Data processing method, device and equipment of enterprise resource planning system, and medium

Abstract

The embodiment of the application provides a data processing method, device and equipment of an enterprise resource planning system and a medium, comprising: obtaining system data to be processed; performing semantic analysis on the system data to be processed through a large language model to generate a data sensitivity grading result; performing algorithm rating on the system data to be processed to generate an algorithm score; performing strategy matching on the system data to be processed according to the data sensitivity grading result and the algorithm score to generate a strategy description file; performing algorithm matching according to the strategy description file to obtain a data privacy protection algorithm; and encrypting the system data to be processed according to the data privacy protection algorithm to complete privacy protection of the enterprise resource planning system data, thereby improving the accuracy of the data processing effect of the enterprise resource planning system.

Description

Technical Field

[0001] This application relates to the field of computer technology, and in particular to a data processing method, apparatus, equipment and medium for an enterprise resource planning system. Background Technology

[0002] Enterprise Resource Planning (ERP) systems, as core platforms for enterprise digital management, are widely used in multiple business areas. ERP systems need to process large amounts of structured and unstructured data, which frequently flows through real-world application scenarios. The data processed by ERP systems includes sensitive data, therefore, it is necessary to encrypt or anonymize this sensitive data.

[0003] In existing technologies, the data processing methods of enterprise resource planning systems mainly use preset data processing rules and fixed algorithms.

[0004] However, in existing technologies, the data in enterprise resource planning systems changes dynamically, and fixed algorithms lack assessments of algorithm security and system performance consumption. There is a risk of leakage of highly sensitive data due to insufficient algorithm strength, or excessive encryption of low-sensitivity data leading to excessive system resource utilization. The methods in existing technologies result in inaccurate data processing effects in enterprise resource planning systems. Summary of the Invention

[0005] This application provides a data processing method, apparatus, device, and medium for an enterprise resource planning system, which addresses the problem of inaccurate data processing results in existing enterprise resource planning systems.

[0006] In a first aspect, embodiments of this application provide a data processing method for an enterprise resource planning system, including:

[0007] Acquire system data to be processed;

[0008] The system data to be processed is semantically parsed using a large language model to generate data sensitivity classification results.

[0009] The system data to be processed is evaluated using an algorithm to generate an algorithm score;

[0010] Based on the data sensitivity classification results and the algorithm score, strategy matching is performed on the system data to be processed to generate a strategy description file;

[0011] Algorithm matching is performed based on the policy description file to obtain a data privacy protection algorithm;

[0012] The system data to be processed is encrypted according to the data privacy protection algorithm to complete the privacy protection of the enterprise resource planning system data.

[0013] In one possible implementation, the step of semantically parsing the system data to be processed using a large language model to generate a data sensitivity classification result includes: obtaining contextual information of the system data to be processed; identifying semantic patterns and privacy cues of data fields based on the contextual information; and performing semantic parsing of the system data to be processed using a large language model based on the semantic patterns and privacy cues of the data fields to generate a data sensitivity classification result.

[0014] In one possible implementation, identifying the semantic patterns and privacy cues of the data field based on the context information includes: extracting multimodal features from the system data to be processed to generate multidimensional semantic features; performing pattern recognition on the multidimensional semantic features based on the context information to generate pattern recognition results; and determining the semantic patterns and privacy cues of the data field based on the pattern recognition results.

[0015] In one possible implementation, before performing semantic parsing on the system data to be processed using a large language model to generate data sensitivity classification results, the method further includes: acquiring raw target data for model training; labeling the raw target data to obtain a training sample set for the large language model, wherein the training sample set of the large language model includes, but is not limited to, a data classification sample set, an algorithm description sample set, and a policy reasoning sample set; converting the formats of the data classification sample set, the algorithm description sample set, and the policy reasoning sample set to generate an adjusted training sample set; training the large language model based on the adjusted training sample set to generate a target model; and adjusting the target model according to a preset model adjustment strategy to generate a trained large language model.

[0016] In one possible implementation, after encrypting the system data to be processed according to the data privacy protection algorithm to complete the privacy protection of the enterprise resource planning system data, the method further includes: obtaining the execution log of the privacy protection policy; optimizing the policy generation logic of the large language model according to the execution log of the privacy protection policy to generate a logic optimization result; evaluating the logic optimization result to generate evaluation information; and updating the policy generation logic of the large language model according to the evaluation information to obtain the updated large language model.

