Structured data automatic classification and grading method, device, equipment, medium and product

Abstract

The invention discloses a structured data automatic classification and grading method, device and equipment, a medium and a product, and relates to the technical field of data security and information processing. The method comprises the steps that after meta-information of to-be-processed structured data is input, the meta-information is subjected to analysis and semantic processing, a semantic information structure is generated, and the semantic information structure is transmitted into a core inference engine which is built based on a large language model and is integrated with a classification and grading strategy knowledge base; and then the engine performs one-time logical reasoning on a semantic information structure according to a knowledge base, directly generates a decision result containing the category and the security level of each field, and finally outputs the result, so that an end-to-end automatic process taking a large language model as an intelligent core is constructed, and the deep semantic understanding and reasoning capability of the end-to-end automatic process is utilized to realize the intelligent decision-making of the large language model. The traditional rule matching and simple machine learning model is replaced, the accuracy, efficiency and interpretability of data classification and grading are remarkably improved, and automatic processing close to the human expert level is achieved.

Description

Technical Field

[0001] This invention belongs to the field of data security and information processing technology, specifically relating to a method, apparatus, equipment, medium and product for automatic classification and grading of structured data. Background Technology

[0002] Data classification and grading are the cornerstones of data security governance and compliance management. Their goal is to categorize data and determine corresponding security protection levels based on its sensitivity, importance, and business attributes. Accurate and efficient data classification and grading are prerequisites for achieving differentiated data security control and meeting relevant legal and regulatory requirements.

[0003] Currently, the industry mainly relies on the following technical means to achieve data classification and grading: (A) Data classification and grading based on rule engines, which involves pre-setting a series of keywords, regular expressions or logical rules, and then classifying and grading database fields by matching data field names and comments with the preset rules. Although simple to implement, this method is essentially a rigid string or pattern matching method, which cannot understand the specific business context and deep semantics of the data. For scenarios with ambiguous, abbreviated or complex field names, it is easy to produce misjudgments and omissions, resulting in unsatisfactory accuracy and recall of the classification results. (B) Data classification and grading based on traditional machine learning, which uses classic machine learning models such as decision trees or support vector machines, requires first manually constructing and labeling a large amount of training data to extract features, and then training the model to classify fields. This approach can reduce the workload of manually writing rules to some extent, but its effectiveness depends heavily on the quality of feature engineering and the scale and representativeness of the training data. This makes the model's generalization ability limited when facing rapid business changes and diverse data patterns, making it difficult to adapt to new and unseen data patterns, and resulting in high maintenance and update costs. (C) Data classification and grading based on manual labeling relies entirely on the personal experience of data administrators or security experts. Experts manually label database tables and fields according to the organization's classification and grading strategy. Although this method can combine the domain knowledge of human experts, it is extremely inefficient and cannot meet the processing needs of massive amounts of data within an enterprise. At the same time, human judgment is highly subjective, and it is difficult to unify standards among different experts. When encountering cases with ambiguous boundaries, repeated discussions and confirmations are still required. The degree of automation is low, the labor cost is high, and it is difficult to ensure the consistency of large-scale data governance projects.

[0004] In summary, existing data classification and grading technologies have the following main shortcomings: (1) Low recognition accuracy and recall: rule-based methods cannot understand semantics, and machine learning methods have weak generalization ability. When faced with complex and ever-changing real enterprise data, both are difficult to achieve high-precision recognition, and mislabeling and missed labeling are serious. (2) Lack of reasoning and explanation ability: Existing methods are mostly a "black box" or a simple matching process, which cannot perform logical reasoning on the decision-making process of classification and grading like human experts, nor can they provide a clear and credible explanation for the judgment results. This is a major shortcoming in the field of data security that emphasizes compliance auditing and accountability. (3) Limited automation and intelligence: The above methods either rely on a large number of manual rule maintenance, or on heavy feature engineering and model training, or are completely completed by human labor. None of them have achieved true end-to-end intelligent processing. Their automation process is fragile and difficult to handle massive data and long-tail cases.

[0005] Therefore, there is an urgent need in this field for a data classification and grading technology that can deeply understand data semantics, possess logical reasoning capabilities, and achieve a high degree of automation, in order to overcome the inherent defects of existing technologies in terms of accuracy, interpretability, and efficiency. Summary of the Invention

[0006] The purpose of this invention is to provide a method, apparatus, computer device, computer-readable storage medium, and computer program product for automatic classification and grading of structured data, in order to solve the problems of low recognition accuracy and recall, lack of reasoning and interpretation capabilities, and / or limited automation and intelligence in existing data classification and grading technologies.

