Enterprise data asset management method and device, computer equipment and storage medium

Abstract

The invention relates to the technical field of data asset governance, and discloses an enterprise data asset governance method and device, computer equipment and a storage medium. Comprising the following steps: constructing a governance model taking real-time data assets as input and taking a target governance strategy as output; inputting the data assets into a governance model for governance analysis, and outputting an original governance strategy; synchronously carrying out toughness analysis, and outputting an original toughness strategy; and in the governance model, fusing the original governance strategy and the original toughness strategy by adopting an attention mechanism, and outputting a target governance strategy. The device, the computer equipment and the storage medium all correspond to the method. According to the method and the device, the adaptability of the original governance strategy, the effectiveness of the original toughness strategy and the quality of the target strategy are effectively improved, so that the risk early response rate and the disaster recovery coverage integrity are optimized.

Description

Technical Field

[0001] This application relates to the field of data asset governance technology, specifically a method, apparatus, computer equipment, and storage medium for enterprise data asset governance. Background Technology

[0002] Existing enterprise data asset governance methods often focus on single-dimensional routine governance, such as data cleaning and storage classification, or implement resilience measures such as data disaster recovery in isolation. This governance approach has two major flaws: First, the governance strategy lacks foresight throughout the entire life cycle, making it difficult to respond to potential risks in advance; Second, the integration of conventional governance with resilience strategies such as data disaster recovery is low, and the time and space adaptability is poor, such as the delay in disaster recovery execution and incomplete coverage.

[0003] When the aforementioned deficiencies lead to the risk of data assets being lost, damaged, or unusable, the effectiveness of governance is greatly reduced.

[0004] Therefore, there is an urgent need for a new technical solution for enterprise data asset governance. Summary of the Invention

[0005] The purpose of this application is to provide a method, apparatus, computer equipment and storage medium for enterprise data asset governance, so as to solve the technical problem of poor coordination between conventional governance and data disaster recovery in the prior art.

[0006] To achieve the above objectives, this application provides an enterprise data asset governance method that optimizes governance in response to the full lifecycle risk management of enterprise-generated data assets. This optimization, based on governance analysis and resilience analysis, outputs a target governance strategy for the data assets. The method includes: Historical data assets and their corresponding historical governance strategies are acquired to construct a governance model, which takes real-time data assets as input and the target governance strategy as output. The data assets are input into the governance model for governance analysis, and the original governance strategy is output; simultaneously, the resilience analysis is performed, and the original resilience strategy is output. In the governance model, an attention mechanism is used to fuse the original governance strategy and the original resilience strategy to output the target governance strategy.

[0007] Preferably, the construction of the governance model includes: Unstructured historical data assets are transformed into structured asset features, which are used to characterize the information required for the governance analysis and the resilience analysis, and the asset features include at least the associated entities and timestamps; The unstructured historical governance strategies are transformed into structured strategy features, which include governance features corresponding to the governance analysis and resilience features corresponding to the resilience analysis. The governance features include at least conventional governance strategies, and the resilience features include at least risk response strategies. The asset characteristics, governance characteristics, and resilience characteristics are reorganized based on time series data, and a governance strategy mapping relationship is constructed based on the reorganization result. The governance strategy mapping relationship is stored in a preset neural network model for training to obtain the governance model.

[0008] Preferably, the construction of the governance strategy mapping relationship includes: Based on the historical data assets and their corresponding historical governance strategies, the feature distribution of the asset characteristics, governance characteristics and resilience characteristics under the same time series is obtained; Feature extraction is performed based on the aforementioned feature distribution to obtain strategy temporal features and strategy spatial features. The strategy temporal features are used to characterize the temporal correspondence between an asset feature and a governance feature and / or a resilience feature, and the strategy spatial features are used to characterize the spatial correspondence between an asset feature and a governance feature and / or a resilience feature. By combining a preset accuracy threshold, the temporal features and spatial features of the strategy are verified. Once the verification is successful, the mapping relationship of the governance strategy is obtained.

[0009] Preferably, the output of the original governance strategy includes: Based on the asset characteristics of real-time data asset transformation, the governance characteristics preset in the governance model are matched, the execution parameters corresponding to the conventional governance strategy are called, and the business ownership and quality rating of the data asset are combined to generate and output the original governance strategy. The original governance strategy includes at least measures corresponding to data cleaning, standardization processing, storage classification and access control.