[0017] In one possible implementation, encrypting the system data to be processed according to the data privacy protection algorithm to achieve privacy protection for the enterprise resource planning system data includes: performing data sharding on the system data to be processed, and distributing the generated multiple data shards to multiple processing nodes; encrypting the generated multiple data shards according to the data privacy protection algorithm to generate multiple shard encryption information; and aggregating the multiple shard encryption information to achieve privacy protection for the enterprise resource planning system data.

[0018] Secondly, embodiments of this application provide a data processing apparatus for an enterprise resource planning system, comprising:

[0019] The first acquisition module is used to acquire system data to be processed;

[0020] The data classification module is used to perform semantic parsing on the system data to be processed using a large language model, and generate data sensitivity classification results.

[0021] The algorithm rating module is used to perform algorithm rating on the system data to be processed and generate an algorithm score;

[0022] The strategy matching module is used to perform strategy matching on the system data to be processed based on the data sensitivity classification results and the algorithm score, and generate a strategy description file.

[0023] The algorithm matching module is used to perform algorithm matching based on the policy description file to obtain a data privacy protection algorithm;

[0024] The encryption module is used to encrypt the system data to be processed according to the data privacy protection algorithm, so as to complete the privacy protection of the enterprise resource planning system data.

[0025] Thirdly, embodiments of this application provide an electronic device, including: a memory and a processor;

[0026] The memory stores computer-executed instructions;

[0027] The processor executes computer execution instructions stored in the memory, causing the processor to perform the first aspect and / or various possible implementations of the first aspect as described above.

[0028] Fourthly, embodiments of this application provide a computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, are used to implement the first aspect and / or various possible implementations of the first aspect.

[0029] Fifthly, embodiments of this application provide a computer program product, including a computer program that, when executed by a processor, implements the first aspect and / or various possible implementations of the first aspect.

[0030] The data processing method, apparatus, device, and medium for an enterprise resource planning (ERP) system provided in this application embodiment utilize the semantic parsing capabilities of a large language model to perform semantic parsing on the system data to be processed, generating data sensitivity classification results. The reasoning capabilities of the large language model are then used to perform algorithmic rating on the sensitivity classification results, generating an algorithm score. Based on the data sensitivity classification results and the algorithm score, strategy matching is performed to generate a strategy description file. An appropriate data privacy protection algorithm is matched based on the strategy description file, and the system data to be processed is encrypted using the data privacy protection algorithm. This achieves privacy protection for the ERP system data and improves the accuracy of the ERP system's data processing performance. Attached Figure Description

[0031] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application.

[0032] Figure 1 This is a schematic diagram of the system structure of a computer device provided in an embodiment of this application;

[0033] Figure 2 Flowchart of the data processing method for the enterprise resource planning system provided in this application Figure 1 ;

[0034] Figure 3 Flowchart of the data processing method for the enterprise resource planning system provided in this application Figure 2 ;

[0035] Figure 4 Flowchart of the data processing method for the enterprise resource planning system provided in this application Figure 3 ;

[0036] Figure 5 A schematic diagram of the data processing device for the enterprise resource planning system provided in this application;

[0037] Figure 6 A schematic diagram of the structure of the electronic device provided in this application.

[0038] The accompanying drawings illustrate specific embodiments of this application, which will be described in more detail below. These drawings and descriptions are not intended to limit the scope of the concept in any way, but rather to illustrate the concept of this application to those skilled in the art through reference to particular embodiments. Detailed Implementation

[0039] Exemplary embodiments will now be described in detail, examples of which are illustrated in the accompanying drawings. When the following description relates to the drawings, unless otherwise indicated, the same numbers in different drawings denote the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this application. Rather, they are merely examples of apparatuses and methods consistent with some aspects of this application as detailed in the appended claims.

[0040] First, let's explain the terms used in this application:

[0041] Enterprise Resource Planning (ERP) system is an integrated software platform used to manage a company's core business processes, connecting data and information flows from multiple departments into a unified system.

[0042] Enterprise Resource Planning (ERP) systems, as core platforms for enterprise digital management, are widely used in multiple business areas. ERP systems need to process large amounts of structured and unstructured data, which frequently flows in real-world application scenarios. Sensitive data is included in the data processed by ERP systems, necessitating encryption or anonymization. Current technologies primarily use preset data processing rules and fixed algorithms for data processing. However, existing technologies suffer from the dynamic nature of ERP data and the lack of assessment of algorithm security and system performance consumption by fixed algorithms. This leads to risks of leakage for highly sensitive data due to insufficient algorithm strength, or excessive encryption of low-sensitivity data resulting in over-utilization of system resources. Consequently, existing methods result in inaccurate data processing performance in ERP systems.