[0007] To achieve the above objectives, the present invention adopts the following technical solution: Firstly, a method for automatically classifying and grading structured data is provided, including: Input the metadata of the structured data to be processed; The metadata is parsed and semantically processed to generate a standardized semantic information structure; The semantic information structure is fed into the core reasoning engine, which is built on a large language model and integrates a classification and grading strategy knowledge base. The core reasoning engine performs a one-time logical reasoning on the semantic information structure based on the classification and grading strategy knowledge base, and directly generates a classification and grading decision result containing the category and security level of each field data in the structured data to be processed. Output the classification and grading decision results.

[0008] Based on the above-mentioned invention, a novel end-to-end automated classification and grading scheme for structured data is provided, based on a large language model and a classification and grading strategy knowledge base. Specifically, after inputting the metadata of the structured data to be processed, the metadata is first parsed and semantically processed to generate a semantic information structure. This structure is then fed into a core reasoning engine built on a large language model and integrated with a classification and grading strategy knowledge base. The engine then performs a one-time logical reasoning on the semantic information structure based on the knowledge base, directly generating a decision result containing the category and security level of each field. Finally, this result is output. By constructing an end-to-end (from input to output) automated process with a large language model as the intelligent core, and utilizing its deep semantic understanding and reasoning capabilities, this scheme replaces traditional rule matching and simple machine learning models, significantly improving the accuracy, efficiency, and interpretability of data classification and grading. It achieves near-human expert-level automated processing, facilitating practical application and promotion.

[0009] In one possible design, the metadata includes database structure information or metadata information.

[0010] In one possible design, the metadata is parsed and semantically processed to generate a standardized semantic information structure, including: The metadata is parsed to obtain a parsing result containing the identified table headers and field meanings; The semantics of the parsing results are transformed into an information structure that can be understood by a core reasoning engine that is built on a large language model and integrates a classification and grading strategy knowledge base, thus serving as a standardized semantic information structure.

[0011] In one possible design, the large language model is replaced by a classification and grading-specific model, which is a specialized model that is pre-trained and fine-tuned for the task of classifying and grading structured data and has fewer parameters than the general large language model.

[0012] In one possible design, the classification and grading decision results are output, including: Output the classification and grading decision results in the form of a structured list and / or a visual chart.

[0013] In one possible design, after outputting the classification and grading decision results, the method includes: For each field of the structured data to be processed, the corresponding security level is extracted from the classification and grading decision results, and a data protection strategy is triggered to protect the corresponding data based on the security level. The data protection strategy includes data encryption and / or data desensitization methods.

[0014] Secondly, an automatic classification and grading device for structured data is provided, comprising an information input unit, an information parsing unit, a logical reasoning unit, and a result output unit that are sequentially connected in communication. The information input unit is used to input the metadata of the structured data to be processed; The information parsing unit is used to parse and semanticize the metadata to generate a standardized semantic information structure. The logical reasoning unit is used to input the semantic information structure into the core reasoning engine, which is built based on a large language model and integrates a classification and grading strategy knowledge base. The core reasoning engine performs a one-time logical reasoning on the semantic information structure according to the classification and grading strategy knowledge base, and directly generates a classification and grading decision result containing the category and security level of each field data in the structured data to be processed. The result output unit is used to output the classification and grading decision results.

[0015] Thirdly, the present invention provides a computer device comprising a storage module, a processing module, and a transceiver module connected in sequence for communication, wherein the storage module is used to store a computer program, the transceiver module is used to send and receive messages, and the processing module is used to read the computer program and execute the structured data automatic classification and grading method as described in the first aspect or any possible design in the first aspect.

[0016] Fourthly, the present invention provides a computer-readable storage medium storing instructions that, when executed on a computer, perform the automatic classification and grading method for structured data as described in the first aspect or any possible design within the first aspect.

[0017] Fifthly, the present invention provides a computer program product, including a computer program or instructions, which, when executed by a computer, implement the automatic classification and grading method for structured data as described in the first aspect or any possible design in the first aspect.