[0010] Preferably, the output of the original resilience strategy includes: Risk scenarios are identified based on associated entities and timestamps of real-time data assets. Risk response strategies in resilience features are matched based on the identification results. The original resilience strategy is generated and output by combining the security level and redundancy requirements of the data assets. The original resilience strategy includes at least the measures corresponding to the fault recovery process, backup scheme, risk warning threshold and fault tolerance adaptation rules.

[0011] Preferably, the construction of the governance model further includes: Based on the long-term series of the entire life cycle of data assets, multiple sets of the time features and spatial features of the policies that have been verified in the mapping relationship of the governance policies are extracted to form a spatiotemporal feature sequence set; A spatiotemporal attention optimization unit is added to the neural network model, and the spatiotemporal feature sequence set is used as the training sample of the spatiotemporal attention optimization unit, with preset weight adjustment rules for the time dimension and the spatial dimension. The spatiotemporal attention optimization unit is trained using the spatiotemporal feature sequence set, enabling the spatiotemporal attention optimization unit to determine the early response weight based on the policy time features of the long-term sequence and the range expansion weight based on the policy space features. The trained spatiotemporal attention optimization unit is associated with the attention mechanism of the governance model, so that the attention mechanism calls the spatiotemporal attention optimization unit to perform early response weights and range expansion weights.

[0012] Preferably, the output of the target governance strategy includes: Obtain the asset characteristics corresponding to the real-time data assets, and combine them with the long-term series of the data assets throughout their entire lifecycle to match the policy time characteristics and policy space characteristics associated in the governance policy mapping relationship; The matched policy temporal features and policy spatial features are input into the spatiotemporal attention optimization unit of the optimized governance model, and the early response weight and the range expansion weight are determined according to the preset weight adjustment rules. The early response weight and the range expansion weight are substituted into the fusion operation process of the attention mechanism to adjust the fusion ratio of the original governance strategy and the original resilience strategy. The early response weight is used to adjust the prediction adaptation of the strategy execution time, and the range expansion weight is used to adjust the adaptation expansion of the strategy execution range. According to the adjusted fusion ratio, the original governance strategy and the original resilience strategy are fused to generate and output the target governance strategy.

[0013] To achieve the above objectives, this application also provides an enterprise data asset governance apparatus, which applies the enterprise data asset governance method described above. The apparatus includes: The model building module is used to acquire historical data assets and their corresponding historical governance strategies, and to build a governance model. The governance model takes real-time data assets as input and the target governance strategy as output. The raw output module is used to input data assets into the governance model for governance analysis and output raw governance strategies; it also performs resilience analysis simultaneously and outputs raw resilience strategies. The fusion output module is used to fuse the original governance strategy and the original resilience strategy in the governance model using an attention mechanism, and output the target governance strategy.

[0014] To achieve the above objectives, this application also provides a computer device for enterprise data asset management, including at least one processor, at least one memory, and a data bus; The processor and the memory communicate with each other via the data bus; The memory stores program instructions that can be executed by the processor, which invokes the program instructions to execute the enterprise data asset governance method described above.

[0015] To achieve the above objectives, this application also provides a storage medium storing a computer program thereon, which, when executed by a processor, implements the enterprise data asset governance method described above.

[0016] Beneficial effects: The enterprise data asset governance method, apparatus, computer equipment and storage medium of this application have achieved effective improvement in the adaptability of the original governance strategy, the effectiveness of the original resilience strategy and the quality of the target strategy, thereby optimizing the risk early response rate and the integrity of disaster recovery coverage. Attached Figure Description

[0017] To more clearly illustrate the technical solutions in the embodiments of this application 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 this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0018] Figure 1 A flowchart illustrating the enterprise data asset governance method provided in this application embodiment; Figure 2 Flowchart for constructing the governance model provided in the embodiments of this application Figure 1 ; Figure 3 A flowchart illustrating the construction of the governance strategy mapping relationship provided in this application embodiment; Figure 4 Flowchart for constructing the governance model provided in the embodiments of this application Figure 2 ; Figure 5 A flowchart illustrating the output of the target governance strategy provided in the embodiments of this application; Figure 6 This is a structural block diagram of the enterprise data asset governance device in this embodiment; In the diagram: 100, Model building module; 200, Original output module; 300, Fusion output module.