[0043] To address the aforementioned technical problems, this application proposes the following technical concept: The inventors, considering the semantic understanding and dynamic reasoning capabilities of a large language model, perform semantic parsing on the system data to be processed, generating data sensitivity classification results. Furthermore, considering the reasoning capabilities of the large language model, they perform algorithmic rating on the sensitivity classification results, generating an algorithm score. Based on the data sensitivity classification results and the algorithm score, they perform strategy matching to generate a strategy description file. Based on the strategy description file, they match an appropriate data privacy protection algorithm, and then encrypt the system data to be processed using this algorithm. This achieves privacy protection for the enterprise resource planning system data and improves the accuracy of the data processing results of the enterprise resource planning system.

[0044] Figure 1 This is a schematic diagram of the system architecture of the computer device provided in an embodiment of this application. Figure 1As shown, the computer device includes: a receiving device 101, a processing device 102, and a display device 103.

[0045] It is understood that the structure illustrated in the embodiments of this application does not constitute a specific limitation on the data processing method of an enterprise resource planning system. In other feasible embodiments of this application, the above architecture may include more or fewer components than illustrated, or combine some components, or split some components, or arrange different components, which can be determined according to the actual application scenario and is not limited here. Figure 1 The components shown can be implemented in hardware, software, or a combination of both.

[0046] In the specific implementation process, the receiving device 101 can be an input / output interface or a communication interface, which can acquire system data to be processed.

[0047] The processing device 102 can perform algorithm matching based on the policy description file to obtain a data privacy protection algorithm, and encrypt the system data to be processed according to the data privacy protection algorithm.

[0048] The display device 103 can be used to display the encryption results generated after the above-mentioned data privacy protection algorithm encrypts the system data to be processed.

[0049] The display device can also be a touch screen, used to receive user commands while displaying the above content, so as to realize operation interaction with the user.

[0050] It should be understood that the above-mentioned processing device can be implemented by a processor reading instructions from memory and executing those instructions, or it can be implemented by a chip circuit.

[0051] Furthermore, the network architecture and business scenarios described in the embodiments of this application are for the purpose of more clearly illustrating the technical solutions of the embodiments of this application, and do not constitute a limitation on the technical solutions provided in the embodiments of this application. As those skilled in the art will know, with the evolution of network architecture and the emergence of new business scenarios, the technical solutions provided in the embodiments of this application are also applicable to similar technical problems.

[0052] Figure 2 Flowchart of the data processing method for the enterprise resource planning system provided in this application Figure 1 ,like Figure 2 As shown, the method includes:

[0053] S201: Obtain system data to be processed.

[0054] Specifically, by establishing a connection with the data bus of the ERP system, system data to be processed is retrieved according to a predetermined cycle.

[0055] In this embodiment, the system data to be processed includes, but is not limited to, employee information, transaction records, and business forms.

[0056] Specifically, after the data is acquired, it is uniformly packaged into a data packet, along with metadata information such as the data source module, data format, and timestamp.

[0057] S202: Semantic parsing of the system data to be processed is performed using a large language model to generate data sensitivity classification results.

[0058] Specifically, the data grading module identifies the semantic patterns and privacy cues of data fields based on the contextual information of the system data to be processed. The large language model then performs semantic parsing on the system data to be processed based on the semantic patterns and privacy cues of the data fields, generating data sensitivity grading results.

[0059] In this embodiment, the data classification module utilizes the semantic recognition and understanding capabilities of the large language model layer to perform context-level semantic analysis and sensitivity judgment on various types of input business data. The model infers the security level of the data and generates the classification reason and confidence label by understanding the relationship between field meaning, semantic pattern and privacy clue.

[0060] In this embodiment, the data grading module supports user-defined data grading systems and judgment criteria. Users can automatically trigger incremental fine-tuning or continuous learning processes of the model by providing labeled data samples or grading description documents.

[0061] In this embodiment, the data classification module is deployed on an internal API service or a local inference endpoint, and automatically performs classification labeling during the data entry, transmission and sharing processes.

[0062] S203: Perform algorithm rating on the system data to be processed and generate an algorithm score.

[0063] Specifically, the system data to be processed is input into a separate algorithm rating module, which performs calculations using a pre-built rating model.