[0018] The beneficial effects of the above scheme are: (1) This invention creatively provides a new scheme for end-to-end automated classification and grading of structured data based on a large language model and a classification and grading strategy knowledge base. That is, after inputting the meta-information of the structured data to be processed, the meta-information is first parsed and semantically processed to generate a semantic information structure, which is then fed into the core reasoning engine built on the large language model and integrated with the classification and grading strategy knowledge base. Then, the engine performs a one-time logical reasoning on the semantic information structure according to the knowledge base, directly generating a decision result containing the category and security level of each field, and finally outputting the result. Thus, by constructing an end-to-end (i.e. from input to output) automated process with the large language model as the intelligent core, and by utilizing its deep semantic understanding and reasoning capabilities, it replaces the traditional rule matching and simple machine learning models, significantly improving the accuracy, efficiency and interpretability of data classification and grading, and achieving automated processing close to the level of human experts. (2) It can achieve a high level of end-to-end automation. That is, by building a core reasoning engine with a large language model that integrates a strategy knowledge base as the "intelligent brain", and integrating data parsing, semantic processing, model reasoning and result output into a coherent process, it realizes full automation from data input to classification and grading result generation. This greatly reduces the reliance on the human experience of data governance engineers, can efficiently and consistently process massive structured data, significantly improve data governance efficiency and reduce operating costs. (3) It can significantly improve the accuracy and reliability of classification and grading. The core innovation lies in using the powerful deep semantic understanding and logical reasoning capabilities of the large language model to replace the traditional rule matching or simple feature classification. This enables the comprehensive analysis of contextual information such as table structure, field description and sample data, accurately capturing the correlation between fields and complex business patterns, thereby accurately judging data with ambiguous field names and diverse expressions. This fundamentally solves the problem of misjudgment and omission caused by the rigidity and insufficient generalization ability of traditional methods, and achieves recognition accuracy and high recall rate close to the level of human experts. (4) It can enhance the interpretability and compliance support of decision-making. That is, the “one-time logical reasoning” process based on the big language model can simulate the decision chain of human experts. The results generated naturally have traceable reasoning logic, which provides a convincing explanation for classification and hierarchical decision-making, greatly facilitates auditing, compliance inspection and accountability, and meets the strict requirements of transparency and auditability in the field of data security. (5) It can promote the automated execution of data security strategies. That is, the generated decision results containing clear security levels can be directly used as the triggering basis for downstream data security control strategies (such as encryption and desensitization), realizing a closed loop from intelligent identification to automatic protection, improving the response speed and accuracy of the overall data security system, and facilitating practical application and promotion. Attached Figure Description

[0019] To more clearly illustrate the technical solutions in the embodiments of the present 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 only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0020] Figure 1 This is a flowchart illustrating the automatic classification and grading method for structured data provided in an embodiment of this application.

[0021] Figure 2 This is a schematic diagram of the structure of the automatic classification and grading device for structured data provided in the embodiments of this application.

[0022] Figure 3 A schematic diagram of the structure of a computer device provided in an embodiment of this application. Detailed Implementation

[0023] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the present invention will be briefly introduced below in conjunction with the accompanying drawings and descriptions of the embodiments or the prior art. Obviously, the following description of the structure of the accompanying drawings is only some embodiments of the present invention. For those skilled in the art, other embodiments can be obtained based on these embodiments without creative effort. It should be noted that the description of these embodiments is for the purpose of helping to understand the present invention, but does not constitute a limitation of the present invention.

[0024] It should be understood that although the terms "first" and "second", etc., may be used herein to describe various objects, these objects should not be limited by these terms. These terms are only used to distinguish one object from another. For example, the first object may be referred to as the second object, and similarly, the second object may be referred to as the first object, without departing from the scope of the exemplary embodiments of the invention.

[0025] It should be understood that the term "and / or" that may appear in this document is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can mean: A exists alone, B exists alone, or A and B exist simultaneously. Another example is A, B and / or C, which can mean that any one of A, B, and C or any combination thereof exists. The term " / and" that may appear in this document describes another relationship between related objects, indicating that two relationships can exist. For example, A / and B can mean: A exists alone or A and B exist simultaneously. In addition, the character " / " that may appear in this document generally indicates that the related objects before and after it are in an "or" relationship.

[0026] Example like Figure 1 As shown, the structured data automatic classification and grading method provided in the first aspect of this embodiment can be executed, but is not limited to, by computer devices with certain computing resources, such as servers, personal computers (PCs, referring to a type of multi-purpose computer of a size, price, and performance suitable for personal use; desktops, laptops, mini-laptops, tablets, and ultrabooks all belong to personal computers), smartphones, personal digital assistants (PDAs), or wearable devices. Figure 1 As shown, the automatic classification and grading method for structured data includes, but is not limited to, the following steps S1 to S4.