[0019] The implementation, functional features, and advantages of this invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0020] The technical solutions in the embodiments of this application will be clearly and completely described below. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments in this application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of this application.

[0021] In this document, the term "comprising" is intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0022] The core of this embodiment lies in constructing a governance model that integrates spatiotemporal characteristics. Based on historical data and governance strategies throughout the entire lifecycle of data assets, it outputs conventional governance strategies through governance analysis and risk response strategies including data disaster recovery through resilience analysis. These two strategies are then integrated through an optimized attention mechanism to output a forward-looking and comprehensive target governance strategy. The implementation process can be simply summarized as the sequential construction of the governance model, the output of the original strategy, and the fusion and output of the target strategy. Each step is closely linked, forming a closed-loop governance system.

[0023] Reference Figure 1 , Figure 1 A flowchart illustrating the enterprise data asset governance method provided in this application embodiment.

[0024] like Figure 1 As shown, this embodiment discloses an enterprise data asset governance method that optimizes governance in response to the full lifecycle risk management of enterprise-generated data assets. This optimization, based on governance analysis and resilience analysis, outputs a target governance strategy for the data assets. The method includes: S1: Obtain historical data assets and their corresponding historical governance strategies to construct a governance model. The governance model takes real-time data assets as input and the target governance strategy as output.

[0025] Reference Figure 2 , Figure 2 Flowchart for constructing the governance model provided in the embodiments of this application Figure 1 .

[0026] like Figure 2 As shown, the specific construction of the governance model includes: S11: Transform unstructured historical data assets into structured asset features. These asset features are used to characterize the information required for governance and resilience analysis, and include at least related entities and timestamps. S12: Transform unstructured historical governance strategies into structured strategy features. The strategy features include governance features corresponding to governance analysis and resilience features corresponding to resilience analysis. The governance features include at least conventional governance strategies, and the resilience features include at least risk response strategies. S13: Reorganize asset characteristics, governance characteristics, and resilience characteristics based on time series data, and construct governance strategy mapping relationships based on the reorganization results; S14: Store the governance strategy mapping relationship into a preset neural network model for training to obtain the governance model.

[0027] In the specific application of this embodiment, we use the Transformer neural network as the basic architecture. The construction process is divided into three steps, focusing on connecting spatiotemporal features and model optimization, while providing a foundation for the output of the original governance strategy and the original resilience strategy.

[0028] In one specific application, we acquire a company's historical data assets. For unstructured data (such as robot operation log text and supply chain Excel ledgers), we transform it into structured asset features through OCR (Optical Character Recognition), NLP (Natural Language Processing) word segmentation, and feature encoding. These features include at least: Related entities: such as equipment number, supplier number, etc., and each is uniquely identified by a code; Timestamp: Retrieves the raw data generation time accurate to the second; Additional features: data type, for example, such as real-time running data and static ledger data; business affiliation, for example, such as production workshop and purchasing department; quality rating, for example, such as A, B, and C, which is based on data integrity, accuracy, etc.; security level, for example, such as confidential and ordinary, which is based on data sensitivity. Based on this, a basis is provided for the strategy adaptation of the original governance strategy and the original resilience strategy.

[0029] In the specific application of this embodiment, historical governance strategies are transformed into structured strategy features, clearly distinguishing the following two types of core features: Governance features, which correspond to conventional governance strategies, include at least: data cleaning rules, standardization specifications, storage tiering strategies, and access control schemes, with each governance feature associated with specific execution parameters; Resilience features, which correspond to risk response strategies for data disaster recovery, include at least: fault recovery process, backup plan, risk warning threshold, and fault tolerance adaptation rules. Among them, the parameters corresponding to data disaster recovery, for example, are core resilience features such as synchronization frequency and disaster recovery center.

[0030] Reference Figure 3 , Figure 3 A flowchart illustrating the construction of the governance strategy mapping relationship provided in this application embodiment.