[0064] In this embodiment, the algorithm score reflects the quantitative requirements of the privacy-preserving algorithm in terms of computational efficiency, storage overhead, and availability.

[0065] In this embodiment, the algorithm rating module calls the large language model layer to perform semantic parsing and comparison of the descriptive text, application documents and performance indicators of different privacy protection algorithms. The model automatically generates a score for the algorithm in three dimensions: usability, sensitivity adaptability and permission dependency by understanding the natural language descriptions of the algorithm's security strength, applicable scenarios and implementation complexity.

[0066] S204: Based on the data sensitivity classification results and algorithm scores, perform policy matching on the system data to be processed and generate a policy description file.

[0067] Specifically, the data sensitivity classification results and algorithm scores are used as input parameters and fed into the policy matching engine to generate a structured policy description file.

[0068] In this embodiment, the content recorded in the policy description file includes, but is not limited to, target data identifier, privacy protection strength, performance loss limit, and compliance requirements.

[0069] S205: Perform algorithm matching based on the policy description file to obtain the data privacy protection algorithm.

[0070] Specifically, the requirements in the policy description file are matched and filtered against the labels of various algorithms in the algorithm library to obtain the optimal data privacy protection algorithm.

[0071] In this embodiment, each algorithm in the algorithm library is configured with an attribute tag.

[0072] S206: Encrypt the system data to be processed according to the data privacy protection algorithm to complete the privacy protection of the enterprise resource planning system data.

[0073] Specifically, the system data to be processed is segmented, and the generated data segments are distributed to multiple processing nodes. The generated data segments are encrypted according to a data privacy protection algorithm, and the encrypted information of the multiple segments is aggregated to complete the privacy protection of the enterprise resource planning system data.

[0074] As can be seen from the above embodiments, by leveraging the semantic parsing capabilities of the large language model, semantic parsing is performed on the system data to be processed to generate data sensitivity classification results. The reasoning capabilities of the large language model are then used to perform algorithm rating on the sensitivity classification results, generating an algorithm score. Based on the data sensitivity classification results and the algorithm score, strategy matching is performed to generate a strategy description file. Based on the strategy description file, an appropriate data privacy protection algorithm is matched, and the system data to be processed is encrypted using the data privacy protection algorithm. This achieves privacy protection for the enterprise resource planning system data and improves the accuracy of the data processing effect of the enterprise resource planning system.

[0075] In one embodiment of this application, step S202 includes:

[0076] S2021: Obtain context information for the system data to be processed.

[0077] Specifically, the context information of each data field in the data packet is extracted from the metadata management library of the ERP system.

[0078] In this embodiment, the context information includes, but is not limited to, the name of the business module to which the data field belongs, the table name and field name definitions in the database, and the business process description.

[0079] S2022: Identify semantic patterns and privacy clues in data fields based on contextual information.

[0080] Specifically, multimodal feature extraction is performed on the system data to be processed to generate multidimensional semantic features. Pattern recognition is then performed on the multidimensional semantic features based on contextual information to generate pattern recognition results. Based on the pattern recognition results, the meaning patterns and privacy clues of the data fields are determined.

[0081] S2023: Using a large language model, semantic parsing is performed on the system data to be processed based on the semantic patterns and privacy clues of the data fields to generate data sensitivity classification results.

[0082] Specifically, the determined semantic patterns, privacy cues, and raw data fragments are constructed into natural language prompts, which are then input into the trained large language model. The large language model parses and outputs the hierarchical results.

[0083] As can be seen from the above embodiments, by introducing contextual information and privacy clues, and by using a large language model to perform dynamic semantic parsing of the system data to be processed based on the semantic patterns and privacy clues of the data fields, data sensitivity classification results are generated, thereby improving the ability of the large language model to judge data sensitivity.

[0084] Figure 3 Flowchart of the data processing method for the enterprise resource planning system provided in this application Figure 2 ,like Figure 3 As shown, in one embodiment of this application, step S2022 includes:

[0085] S301: Extract multimodal features from the system data to be processed to generate multidimensional semantic features.

[0086] In this embodiment, the extracted multimodal features include, but are not limited to, named entities of field names, numerical ranges, and data types.

[0087] S302: Perform pattern recognition on multi-dimensional semantic features based on contextual information and generate pattern recognition results.

[0088] Specifically, the extracted multi-dimensional semantic features are combined with the acquired contextual information and input into a predefined pattern recognition rule engine, which then generates pattern recognition results based on the built-in recognition rules.