[0027] S1. Input the metadata of the structured data to be processed.

[0028] In step S1, the structured data to be processed is the object to be automatically classified and graded. Subsequent steps S2-S3 are needed to obtain the category and security level of each field. Specifically, the metadata includes, but is not limited to, database structure information or metadata information. Since the database structure is the framework for organizing related data in a computer storage device, its core components include the database, data tables, fields, and indexes. Therefore, the database structure information contains specific information about these core components. The metadata information specifically includes, but is not limited to, at least one of the following: data table structure information, table name, field name, field description, and sample data. Furthermore, the metadata can be specifically carried in the data processing request and obtained through regular input at the data input terminal.

[0029] S2. The metadata is parsed and semantically processed to generate a standardized semantic information structure.

[0030] In step S2, the standardized semantic information structure is enriched with the semantics of the meta-information to facilitate understanding by the subsequent core inference engine; therefore, this step is equivalent to a "sensory" stage. Specifically, the meta-information is parsed and semantically processed to generate a standardized semantic information structure, including but not limited to: parsing the meta-information to obtain a parsing result containing the identified headers and field meanings; and then semantically transforming the parsing result into an information structure that can be understood by the core inference engine, which is built based on a large language model and integrates a classification and grading strategy knowledge base, as the standardized semantic information structure. The aforementioned parsing process specifically involves identifying and extracting the headers, fields, and their descriptive content from the meta-information, while the semantic transformation involves organizing the extracted information into a standardized data structure that includes field semantics, data types, and data sample context.

[0031] S3. The semantic information structure is fed into the core reasoning engine, which is built based on a large language model and integrates a classification and grading strategy knowledge base. The core reasoning engine performs a one-time logical reasoning on the semantic information structure according to the classification and grading strategy knowledge base, and directly generates a classification and grading decision result containing the category and security level of each field data in the structured data to be processed.

[0032] In step S3, the large language model plays the role of a "bionic brain" in this embodiment. Its core function is to transform general cognitive abilities into specialized reasoning abilities to solve professional problems. Specifically, it is manifested in: (11) deep semantic understanding and contextual association, that is, it can go beyond keyword matching, deeply understand the natural language descriptions in field names, annotations and sample data values, and accurately capture the true business meaning of fields by combining contextual information such as operating system names and table names; (12) complex logical reasoning and decision generation, that is, it can simulate expert thinking, perform step-by-step reasoning on ambiguous, ambiguous or boundary cases (for example, determine whether the "Cust_ID" field is a personal identifier or an institutional identifier), and output a structured decision result containing classification path, security level and reasoning basis at one time; (13) powerful generalization and association ability, that is, it can handle unseen data expressions without pre-training for each specific field name or business term, and adapt to the rapid changes in business and data. The classification and grading strategy knowledge base plays the role of "domain expert manual" or "compliance code" in this step. Its core function is to ensure the professionalism, accuracy and compliance of the reasoning process. Specifically, it is reflected in: (21) defining a professional classification system framework, that is, storing the complete classification and grading directory under the enterprise or industry standard in a structured form (such as a tree structure). For example, it includes multi-level subcategories such as first-level subcategories, second-level subcategories, third-level subcategories and fourth-level subcategories, and security levels of 1 to 4. Examples of first-level subcategories include "customer", "business", "operation management" and "supervision", examples of second-level subcategories include "individual" and "unit" under the first-level subcategory of "customer", and examples of third-level subcategories include "personal natural information", "personal identity information", "personal information information" and "personal relationship information" under the second-level subcategory of "person". "Personal behavior information" and "personal device information", etc. Examples of the four-level subcategories include "personal basic information", "personal identity information", "personal property information", "personal contact information", "personal health and physiological information" and "personal internet access record information" under the three-level subcategories of "personal natural information", providing precise "target options" for decision-making; (22) solidify domain rules and judgment logic, that is, embedding judgment rules in the form of business expert experience and compliance requirements (for example, marking "fields containing personal direct identifiers and used for payment business as 'sensitive level 4'"), providing logical basis for reasoning; (23) constrain and guide the output of the large language model, that is, strictly guide and constrain the open domain knowledge output of the large language model within the professional classification and grading framework, ensuring that the decisions generated conform to the specific strategies of the enterprise, rather than unrestrained general answers.Based on the aforementioned large language model and classification and grading strategy knowledge base, the core reasoning engine can achieve the following one-time logical reasoning: After the semantic information structure is input, the large language model uses its understanding and reasoning capabilities to analyze the input information, and queries and applies the frameworks and rules in the classification and grading strategy knowledge base in real time as a guide for reasoning and a benchmark for judgment. Finally, under the constraints and guidance of the knowledge base, the large language model performs one-time and interpretable logical reasoning to directly generate a classification and grading decision that conforms to professional standards and is structured (i.e., includes, but is not limited to, the category and security level of each field data in the structured data to be processed).