[0031] like Figure 3 As shown, the specific construction of the governance strategy mapping relationship includes: S131: Based on historical data assets and their corresponding historical governance strategies, obtain the feature distribution of asset characteristics, governance characteristics and resilience characteristics of the two under the same time series; S132: Perform feature extraction based on the feature distribution to obtain strategy temporal features and strategy spatial features. Strategy temporal features are used to characterize the temporal correspondence between an asset feature and a governance feature and / or a resilience feature, while strategy spatial features are used to characterize the spatial correspondence between an asset feature and a governance feature and / or a resilience feature. S133: Combine the preset accuracy threshold to verify the temporal and spatial characteristics of the strategy. After the verification is successful, the governance strategy mapping relationship is obtained.

[0032] In the specific application of this embodiment, feature distribution association is first performed based on time series. In a simple example, we associate asset features with strategy features at the same time point using "hours" as the unit, forming a "time-asset-strategy" correspondence table. Secondly, a convolutional neural network, but not limited to, is used to extract features from the above association table, yielding the following two types of core features: Strategy time characteristics are used to characterize the temporal relationship between asset characteristics and strategy characteristics. Examples include "the operation data of a certain device is cleaned at a certain time and disaster recovery is performed after a preset time interval" and "in historical risk events, most data loss occurs within a short period of time after cleaning". Strategy space features are used to characterize the spatial correlation between asset features and strategy features, i.e., the scope of strategy coverage. For example, "the operational data of a certain device is associated with production plan data and supply chain component data. The current disaster recovery only covers the operational data and does not cover the associated data," or "the associated entity hierarchy is level 3: a certain device → a certain production workshop → a certain factory." Finally, feature verification is performed. In this embodiment, a preset accuracy threshold is set (this accuracy threshold is set based on actual needs; theoretically, we set it to 95%). Taking the time dimension as an example, the effectiveness of the features is verified by comparing the "strategy execution effect guided by spatiotemporal features" with the "actual risk response needs" in historical risk events. For example, based on the time feature of "disaster recovery 2 hours after cleaning," in 100 historical risk simulations, data loss can be avoided in 98 of them, with an accuracy rate of 98% > 95%. The verification is passed, forming a governance strategy mapping relationship. This governance strategy mapping relationship includes asset features, governance features, resilience features, strategy time features, and strategy space features.

[0033] Reference Figure 4 , Figure 4 Flowchart for constructing the governance model provided in the embodiments of this application Figure 2 .

[0034] like Figure 4 As shown, the construction of the governance model specifically also includes: S15: Based on the long-term series of the entire life cycle of data assets, extract multiple sets of policy time features and policy space features that have been verified in the governance policy mapping relationship to form a spatiotemporal feature sequence set; S16: Add a spatiotemporal attention optimization unit to the neural network model, use the spatiotemporal feature sequence set as the training sample of the spatiotemporal attention optimization unit, and preset the weight adjustment rules of the time dimension and the space dimension. S17: Train the spatiotemporal attention optimization unit through a set of spatiotemporal feature sequences, so that the spatiotemporal attention optimization unit determines the early response weight based on the policy time features of long time series and the range expansion weight based on the policy space features. S18: Associate the trained spatiotemporal attention optimization unit with the attention mechanism of the governance model, so that the attention mechanism calls the spatiotemporal attention optimization unit to perform early response weights and range expansion weights.

[0035] In the specific application of this embodiment: For the spatiotemporal feature sequence set formed by S15, we extract the combination of the above-verified policy time features and policy space features based on the time series corresponding to the entire life cycle of data assets (which includes generation, storage, use, and destruction) (for example, a 12-month data retention period), forming a spatiotemporal feature sequence set. For example, the feature combination consisting of "cleaning-disaster recovery time interval" and "associated entity coverage level" of a certain device in each month within 12 months, ensuring coverage of the spatiotemporal feature differences at different stages of the entire life cycle.

[0036] Regarding the spatiotemporal attention optimization unit added in S16, we added a spatiotemporal attention optimization unit to the existing Transformer neural network model. This unit consists of two sub-units: The time weight adjustment subunit presets adjustment rules based on the time characteristics of the input strategy. For example, if historical risks mostly occur within 1-3 hours after cleanup, the disaster recovery advance weight is positively correlated with the time interval, with a weight of 0.7 for a 2-hour interval and 0.8 for a 3-hour interval. The spatial weight adjustment submodule presets adjustment rules based on the input policy space characteristics. For example, if the associated entity level is 3, the disaster recovery scope expansion weight is positively correlated with the level number, with a weight of 0.7 for level 3 and 0.8 for level 4.