[0089] S303: Determine the semantic patterns and privacy clues of the data fields based on the pattern recognition results.

[0090] In this embodiment, the semantic pattern is the business meaning category of the data.

[0091] In this embodiment, privacy clues are data that implies data sensitivity.

[0092] As can be seen from the above embodiments, by introducing multimodal feature extraction technology, multi-dimensional semantic features are generated from the data to be processed, and pattern recognition is performed in combination with the context, thereby realizing feature mining of data structure and data content. This enables the identification of more complex semantic patterns and more hidden privacy clues, enhancing the depth and breadth of semantic parsing of large language models.

[0093] Figure 4 Flowchart of the data processing method for the enterprise resource planning system provided in this application Figure 3 ,like Figure 4 As shown, the procedure before step S202 also includes:

[0094] S401: Obtain the raw target data for model training.

[0095] In this embodiment, the sources for obtaining the original target data include, but are not limited to, ERP data warehouses, publicly available datasets, and business knowledge documents.

[0096] S402: Label the original target data to obtain the training sample set of the large language model, wherein the training sample set of the large language model includes, but is not limited to, the data hierarchical sample set, the algorithm description sample set, and the policy reasoning sample set.

[0097] In this embodiment, the data classification sample set consists of data fields and their contexts, labeled with their respective sensitivity levels.

[0098] In this embodiment, the samples in the algorithm description sample set are the algorithm name and its technical document summary, labeled with the applicable privacy protection scenarios, computational overhead and strength level.

[0099] In this embodiment, the samples in the policy reasoning sample set are composite texts containing descriptions of data sensitivity and business scenarios, labeled with the privacy protection policy framework that should be matched.

[0100] In this embodiment, a training corpus consisting of three types of sub-datasets is constructed by combining manual annotation with automatic sampling.

[0101] S403: Convert the format of the data hierarchical sample set, algorithm description sample set, and policy inference sample set to generate an adjusted training sample set.

[0102] Specifically, the natural language annotations in the three sample sets are uniformly converted into a standardized format required for fine-tuning the large language model instructions. The adjusted format facilitates the model's understanding and learning of task logic.

[0103] S404: Train the large language model based on the adjusted training sample set to generate the target model.

[0104] Specifically, a supervised fine-tuning method is used to train the large language model. Through model training, the large language model learns the mapping relationship from data content to sensitivity classification and from scene description to strategy recommendation.

[0105] S405: Adjust the target model according to the preset model adjustment strategy to generate a trained large language model.

[0106] Specifically, a reinforcement learning strategy is used to adjust the target model, the output results of the model are ranked according to preferences, a reward model is trained, and the target model is iteratively optimized through a proximal policy optimization algorithm, ultimately generating a large language model after training.

[0107] As can be seen from the above embodiments, by constructing a specialized training sample set that includes multiple tasks such as data classification, algorithm description, and strategy reasoning, and by conducting targeted training and adjustments on the model, the model's deep understanding of domain terminology, business logic, and privacy rules is ensured, overcoming the problem of insufficient knowledge or accuracy that may exist in general models in specific domains.

[0108] In one embodiment of this application, after step S206, the method further includes:

[0109] S207: Obtain the execution log of the privacy protection policy.

[0110] In this embodiment, the execution log records, including but not limited to, the content of the policy description file, the algorithm selected for final matching, the encryption execution time, the amount of data processed, whether errors or alarms occurred during the encryption process, and the performance when querying data.

[0111] S208: Optimize the policy generation logic of the large language model based on the execution log of the privacy protection policy, and generate the logic optimization result.

[0112] Specifically, the strategy generation logic of the large language model is optimized based on the exceptions in the execution log, resulting in an optimized logic result.

[0113] S209: Evaluate the logic optimization results and generate evaluation information.

[0114] Specifically, in an isolated test environment, the rules of the policy matching engine are adjusted, and historical data is used for backtesting to evaluate whether the optimized policy effectively reduces business latency while ensuring security.

[0115] In this embodiment, the evaluation information includes quantitative indicators and review comments.

[0116] In this embodiment, the policy matching engine uses the large language model layer to analyze the semantic content, sensitivity level, and calling context of the transmitted data in real time, compares the existing algorithm ratings with historical matching records, and automatically determines the optimal combination of security algorithms and policies that should be used for this transmission.

