aspect, the present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the data abstraction and classification processing method provided by the above methods, the method comprising: Obtain a data block, perform attribute parsing on the data block, and obtain the attribute values ​​of each attribute in the data block; Based on the composition information of each unique feature attribute of each category, and the attribute value of each attribute in the data block, a unique feature attribute is generated for the data block; the composition information is based on unique attributes or based on a combination of original attributes, the combination of original attributes includes multiple original attributes, or includes multiple original attributes and the relationship between the multiple original attributes. Based on the generation result generated by the unique feature attribute, the category to which the data block belongs is determined; The data block is processed based on the category to which it belongs.

[0146] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0147] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0148] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A data abstraction and classification processing method, characterized in that, include: Obtain a data block, parse the attributes of the data block to obtain the attribute values ​​of each attribute in the data block, and the data block includes any one of xml, html, txt text, layout document, streaming document, image, audio, and video; Based on the composition information of each unique feature attribute of each category, and the attribute value of each attribute in the data block, a unique feature attribute is generated for the data block; the composition information is based on unique attributes or based on a combination of original attributes, the combination of original attributes includes multiple original attributes, or includes multiple original attributes and the relationship between the multiple original attributes. Based on the results generated from the unique feature attributes, the category to which the data block belongs is determined; The data block is processed based on its category. The composition information of each unique feature attribute of each category is represented as a first knowledge graph. The first knowledge graph uses categories and attributes as nodes. The attributes connected to the category include unique attributes and / or primitive attributes. Any unique feature attribute of the category is the unique attribute, or is a combination of primitive attributes. The combination of primitive attributes is determined based on multiple primitive attributes connected to the category, or based on multiple primitive attributes connected to the category and the relationship between multiple primitive attributes connected to the category. The process of generating unique feature attributes for the data block based on the composition information of each unique feature attribute of each category and the attribute values ​​of each attribute in the data block includes: The attribute values ​​of each attribute in the data block are filled into the first knowledge graph to obtain the second knowledge graph; Based on the attribute nodes filled with attribute values ​​in the second knowledge graph and the node connection relationships in the second knowledge graph, a unique feature attribute is generated for the data block; The processing of the data block based on its category includes: Based on the category to which the data block belongs and the data processing environment corresponding to the data block, determine the level of the category to which the data block belongs in the data processing environment; The data block is processed based on the category to which it belongs, and / or based on the level of the category to which it belongs in the data processing environment; The data processing environment includes at least one of the following: the scenario, time, location, business system, application system, and industry in which the data block is processed.

2. The data abstraction and classification processing method according to claim 1, characterized in that, Based on the composition information of each unique feature attribute of each category and the attribute values ​​of each attribute in the data block, the unique feature attributes of the data block are generated, and then the process further includes: Obtain the newly added data block of the data block, perform attribute parsing on the newly added data block, and obtain the attribute values ​​of each attribute in the newly added data block; The attribute values ​​of each attribute in the newly added data block are filled into the second knowledge graph to obtain the third knowledge graph; Based on the attribute nodes filled with attribute values ​​in the third knowledge graph and the node connection relationships in the third knowledge graph, a unique feature attribute is generated for the newly added data block.

3. The data abstraction and classification processing method according to claim 1, characterized in that, The steps for acquiring the first knowledge graph include: Obtain the target category and the attributes under the target category; From the attributes, determine the unique attributes and / or original attributes of the target category; The original attributes are combined to obtain the original attribute combination, which serves as the unique feature attribute of the target category; The first knowledge graph is constructed based on the unique attributes of the target category and / or the combination of the original attributes.

4. The data abstraction and classification processing method according to claim 1, characterized in that, The processing of the data block based on its category and / or the level of its category in the data processing environment includes: The data block is processed based on the data processing environment corresponding to the data block, the category to which the data block belongs, and / or the level of the category to which the data block belongs in the data processing environment.

5. The data abstraction and classification processing method according to claim 1, characterized in that, The processing of the data block based on its category and / or the level of its category in the data processing environment includes: Based on the category to which the data block belongs and / or the level of the category to which the data block belongs in the data processing environment, or based on the data processing environment corresponding to the data block, and the category to which the data block belongs and / or the level of the category to which the data block belongs in the data processing environment, a risk prediction is made on the data block to obtain the risk prediction result of the data block. Based on the risk assessment results of the data block, the data block is processed.

6. A data abstraction and classification processing device, characterized in that, include: A data acquisition unit is used to acquire data blocks, perform attribute parsing on the data blocks, and obtain the attribute values ​​of each attribute in the data blocks. The data blocks include any one of the following: XML, HTML, TXT text, formatted document, streaming document, image, audio, and video. A unique feature attribute generation unit is used to generate unique feature attributes for the data block based on the composition information of each unique feature attribute of each category and the attribute value of each attribute in the data block; the composition information is based on unique attributes or based on a combination of original attributes, the combination of original attributes includes multiple original attributes, or includes multiple original attributes and the relationship between the multiple original attributes. A category labeling unit is used to determine the category to which the data block belongs based on the result generated by the unique feature attribute; A data processing unit is used to process the data block based on the category to which the data block belongs; The composition information of each unique feature attribute of each category is represented as a first knowledge graph. The first knowledge graph uses categories and attributes as nodes. The attributes connected to the category include unique attributes and / or primitive attributes. Any unique feature attribute of the category is the unique attribute, or is a combination of primitive attributes. The combination of primitive attributes is determined based on multiple primitive attributes connected to the category, or based on multiple primitive attributes connected to the category and the relationship between multiple primitive attributes connected to the category. The unique feature attribute generation unit is specifically used to: fill the attribute values ​​of each attribute in the data block into the first knowledge graph to obtain the second knowledge graph; and generate unique feature attributes for the data block based on the attribute nodes filled with attribute values ​​in the second knowledge graph and the node connection relationships in the second knowledge graph. The data processing unit is specifically used to: determine the level of the data block's category in the data processing environment based on the category to which the data block belongs and the data processing environment corresponding to the data block; The data block is processed based on the category to which it belongs, and / or based on the level of the category to which it belongs in the data processing environment; The data processing environment includes at least one of the following: the scenario, time, location, business system, application system, and industry in which the data block is processed.

7. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the data abstraction and classification processing method as described in any one of claims 1 to 5.

8. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the data abstraction and classification processing method as described in any one of claims 1 to 5.

Images