[0033] In step S3, it is also considered that the execution entity of this method may not have strong computing power. Therefore, the core inference engine may not use a general large language model, but instead use a classification and grading-specific model. The classification and grading-specific model refers to a special model that is pre-trained and fine-tuned for the classification and grading task of structured data and has fewer parameters than a general large language model. Since the structure of this special model can be more streamlined, it is possible to sacrifice some generalization ability in exchange for faster response speed and lower deployment cost.

[0034] S4. Output the classification and grading decision results.

[0035] In step S4, the classification and grading decision results may be output in the form of a structured list and / or a visual chart, but are not limited to. Furthermore, to achieve automatic triggering of field data protection, preferably, after outputting the classification and grading decision results, the method includes, but is not limited to: extracting the corresponding security level from the classification and grading decision results for each field in the structured data to be processed, and triggering the execution of a data protection strategy based on the security level to protect the corresponding data. The data protection strategy includes, but is not limited to, data encryption methods and / or data anonymization methods.

[0036] Therefore, based on the structured data automatic classification and grading method described in steps S1 to S4 above, a new scheme for end-to-end automated classification and grading of structured data based on a large language model and a classification and grading strategy knowledge base is provided. Specifically, after inputting the metadata of the structured data to be processed, the metadata is first parsed and semantically processed to generate a semantic information structure. This structure is then fed into a core reasoning engine built on a large language model and integrated with a classification and grading strategy knowledge base. The engine then performs a one-time logical reasoning on the semantic information structure based on the knowledge base, directly generating a decision result containing the category and security level of each field. Finally, this result is output. Thus, by constructing an end-to-end (from input to output) automated process with a large language model as the intelligent core, and utilizing its deep semantic understanding and reasoning capabilities, it replaces traditional rule matching and simple machine learning models, significantly improving the accuracy, efficiency, and interpretability of data classification and grading. It achieves near-human expert-level automated processing, facilitating practical application and promotion.

[0037] like Figure 2 As shown, the second aspect of this embodiment provides a virtual device for implementing the automatic classification and grading method of structured data described in the first aspect, including an information input unit, an information parsing unit, a logical reasoning unit, and a result output unit that are sequentially connected in communication. The information input unit is used to input the metadata of the structured data to be processed; The information parsing unit is used to parse and semanticize the metadata to generate a standardized semantic information structure. The logical reasoning unit is used to input the semantic information structure into the core reasoning engine, which is built based on a large language model and integrates a classification and grading strategy knowledge base. The core reasoning engine performs a one-time logical reasoning on the semantic information structure according to the classification and grading strategy knowledge base, and directly generates a classification and grading decision result containing the category and security level of each field data in the structured data to be processed. The result output unit is used to output the classification and grading decision results.

[0038] The working process, working details and technical effects of the aforementioned device provided in the second aspect of this embodiment can be found in the structured data automatic classification and grading method described in the first aspect, and will not be repeated here.

[0039] like Figure 3As shown, the third aspect of this embodiment provides a computer device for executing the automatic classification and grading method for structured data as described in the first aspect. The device includes a storage module, a processing module, and a transceiver module connected in sequence. The storage module stores a computer program, the transceiver module sends and receives messages, and the processing module reads the computer program and executes the automatic classification and grading method for structured data as described in the first aspect. Specifically, the storage module may include, but is not limited to, random-access memory (RAM), read-only memory (ROM), flash memory, first-in-first-out (FIFO) memory, and / or first-in-last-out (FILO) memory, etc.; the processing module may, but is not limited to, use a microprocessor of the STM32F105 series. Furthermore, the computer device may also include, but is not limited to, a power supply module, a display screen, and other necessary components.