[0117] S210: Update the strategy generation logic of the large language model based on the evaluation information to obtain the updated large language model.

[0118] Specifically, the logic optimization results are concretized into new training samples, added to the policy reasoning sample set, and an incremental training process is initiated. The existing large language model is then trained for a short period using the new mixed sample set to obtain the updated large language model.

[0119] As can be seen from the above embodiments, by adding a closed-loop feedback mechanism based on execution logs to continuously optimize and update the model strategy generation logic after encryption protection is implemented, the practical effect evaluation of privacy protection strategies and iterative learning of the model are realized. By analyzing the execution logs in real scenarios, the decision logic of the large language model is continuously optimized, thereby improving the performance of the large language model.

[0120] In one embodiment of this application, step S206 includes:

[0121] S2061: The system data to be processed is fragmented, and the generated data fragments are distributed to multiple processing nodes.

[0122] Specifically, based on the total amount of data and the system load, the acquired system data to be processed is divided into multiple data shards of balanced size, and the data shards are scheduled and distributed to different processing nodes in a distributed computing cluster consisting of multiple servers.

[0123] S2062: Encrypt the generated multiple data fragments according to the data privacy protection algorithm to generate multiple fragment encryption information.

[0124] Specifically, at each processing node, the specific implementation of the determined data privacy protection algorithm is invoked to perform encryption operations on the allocated data fragments, and each fragment generates corresponding fragment encryption information after encryption.

[0125] In this embodiment, the fragmented encryption information includes ciphertext data and metadata required for data fragmented encryption.

[0126] S2063: Aggregate multiple fragmented encrypted information to achieve privacy protection for enterprise resource planning system data.

[0127] Specifically, a coordinating node is set up to collect the fragmented encryption information generated by all processing nodes. The coordinating node verifies the integrity of each fragment of ciphertext and, according to the order and logical relationship of the original data fragments, re-aggregates and encapsulates all ciphertext fragments and global encrypted metadata into a privacy-protected data packet.

[0128] As can be seen from the above embodiments, by adopting data sharding technology, large-scale data processing tasks are decomposed and processed in parallel, and the resources of multiple processing nodes are used to encrypt the data shards, thereby improving the system's processing speed for enterprise resource planning system data.

[0129] Figure 5 A schematic diagram of the data processing device for the enterprise resource planning system provided in this application is shown below. Figure 5 As shown, the data processing device 50 of the enterprise resource planning system provided in this embodiment includes: a first acquisition module 501, a data classification module 502, an algorithm rating module 503, a strategy matching module 504, an algorithm matching module 505, and an encryption module 506.

[0130] The first acquisition module 501 is used to acquire system data to be processed.

[0131] The data classification module 502 is used to perform semantic parsing on the system data to be processed through a large language model and generate data sensitivity classification results.

[0132] The algorithm rating module 503 is used to perform algorithm rating on the system data to be processed and generate an algorithm score.

[0133] The strategy matching module 504 is used to perform strategy matching on the system data to be processed based on the data sensitivity classification results and algorithm scores, and generate a strategy description file.

[0134] The algorithm matching module 505 is used to perform algorithm matching based on the policy description file to obtain the data privacy protection algorithm.

[0135] The encryption module 506 is used to encrypt the system data to be processed according to the data privacy protection algorithm in order to complete the privacy protection of the enterprise resource planning system data.

[0136] In one embodiment of this application, the data classification module 502 includes:

[0137] The acquisition unit is used to acquire context information of the system data to be processed.

[0138] The identification unit is used to identify semantic patterns and privacy clues in data fields based on contextual information.

[0139] The semantic parsing unit is used to perform semantic parsing on the system data to be processed based on the semantic patterns and privacy clues of the data fields using a large language model, and generate data sensitivity classification results.

[0140] In one embodiment of this application, the identification unit includes:

[0141] The feature extraction subunit is used to extract multimodal features from the system data to be processed, generating multidimensional semantic features.

[0142] The pattern recognition subunit is used to perform pattern recognition on multi-dimensional semantic features based on contextual information and generate pattern recognition results.

[0143] The determination sub-unit is used to determine the semantic patterns and privacy clues of data fields based on the pattern recognition results.

[0144] In one embodiment of this application, the data processing apparatus 50 of the enterprise resource planning system further includes:

[0145] The second acquisition module is used to acquire the raw target data for model training.