[0037] For S17, the spatiotemporal feature sequence set is used as the training sample for this unit. The training objective is to enable the unit to output stable temporal advance weights and spatial expansion weights based on the input policy temporal features and / or policy spatial features.

[0038] For S18, the trained spatiotemporal attention optimization unit is associated with the attention mechanism module of the Transformer model. This allows the attention mechanism to call the temporal advance weights and spatial expansion weights output by the unit in real time when fusing strategies, thereby completing the governance model optimization and providing a technical foundation for subsequent strategy fusion.

[0039] S2: Input data assets into the governance model for governance analysis and output the original governance strategy; simultaneously perform resilience analysis and output the original resilience strategy.

[0040] Specifically, the output of the original governance strategy includes: Based on the asset characteristics of real-time data asset transformation, the governance characteristics preset in the governance model are matched, the execution parameters corresponding to the conventional governance strategy are called, and the business ownership and quality rating of the data asset are combined to generate and output the original governance strategy. The original governance strategy includes at least the measures corresponding to data cleaning, standardization processing, storage classification and access control.

[0041] Specifically, the output of the original resilience strategy includes: Risk scenarios are identified based on the associated entities and timestamps of real-time data assets. Based on the identification results, risk response strategies in resilience features are matched. Combined with the security level and redundancy requirements of data assets, original resilience strategies are generated and output. The original resilience strategies include at least the fault recovery process, backup plan, risk warning threshold, and measures corresponding to fault tolerance adaptation rules.

[0042] In this embodiment, we input real-time data assets into the governance model before optimization and output two types of original policies, thereby providing input for policy fusion. In a simple example, real-time data assets can be the operating parameters generated by a device at a certain moment and its corresponding unstructured data.

[0043] For the original governance strategy output: First, real-time asset feature matching is performed to transform real-time data assets into structured features that match the pre-set governance features in the governance model; second, based on the matched governance features, specific execution parameters are determined; finally, based on the execution parameters, a directly executable strategy document is generated, clearly defining the implementing entity, execution time, verification standards, etc., for each measure.

[0044] For the output of the original resilience strategy: First, risk scenario identification and feature matching are performed. Based on real-time asset characteristics, core risk scenarios are identified and matched with the resilience features preset in the governance model. Second, the specific execution parameters are determined by combining the security level and redundancy requirements of data assets. Finally, a resilience strategy execution manual is formed based on the specific parameters, which clarifies the triggering conditions for disaster recovery synchronization, the early warning response process, the calculation logic for fault tolerance filling, etc. The triggering conditions may be, but are not limited to, timed triggering and / or data volume triggering.

[0045] S3: In the governance model, an attention mechanism is used to integrate the original governance strategy and the original resilience strategy to output the target governance strategy.

[0046] Reference Figure 5 , Figure 5 A flowchart illustrating the output of the target governance strategy provided in the embodiments of this application.

[0047] like Figure 5 As shown, the specific outputs of the target governance strategy include: S181: Obtain the asset characteristics corresponding to real-time data assets, and combine them with the long-term series of the entire life cycle of the data assets to match the policy time characteristics and policy space characteristics associated in the governance policy mapping relationship; S182: Input the matched policy temporal features and policy spatial features into the spatiotemporal attention optimization unit of the optimized governance model, and determine the early response weight and range expansion weight respectively according to the preset weight adjustment rules. S183: Substitute the early response weight and the scope expansion weight into the fusion operation process of the attention mechanism to adjust the fusion ratio of the original governance strategy and the original resilience strategy. The early response weight is used to adjust the prediction adaptation of the strategy execution time, and the scope expansion weight is used to adjust the adaptation expansion of the strategy execution range. S184: According to the adjusted fusion ratio, the original governance strategy and the original resilience strategy are fused to generate and output the target governance strategy.

