Cloud-edge collaborative distributed model construction and scheduling method and device and storage medium

Abstract

The invention provides a cloud edge collaborative distributed model construction and scheduling method and device and a storage medium, and the method comprises the steps: building a distributed processing layer virtual database based on a virtualization technology according to a physical data table of each data node; on the basis of a set permission table, authorizing the virtual data table access permission of the visiting user; according to the packaged algorithm library, constructing a distributed model in combination with the standardized virtual data table; based on a virtualization technology, according to the distributed model of each data node, establishing a comprehensive processing layer virtual database; according to the method, the comprehensive model is constructed based on the distributed model of each data node, and through the method, the problem of how to construct the supervision model and perform scheduling operation based on the distributed heterogeneous data under the condition that related data of a supervised department is difficult to physically concentrate in a big data law supervision scene is solved.

Description

Background Art

[0002] Today, with the rapid development of information technology, the requirements for data processing and model construction are becoming increasingly complex and diverse. With the wide application of the Internet of Things (IoT), the fifth-generation mobile communication technology (5G), and big data technology, the generation and transmission speed of data have reached an unprecedented level. These data come from different sources, are difficult to share, and are heterogeneous and diverse, posing higher requirements for the efficiency and accuracy of data processing and model construction. With the enhancement of data privacy protection awareness and the progress of technology, federated computing, as an emerging data processing paradigm, has become a research hotspot. Federated computing allows for the construction of a global model by aggregating model updates from each participating party without exposing the original data. This approach not only improves the security and privacy protection level of data but also enables the full utilization of dispersed data resources. However, existing federated computing methods still face a series of challenges in practice.

[0003] There are multiple ways to implement federated computing. Among them, federated averaging is one of the most common federated learning algorithms, proposed by Google. In FedAvg, the central server first distributes the initial version of the global model to the clients. The clients use local data to train the model and send the updated model parameters back to the central server. The central server collects all the model updates from the clients and performs a weighted average of these updates to obtain a new global model. Federated optimization methods attempt to improve FedAvg by introducing additional optimization strategies. For example, using adaptive learning rate adjustment strategies (such as Adam or Adagrad) to accelerate the convergence speed. Improving the model performance by introducing momentum or other optimization techniques. Federated transfer learning utilizes the knowledge of pre-trained models to assist in the learning of new tasks. In FedTL, the clients can benefit from the pre-trained models of other tasks to accelerate the model training process. Federated reinforcement learning allows for collaborative learning in a multi-agent system. Through the federated learning framework, each agent can learn in its local environment and then aggregate the learned knowledge into the global model.

[0004] There are still many deficiencies in federated computing technology. For example, data heterogeneity: Most federated computing methods assume that the data is independently and identically distributed, but in reality, the data usually has different distributions and characteristics. How to effectively handle these differences remains a challenge; communication efficiency: During the federated computing process, a large amount of data transmission will lead to high communication costs. Especially in edge computing scenarios where the network bandwidth is limited, how to reduce the communication overhead is a key issue; resource allocation: In a distributed environment, reasonably allocating computing resources and tasks to maximize the overall efficiency of the system is a difficult point; security and privacy protection: Although federated computing aims to protect data privacy, the risk of data leakage still needs to be considered in actual operations, especially during the model aggregation stage.

[0005] It can be seen from this that in the scenario of big data legal supervision, when it is difficult to physically centralize the relevant data of the supervised department, how to build a supervision model based on distributed heterogeneous data and schedule its operation has become one of the existing technical problems to be solved urgently. Summary of the Invention

[0006] The present invention provides a cloud-edge collaborative distributed model construction and scheduling method, device and storage medium, which are used to build a supervision model based on distributed heterogeneous data and schedule its operation when it is difficult to physically centralize the relevant data of the supervised department.

[0007] In the first aspect, a cloud-edge collaborative distributed model construction and scheduling method is provided, including:

[0008] Based on virtualization technology, establish a virtual database for the distributed processing layer according to the physical data tables of each data node;

[0009] Based on the set permission table, authorize the virtual data table access permission for the visiting users;

[0010] According to the encapsulated algorithm library, combine with the standardized virtual data tables to build a distributed model;

[0011] Based on virtualization technology, establish a virtual database for the comprehensive processing layer according to the distributed models of each data node;

[0012] Based on the distributed models of each data node, build a comprehensive model.