[0146] The annotation module is used to annotate the original target data to obtain the training sample set of the large language model. The training sample set of the large language model includes, but is not limited to, the data hierarchical sample set, the algorithm description sample set, and the policy inference sample set.

[0147] The format conversion module is used to convert the formats of the data hierarchical sample set, algorithm description sample set, and policy inference sample set to generate an adjusted training sample set.

[0148] The training module is used to train the large language model based on the adjusted training sample set to generate the target model.

[0149] The adjustment module is used to adjust the target model according to the preset model adjustment strategy to generate a trained large language model.

[0150] In one embodiment of this application, the data processing apparatus 50 of the enterprise resource planning system further includes:

[0151] The third acquisition module is used to acquire the execution logs of the privacy protection policy.

[0152] The optimization module is used to optimize the policy generation logic of the large language model based on the execution log of the privacy protection policy, and generate the logic optimization result.

[0153] The evaluation module is used to evaluate the logic optimization results and generate evaluation information.

[0154] The update module is used to update the strategy generation logic of the large language model based on the evaluation information, so as to obtain the updated large language model.

[0155] In one embodiment of this application, the encryption module 506 includes:

[0156] The sharding unit is used to shard the system data to be processed, and distribute the generated data shards to multiple processing nodes.

[0157] The encryption unit is used to encrypt multiple data fragments generated according to a data privacy protection algorithm, thereby generating multiple fragment encryption information.

[0158] The aggregation unit is used to aggregate multiple fragmented encrypted information to achieve privacy protection for enterprise resource planning system data.

[0159] The data processing device for the enterprise resource planning system provided in this embodiment can execute the method provided in the above method embodiment. Its implementation principle and technical effect are similar, and will not be described in detail here.

[0160] Figure 6 A schematic diagram of the structure of the electronic device provided in this application. Figure 6 As shown, the electronic device 60 provided in this embodiment includes at least one processor 601 and a memory 602. Optionally, the electronic device 60 further includes a communication component 603. The processor 601, memory 602, and communication component 603 are connected via a bus.

[0161] In the specific implementation process, at least one processor 601 executes computer execution instructions stored in memory 602, causing at least one processor 601 to execute the data processing method of the enterprise resource planning system described above.

[0162] The specific implementation process of processor 601 can be found in the above method embodiments, and its implementation principle and technical effect are similar. It will not be repeated here.

[0163] In the above embodiments, it should be understood that the processor can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), etc. The general-purpose processor can be a microprocessor or any conventional processor. The steps of the method disclosed in this invention can be directly implemented by a hardware processor, or implemented by a combination of hardware and software modules within the processor.

[0164] The memory may include random access memory (RAM) and may also include non-volatile memory (NVM), such as at least one disk storage device.

[0165] The bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, or an Extended Industry Standard Architecture (EISA) bus, etc. Buses can be categorized as address buses, data buses, control buses, etc. For ease of illustration, the buses shown in the accompanying drawings are not limited to a single bus or a single type of bus.

[0166] This application also provides a computer program product, including a computer program that, when executed by a processor, implements the data processing method of the enterprise resource planning system described above.

[0167] This application also provides a computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, implement the data processing method of the enterprise resource planning system described above.

[0168] The aforementioned readable storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk. The readable storage medium can be any available medium accessible to a general-purpose or special-purpose computer.

[0169] An exemplary readable storage medium is coupled to a processor, enabling the processor to read information from and write information to the readable storage medium. Of course, the readable storage medium can also be a component of the processor. The processor and the readable storage medium can reside in an Application Specific Integrated Circuit (ASIC). Alternatively, the processor and the readable storage medium can exist as discrete components in the device.

[0170] The division of units is merely a logical functional division; in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be indirect coupling or communication connection through some interfaces, devices, or units, and may be electrical, mechanical, or other forms.

[0171] 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 units can be selected to achieve the purpose of this embodiment according to actual needs.

[0172] In addition, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0173] If a function is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, 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 of the various embodiments of this 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.

[0174] Those skilled in the art will understand that all or part of the steps of the above-described method embodiments can be implemented by hardware related to program instructions. The aforementioned program can be stored in a computer-readable storage medium. When executed, the program performs the steps of the above-described method embodiments; and the aforementioned storage medium includes various media capable of storing program code, such as ROM, RAM, magnetic disks, or optical disks.