[0048] In a specific application of this embodiment, the features of the aforementioned real-time data assets are input into the optimized governance model, which includes a spatiotemporal attention optimization unit. The target governance strategy is then output through the following steps: For S181, spatiotemporal feature matching and weight calculation are performed. In a specific application: Combine this with a long-term time series of the entire lifecycle of data assets, for example, the aforementioned 12 months; Match the time characteristics of the policies associated in the governance policy mapping relationship. For example, as mentioned above, "Disaster recovery is performed 2 hours after cleaning of a certain device, and historical risks mostly occur within 1-3 hours after cleaning." Match the spatial characteristics of the policies associated in the governance policy mapping relationship. For example, as mentioned above, "Associated entity level 3: a certain device → a certain production workshop → a certain factory, and the current disaster recovery only covers the data of a certain device." Two types of features are input into the spatiotemporal attention optimization unit, and the output is: time advance weight 0.7, which means that the disaster recovery needs to be brought forward by 70% of the original interval time. If the original interval is 2 hours, it will be brought forward to 1.4 hours; spatial expansion weight 0.7, which means that the disaster recovery scope needs to be expanded by 70% of the associated entity levels. If the original coverage is level 1, it will be expanded to level 1.7, which is rounded up to level 2, i.e. "a certain device → a certain production workshop".

[0049] For S182 to S183, time dimension adjustment, spatial dimension adjustment, and weight fusion are performed. In a specific application: In terms of time dimension adjustment, based on a time advance weight of 0.7, the disaster recovery synchronization time in the original resilience strategy is advanced by 1.4 hours, and it is connected with the data cleaning time of the original governance strategy this year. At this time, the synchronization time after cleaning is greatly advanced, thus avoiding the high-risk period. In terms of spatial dimension adjustment, based on the spatial expansion weight of 0.7, the original resilience strategy of "only covering the data of a certain device" is expanded to "covering the associated data of a certain device → a certain production workshop" (i.e., level 2 associated entities), thereby avoiding the governance gap caused by the lack of associated data. Regarding weight fusion, the attention mechanism adjusts the fusion ratio of the original governance strategy and the original resilience strategy based on the aforementioned spatiotemporal weights. For example, the weight of conventional governance is 40%, thereby retaining core cleaning, storage and other measures; the weight of resilience strategy is 60%, thereby strengthening disaster recovery, early warning and other measures.

[0050] For S184, distinguish between the routine governance component, the resilience governance component, and the collaborative verification component. In a specific application: For the routine governance portion, the core measures of the original governance strategy are retained, and the timing of their execution is optimized; for example, the timing of data cleaning, standardization processing, storage execution, and control permissions of the original governance strategy are maintained. For the resilience governance component, optimize disaster recovery and early warning measures; integrate the core measures of the optimized original resilience strategy with those of the retained original governance strategy; for example, the connection between data disaster recovery timing and the expansion of disaster recovery scope; For the collaborative verification part, after the disaster recovery synchronization is completed, the system automatically compares the verification codes of the local data with those of the disaster recovery center. If the verification passes, a governance completion report is generated; if the verification fails, a resynchronization is triggered to ensure the effectiveness of policy execution.

[0051] The enterprise data asset governance method based on this embodiment effectively improves the adaptability of the original governance strategy, the effectiveness of the original resilience strategy, and the quality of the target strategy, thereby optimizing the risk early response rate and the integrity of disaster recovery coverage.

[0052] The enterprise data asset governance method based on this embodiment achieves at least the following technical effects: By precisely matching conventional governance strategies with the characteristics of data assets (business ownership, quality rating), and determining execution parameters through quantitative thresholds (such as cleaning thresholds), the poor adaptability of traditional fixed rules can be avoided. Data disaster recovery is incorporated into the core parameters of resilience strategy, the synchronization frequency and off-site disaster recovery level are clearly defined, and the recovery process is dynamically adjusted in combination with risk scenarios to solve the problem of disconnect between traditional disaster recovery and governance. Based on the spatiotemporal characteristics, the governance model is optimized and the strategy of "time prediction (early disaster recovery) + spatial expansion (expanded scope)" is integrated through the attention mechanism to form a closed loop of full life cycle governance, breaking through the bottlenecks of post-event response and scope limitation of existing technologies.

[0053] Reference Figure 6 , Figure 6 This is a structural block diagram of the enterprise data asset governance device in this embodiment.