[0013] In one implementation, based on virtualization technology, establishing a virtual database for the distributed processing layer according to the physical data tables of each data node specifically includes:

[0014] Based on virtualization technology, map the physical data tables of each data node to virtual data tables. The virtual data tables are virtual database tables after standardizing the node physical data according to the agreed data structure standard. Use data virtualization technology to shield the technical differences of data sources, and enable the model to perform cross-data source node operations and be quickly transplanted through data structure standardization.

[0015] In one implementation, based on the set permission table, authorizing the virtual data table access permission for the visiting users specifically includes:

[0016] The distributed model accesses the data sources of each node through data users, and controls the access permission to the source data by authorizing the data users. If the data user can be queried and obtained through the set permission table, it has the access permission. Authorize the visiting users to access the source data, and at the same time map the data source to the virtual table.

[0017] In one embodiment, the distributed model is constructed by combining each encapsulated algorithm library with a standardized virtual data table for modeling; the construction of the distributed model introduces an existing model, which has clear inputs and outputs and provides WebService or Restful interfaces.

[0018] In one embodiment, the virtual database in the integrated processing layer includes an integrated processing layer virtual data table, and an integrated model is constructed in the form of a data table, and the calculation is performed using the combined query calculation ability of the sql language.

[0019] In one embodiment, the integrated model provides integrated services to the application through WebService or RestFul interfaces; the integrated model includes an integrated processing layer virtual database, and according to the needs of the model, the execution results are cached and the model result data is accumulated.

[0020] In a second aspect, a cloud-edge collaborative distributed model construction and scheduling device is provided, including:

[0021] A first database establishment module, configured to establish a distributed processing layer virtual database based on virtualization technology according to the physical data tables of each data node;

[0022] An authorization module, configured to authorize the virtual data table access rights of visiting users based on a set permission table;

[0023] A distributed model construction module, configured to construct a distributed model by combining an encapsulated algorithm library with a standardized virtual data table;

[0024] A second database establishment module, configured to establish an integrated processing layer virtual database based on virtualization technology according to the distributed models of each data node;

[0025] An integrated model construction module, configured to construct an integrated model based on the distributed models of each data node.

[0026] In one embodiment, the distributed model is constructed by combining each encapsulated algorithm library with a standardized virtual data table for modeling; the construction of the distributed model introduces an existing model, which has clear inputs and outputs and provides WebService or Restful interfaces.

[0027] In a third aspect, a computing device is provided, including at least one processor and at least one memory, wherein the memory stores a computer program, and the memory is configured to read the computer program in the memory and execute any step of the cloud-edge collaborative distributed model construction and scheduling method provided in the first aspect above.

[0028] Fourthly, a computer-readable storage medium is provided. The computer-readable storage medium stores computer-executable instructions for causing a computer to execute any step of the cloud-edge collaborative distributed model construction and scheduling method provided in the first aspect above.

[0029] A cloud-edge collaborative distributed model construction and scheduling method, device and storage medium provided by an embodiment of the present invention. The method includes: based on virtualization technology, establishing a virtual database for the distributed processing layer according to the physical data tables of each data node; authorizing the virtual data table access rights of visiting users based on a set permission table; constructing a distributed model by combining a standardized virtual data table according to a packaged algorithm library; based on virtualization technology, establishing a virtual database for the comprehensive processing layer according to the distributed models of each data node; constructing a comprehensive model based on the distributed models of each data node. Through the above method, the problem of how to construct a supervision model based on distributed heterogeneous data and schedule its operation in the case where it is difficult to physically centralize the relevant data of the supervised department in the big data legal supervision scenario is solved.