[0175] Finally, it should be noted that other embodiments of the invention will readily occur to those skilled in the art upon consideration of the specification and practice of the invention disclosed herein. This invention is intended to cover any variations, uses, or adaptations of the invention that follow the general principles of the invention and include common knowledge or customary techniques in the art not disclosed herein, and is not limited to the precise structures described above and shown in the accompanying drawings, and various modifications and changes can be made without departing from its scope. The scope of the invention is limited only by the appended claims.

Claims

1. A data processing method for an enterprise resource planning system, characterized in that, include: Acquire system data to be processed; The system data to be processed is semantically parsed using a large language model to generate data sensitivity classification results. The system data to be processed is evaluated using an algorithm to generate an algorithm score; Based on the data sensitivity classification results and the algorithm score, strategy matching is performed on the system data to be processed to generate a strategy description file; Algorithm matching is performed based on the policy description file to obtain a data privacy protection algorithm; The system data to be processed is encrypted according to the data privacy protection algorithm to complete the privacy protection of the enterprise resource planning system data.

2. The method according to claim 1, characterized in that, The step of performing semantic parsing on the system data to be processed using a large language model to generate data sensitivity classification results includes: Obtain the context information of the system data to be processed; Identify semantic patterns and privacy clues in the data fields based on the context information; The system data to be processed is semantically parsed using a large language model based on the semantic patterns and privacy cues of the data fields, generating data sensitivity classification results.

3. The method according to claim 2, characterized in that, The step of identifying semantic patterns and privacy clues in data fields based on the context information includes: Multimodal feature extraction is performed on the system data to be processed to generate multidimensional semantic features; Based on the context information, pattern recognition is performed on the multi-dimensional semantic features to generate pattern recognition results; The semantic patterns and privacy clues of the data fields are determined based on the pattern recognition results.

4. The method according to claim 1, characterized in that, Before performing semantic parsing on the system data to be processed using a large language model to generate data sensitivity classification results, the process also includes: Obtain the raw target data for model training; The original target data is labeled to obtain a training sample set for the large language model, wherein the training sample set for the large language model includes, but is not limited to, a data hierarchical sample set, an algorithm description sample set, and a policy reasoning sample set; The data hierarchical sample set, algorithm description sample set, and policy inference sample set are converted into formats to generate an adjusted training sample set. The large language model is trained based on the adjusted training sample set to generate the target model; The target model is adjusted according to a preset model adjustment strategy to generate a trained large language model.

5. The method according to claim 1, characterized in that, After encrypting the system data to be processed according to the data privacy protection algorithm to complete the privacy protection of the enterprise resource planning system data, the method further includes: Obtain the execution logs of the privacy protection policy; The policy generation logic of the large language model is optimized based on the execution log of the privacy protection policy, and the logic optimization result is generated. The results of the logic optimization are evaluated to generate evaluation information; The strategy generation logic for updating the large language model is based on the evaluation information to obtain the updated large language model.

6. The method according to any one of claims 1 to 5, characterized in that, The step of encrypting the system data to be processed according to the data privacy protection algorithm to achieve privacy protection for the enterprise resource planning system data includes: The system data to be processed is segmented, and the generated data segments are distributed to multiple processing nodes. The generated multiple data fragments are encrypted according to the data privacy protection algorithm to generate multiple fragment encryption information; The multiple encrypted data fragments are aggregated to protect the privacy of enterprise resource planning system data.

7. A data processing device for an enterprise resource planning system, characterized in that, include: The first acquisition module is used to acquire system data to be processed; The data classification module is used to perform semantic parsing on the system data to be processed using a large language model, and generate data sensitivity classification results. The algorithm rating module is used to perform algorithm rating on the system data to be processed and generate an algorithm score; The strategy matching module is used to perform strategy matching on the system data to be processed based on the data sensitivity classification results and the algorithm score, and generate a strategy description file; The algorithm matching module is used to perform algorithm matching based on the policy description file to obtain a data privacy protection algorithm; The encryption module is used to encrypt the system data to be processed according to the data privacy protection algorithm, so as to complete the privacy protection of the enterprise resource planning system data.

8. An electronic device, characterized in that, include: Memory, processor; The memory stores computer-executed instructions; The processor executes computer execution instructions stored in the memory, causing the processor to perform the data processing method of the enterprise resource planning system 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 computer-executable instructions, which, when executed by a processor, are used to implement the data processing method of the enterprise resource planning system as described in any one of claims 1 to 6.

10. A computer program product, characterized in that, Includes a computer program that, when executed by a processor, implements the data processing method of the enterprise resource planning system according to any one of claims 1 to 6.

Images