[0054] like Figure 6 As shown, this embodiment also discloses an enterprise data asset governance device, which applies the enterprise data asset governance method described above. The device includes: The model building module is used to acquire historical data assets and their corresponding historical governance strategies, and to build a governance model. The governance model takes real-time data assets as input and the target governance strategy as output. The raw output module is used to input data assets into the governance model for governance analysis and output raw governance strategies; it also performs resilience analysis and outputs raw resilience strategies simultaneously. The fusion output module is used in the governance model to fuse the original governance strategy and the original resilience strategy using an attention mechanism, and output the target governance strategy.

[0055] This embodiment also discloses a computer device for enterprise data asset governance, including at least one processor, at least one memory, and a data bus; The processor and memory communicate with each other via a data bus; The memory stores program instructions that can be executed by the processor, which invokes the program instructions to execute the enterprise data asset governance method described above.

[0056] This embodiment also discloses a storage medium storing a computer program, characterized in that the computer program, when executed by a processor, implements the enterprise data asset governance method described above.

[0057] It should be noted that the enterprise data asset governance device, computer equipment, and storage medium in this embodiment correspond to the aforementioned enterprise data asset governance method. Therefore, any content not specifically described in the enterprise data asset governance device, computer equipment, and storage medium in this embodiment, including but not limited to functional definitions, working principles, and technical effects, can be referred to the description in the aforementioned enterprise data asset governance method, and will not be repeated here.

[0058] In the embodiments provided in this application, it should be understood that the embodiments described herein can be implemented in hardware, software, firmware, middleware, code, or any suitable combination thereof. For hardware implementation, the processor may be implemented in one or more of the following: application-specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field-programmable gate arrays (FPGAs), processors, controllers, microcontrollers, microprocessors, other electronic units designed to implement the functions described herein, or combinations thereof. For software implementation, some or all of the processes of the embodiments may be performed by a computer program instructing the associated hardware. During implementation, the program may be stored in a computer-readable storage medium or transmitted as one or more instructions or code on a computer-readable storage medium. Computer-readable storage media include computer storage media and communication media, wherein communication media include any medium that facilitates the transmission of a computer program from one place to another. Storage media may be any available medium accessible to a computer. Computer-readable storage media may include, but are not limited to, RAM, ROM, EEPROM, CD-ROM or other optical disk storage, magnetic disk storage media or other magnetic storage devices, or any other medium capable of carrying or storing desired program code having the form of instructions or data structures and accessible to a computer.

[0059] Finally, it should be noted that the above description is only a preferred embodiment of this application and is not intended to limit this application. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the protection scope of this application.

Claims

1. A method for enterprise data asset governance, characterized in that, Optimizing the full lifecycle risk management of enterprise-generated data assets, this optimization is based on governance analysis and resilience analysis to output target governance strategies for data assets, the method including: Historical data assets and their corresponding historical governance strategies are acquired to construct a governance model, which takes real-time data assets as input and the target governance strategy as output. The data assets are input into the governance model for governance analysis, and the original governance strategy is output; simultaneously, the resilience analysis is performed, and the original resilience strategy is output. In the governance model, an attention mechanism is used to fuse the original governance strategy and the original resilience strategy to output the target governance strategy.

2. The enterprise data asset governance method according to claim 1, characterized in that, The construction of the governance model includes: Unstructured historical data assets are transformed into structured asset features, which are used to characterize the information required for the governance analysis and the resilience analysis, and the asset features include at least the associated entities and timestamps; The unstructured historical governance strategies are transformed into structured strategy features, which include governance features corresponding to the governance analysis and resilience features corresponding to the resilience analysis. The governance features include at least conventional governance strategies, and the resilience features include at least risk response strategies. The asset characteristics, governance characteristics, and resilience characteristics are reorganized based on time series data, and a governance strategy mapping relationship is constructed based on the reorganization result. The governance strategy mapping relationship is stored in a preset neural network model for training to obtain the governance model.