[0030] Other features and advantages of the present invention will be described in the following specification, and part of them will become obvious from the specification or be understood by implementing the present invention. The objectives and other advantages of the present invention can be achieved and obtained by the structures specifically pointed out in the written specification, claims and their drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] The drawings described herein are used to provide a further understanding of the present invention, and constitute a part of the present invention. The illustrative embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute an improper limitation to the present invention. In the drawings:

[0032] Figure 1 is a flowchart of the steps of the cloud-edge collaborative distributed model construction and scheduling method according to an embodiment of the present invention;

[0033] Figure 2 is a structural relationship diagram of the processing layer according to an embodiment of the present invention;

[0034] Figure 3 is a structural diagram of the cloud-edge collaborative distributed model construction and scheduling device according to an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0035] In order to construct a supervision model based on distributed heterogeneous data and schedule its operation in the case where it is difficult to physically centralize the relevant data of the supervised department in the big data legal supervision scenario, a cloud-edge collaborative distributed model construction and scheduling method, device and storage medium are provided.

[0036] The preferred embodiments of the present invention will be described below in conjunction with the accompanying drawings of the specification. It should be understood that the preferred embodiments described herein are only for the purpose of illustrating and explaining the present invention, and are not used to limit the present invention. And without conflict, the embodiments in the present invention and the features in the embodiments can be combined with each other.

[0037] It should be noted that the terms "first", "second", etc. in the specification, claims and the above-mentioned drawings of the embodiments of the present invention are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that such data used can be interchanged under appropriate circumstances, so that the embodiments described herein can be implemented in an order other than that illustrated or described herein.

[0038] As Figure 1 shown, the embodiment provides a cloud-edge collaborative distributed model construction and scheduling method, and the specific implementation steps include:

[0039] S11. Based on virtualization technology, establish a virtual database for the distributed processing layer according to the physical data tables of each data node.

[0040] In specific implementation, the node virtual data table is based on virtualization technology to map the physical data tables of each data node into virtual data tables. The virtual data table is a (virtual) database table after standardizing the node physical data according to the agreed data structure standard. The purpose is to use data virtualization technology to shield the technical differences of data sources, and through data structure standardization, the model has the computing ability and fast transplantation ability across data source nodes.

[0041] S12. Based on the set permission table, authorize the virtual data table access rights of the visiting users.

[0042] In specific implementation, the distributed model accesses the data sources of each node through data users, and controls the access rights to the source data by authorizing the data users. If the data user can be queried and obtained through the set permission table, it has the access right, authorizes the visiting user to access the source data, and at the same time maps the data source to the virtual table.

[0043] In one implementation manner, the access rights to the source data are controlled by authorizing the data users, that is, the access of each data user to the data source requires the approval of the node administrator. And the mapping of the data source to the virtual table is completed by the node administrator.

[0044] S13. According to the encapsulated algorithm library, construct a distributed model in combination with the standardized virtual data table.

[0045] In specific implementation, the process of building a model is a process of modeling based on a series of encapsulated algorithm libraries in combination with a standardized virtual data table. The construction of the model can also introduce existing models, requiring that the existing models have clear inputs and outputs and provide WebService or Restful interfaces. After the model is built, in order to serve the application, a standardized interface needs to be provided. The process of forming this standardized interface is the release of the service, and the model service interface is of the WebServcie type.

[0046] S14. Based on virtualization technology, establish a virtual database for the integrated processing layer according to the distributed models of each data node.

[0047] In specific implementation, the virtual data table of the integrated processing layer is based on virtualization technology to map the outputs of the distributed models of each data node into virtual data tables. The purpose is to provide the greatest flexibility for the construction of the integrated model in the form of data tables and make full use of the computing capabilities such as the combined query of the sql language.

[0048] S15. Build an integrated model based on the distributed models of each data node.

[0049] In specific implementation, based on the outputs of the distributed models of each node, apply various algorithms to establish an integrated model, and provide integrated services for applications through interfaces such as WebService or RestFul. The virtual database of the integrated processing layer has caching capabilities. According to the needs of the model, the model can control the caching of the model execution results and accumulate the model result data.

[0050] For easy understanding, the following gives a specific example:

[0051] 1. Establish model input specifications.

[0052] According to the needs of prosecutors to review the data of administrative penalty cases, establish a unified data model for administrative penalty cases. The case data model includes two parts: basic case information and case elements. The basic information fields include structured data fields such as case number, administrative division number, case entry time, administrative law enforcement unit number, administrative matter number, and administrative law enforcement unit name. The case elements include case number, element name, and element value. (The so-called elements are the rule elements used to judge whether a case involves a crime.)