[0040] The working process, working details and technical effects of the aforementioned computer device provided in the third aspect of this embodiment can be found in the automatic classification and grading method of structured data described in the first aspect, and will not be repeated here.

[0041] This fourth aspect of the embodiment provides a computer-readable storage medium storing instructions comprising the structured data automatic classification and grading method as described in the first aspect. Specifically, the computer-readable storage medium stores instructions that, when executed on a computer, perform the structured data automatic classification and grading method as described in the first aspect. The computer-readable storage medium refers to a data storage medium, which may include, but is not limited to, floppy disks, optical disks, hard disks, flash memory, USB flash drives, and / or Memory Sticks. The computer may be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices.

[0042] The working process, working details and technical effects of the aforementioned computer-readable storage medium provided in the fourth aspect of this embodiment can be found in the automatic classification and grading method of structured data as described in the first aspect, and will not be repeated here.

[0043] This fifth aspect of the embodiment provides a computer program product, including a computer program or instructions, which, when executed by a computer, implements the automatic classification and grading method for structured data as described in the first aspect. The computer may be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device.

[0044] Finally, it should be noted that the above description is merely a preferred embodiment of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for automatic classification and grading of structured data, characterized in that, include: Input the metadata of the structured data to be processed; The metadata is parsed and semantically processed to generate a standardized semantic information structure; The semantic information structure is fed into the core reasoning engine, which is built on a large language model and integrates a classification and grading strategy knowledge base. The core reasoning engine performs a one-time logical reasoning on the semantic information structure based on the classification and grading strategy knowledge base, and directly generates a classification and grading decision result containing the category and security level of each field data in the structured data to be processed. Output the classification and grading decision results.

2. The automatic classification and grading method for structured data according to claim 1, characterized in that, The metadata includes database structure information or metadata information.

3. The automatic classification and grading method for structured data according to claim 1, characterized in that, The metadata is parsed and semantically processed to generate a standardized semantic information structure, including: The metadata is parsed to obtain a parsing result containing the identified table headers and field meanings; The semantics of the parsing results are transformed into an information structure that can be understood by a core reasoning engine that is built on a large language model and integrates a classification and grading strategy knowledge base, thus serving as a standardized semantic information structure.

4. The automatic classification and grading method for structured data according to claim 1, characterized in that, The large language model is replaced by a classification and grading-specific model, which is a special model that is pre-trained and fine-tuned for the task of classifying and grading structured data and has fewer parameters than the general large language model.

5. The automatic classification and grading method for structured data according to claim 1, characterized in that, The classification and grading decision results are output, including: Output the classification and grading decision results in the form of a structured list and / or a visual chart.

6. The automatic classification and grading method for structured data according to claim 1, characterized in that, After outputting the classification and grading decision results, the method includes: For each field of the structured data to be processed, the corresponding security level is extracted from the classification and grading decision results, and a data protection strategy is triggered to protect the corresponding data based on the security level. The data protection strategy includes data encryption and / or data desensitization methods.

7. An automatic classification and grading device for structured data, characterized in that, It includes an information input unit, an information parsing unit, a logical reasoning unit, and a result output unit that are sequentially connected by communication. The information input unit is used to input the metadata of the structured data to be processed; The information parsing unit is used to parse and semanticize the metadata to generate a standardized semantic information structure. The logical reasoning unit is used to input the semantic information structure into the core reasoning engine, which is built based on a large language model and integrates a classification and grading strategy knowledge base. The core reasoning engine performs a one-time logical reasoning on the semantic information structure according to the classification and grading strategy knowledge base, and directly generates a classification and grading decision result containing the category and security level of each field data in the structured data to be processed. The result output unit is used to output the classification and grading decision results.

8. A computer device, characterized in that, It includes a storage module, a processing module, and a transceiver module that are sequentially connected in communication. The storage module is used to store computer programs, the transceiver module is used to send and receive messages, and the processing module is used to read the computer programs and execute the automatic classification and grading method for structured data as described in any one of claims 1 to 6.

9. A computer-readable storage medium, characterized in that... The computer-readable storage medium stores instructions that, when executed on a computer, perform the automatic classification and grading method for structured data as described in any one of claims 1 to 6.

10. A computer program product, comprising a computer program or instructions, characterized in that, When the computer program or the instructions are executed by the computer, they implement the automatic classification and grading method for structured data as described in any one of claims 1 to 6.

Images