3. The enterprise data asset governance method according to claim 2, characterized in that, The construction of the governance strategy mapping relationship includes: Based on the historical data assets and their corresponding historical governance strategies, the feature distribution of the asset characteristics, governance characteristics and resilience characteristics under the same time series is obtained; Feature extraction is performed based on the aforementioned feature distribution to obtain strategy temporal features and strategy spatial features. The strategy temporal features are used to characterize the temporal correspondence between an asset feature and a governance feature and / or a resilience feature, and the strategy spatial features are used to characterize the spatial correspondence between an asset feature and a governance feature and / or a resilience feature. By combining a preset accuracy threshold, the temporal features and spatial features of the strategy are verified. Once the verification is successful, the mapping relationship of the governance strategy is obtained.

4. The enterprise data asset governance method according to claim 2, characterized in that, The output of the original governance strategy includes: Based on the asset characteristics of real-time data asset transformation, the governance characteristics preset in the governance model are matched, the execution parameters corresponding to the conventional governance strategy are called, and the business ownership and quality rating of the data asset are combined to generate and output the original governance strategy. The original governance strategy includes at least measures corresponding to data cleaning, standardization processing, storage classification and access control.

5. The enterprise data asset governance method according to claim 2, characterized in that, The output of the original resilience strategy includes: Risk scenarios are identified based on associated entities and timestamps of real-time data assets. Risk response strategies in resilience features are matched based on the identification results. The original resilience strategy is generated and output by combining the security level and redundancy requirements of the data assets. The original resilience strategy includes at least the measures corresponding to the fault recovery process, backup scheme, risk warning threshold and fault tolerance adaptation rules.

6. The enterprise data asset governance method according to claim 3, characterized in that, The construction of the governance model also includes: Based on the long-term series of the entire life cycle of data assets, multiple sets of the time features and spatial features of the policies that have been verified in the mapping relationship of the governance policies are extracted to form a spatiotemporal feature sequence set; A spatiotemporal attention optimization unit is added to the neural network model, and the spatiotemporal feature sequence set is used as the training sample of the spatiotemporal attention optimization unit, with preset weight adjustment rules for the time dimension and the spatial dimension. The spatiotemporal attention optimization unit is trained using the spatiotemporal feature sequence set, enabling the spatiotemporal attention optimization unit to determine the early response weight based on the policy time features of the long-term sequence and the range expansion weight based on the policy space features. The trained spatiotemporal attention optimization unit is associated with the attention mechanism of the governance model, so that the attention mechanism calls the spatiotemporal attention optimization unit to perform early response weights and range expansion weights.

7. The enterprise data asset governance method according to claim 6, characterized in that, The output of the target governance strategy includes: Obtain the asset characteristics corresponding to the real-time data assets, and combine them with the long-term series of the data assets throughout their entire lifecycle to match the policy time characteristics and policy space characteristics associated in the governance policy mapping relationship; The matched policy temporal features and policy spatial features are input into the spatiotemporal attention optimization unit of the optimized governance model, and the early response weight and the range expansion weight are determined according to the preset weight adjustment rules. The early response weight and the range expansion weight are substituted into the fusion operation process of the attention mechanism to adjust the fusion ratio of the original governance strategy and the original resilience strategy. The early response weight is used to adjust the prediction adaptation of the strategy execution time, and the range expansion weight is used to adjust the adaptation expansion of the strategy execution range. According to the adjusted fusion ratio, the original governance strategy and the original resilience strategy are fused to generate and output the target governance strategy.

8. An enterprise data asset governance device, employing the enterprise data asset governance method as described in any one of claims 1 to 7, characterized in that, The device includes: The model building module is used to acquire historical data assets and their corresponding historical governance strategies, and to build a governance model. The governance model takes real-time data assets as input and the target governance strategy as output. The raw output module is used to input data assets into the governance model for governance analysis and output raw governance strategies; it also performs resilience analysis simultaneously and outputs raw resilience strategies. The fusion output module is used to fuse the original governance strategy and the original resilience strategy in the governance model using an attention mechanism, and output the target governance strategy.

9. A computer device for enterprise data asset governance, characterized in that, Includes at least one processor, at least one memory, and a data bus; The processor and the memory communicate with each other via the data bus; The memory stores program instructions that can be executed by the processor, which invokes the program instructions to execute the enterprise data asset governance method according to any one of claims 1 to 7.

10. A storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the enterprise data asset governance method according to any one of claims 1 to 7.

Images