[0053] 2. System deployment.

[0054] An intelligent collaborative review system for intellectual property rights is deployed in the provincial procuratorate, which includes an integrated model part and a cloud virtual data platform. The distributed model and the edge virtual data platform are deployed in administrative law enforcement units (such as the Market Supervision Bureau). As Figure 2 shown.

[0055] 3. Edge - side data preparation.

[0056] Build an edge - side data management system in the front - end computer of relevant provincial administrative law - enforcement units (such as the Market Supervision Bureau). Use a relational database to manage administrative penalty case data. The edge - side data management system synchronizes incremental data with the source data at 24:00 every day, and the synchronization program is controlled by the administrative law - enforcement unit. Administrative penalty case data includes two parts: basic case information and case elements. The basic information fields include structured data fields such as case number, administrative division code, case entry time, administrative law - enforcement unit code, administrative matter code, and administrative law - enforcement unit name. Case elements include case number, element name, and element value. Due to different database technologies of each administrative law - enforcement unit, the definition of field names is also different. Edge - side data maps physical database tables into virtual tables in the standard format required for model input through the virtual data platform deployed on the edge - side.

[0057] 4. Distributed model construction and deployment.

[0058] The distributed model of this application is a model developed based on Python. Its function is to determine whether a case involves criminal offenses based on administrative penalty case elements and provide a Restful interface. The model is deployed in each administrative law - enforcement unit and connects to the standardized virtual data tables of this node.

[0059] 5. Comprehensive processing module

[0060] The comprehensive processing module mainly accesses the output clues of the distributed models of each node based on virtual tables, and summarizes the outputs of each node based on the Sql calculation function of the virtual data platform to form a comprehensive clue discovery model, which is published as a WebService service through the platform.

[0061] 6. Clue management function

[0062] Based on the WebService service of the comprehensive clue discovery model, provide a clue management function for prosecutors. First, display the list of criminal - involved administrative case clues within the jurisdiction of the prosecutor according to the prosecutor's authority. The prosecutor can view the details of the clues and learn more information such as the basis for judging the case clues. Based on this, the prosecutor can initiate the review of more case files and then determine whether to initiate supervision of the case.

[0063] The above - mentioned method helps to promote the smoother development of big - data legal supervision. In the process of big - data legal supervision, there are problems such as many data sources, complex structures, and difficult sharing. For example, intellectual property case data is usually scattered in different business domains, such as the Intellectual Property Office (administrative law - enforcement), courts, procuratorates, etc. Due to the limitations of management boundaries, it is difficult to physically centralize data into a unified data center (data lake / database). Therefore, traditional business model construction methods are not applicable.

[0064] The embodiment provides a cloud-edge collaborative distributed model construction and scheduling method, which combines federated computing and virtualization technologies. Without the original data leaving the domain, the distributed model calculates based on the original data and obtains preliminary results for joint computing. Then, the comprehensive model performs the second comprehensive calculation. Different from existing federated computing technologies, in this process, the present invention uses data virtualization technology to first convert the original data into a standardized virtual table and then provide it to the distributed model, making the entire model have standardized inputs and thus being more portable. The method of using the data source node administrator to approve access rights enhances the control ability of the data ownership party over the data. By mapping the output of the distributed model to a virtual table, the SQL computing power brought by data virtualization technology is fully utilized, improving the construction efficiency of the comprehensive model.

[0065] Based on the same technical concept, the embodiment of the present application also provides a cloud-edge collaborative distributed model construction and scheduling device. Since the principle of the device for solving problems is similar to that of the cloud-edge collaborative distributed model construction and scheduling method, the implementation of the device can refer to the implementation of the method, and the repeated parts will not be elaborated.

[0066] As Figure 3 shown, the embodiment provides a cloud-edge collaborative distributed model construction and scheduling device, including:

[0067] The first database establishment module 31 is used to establish a distributed processing layer virtual database based on virtualization technology according to the physical data tables of each data node;

[0068] The authorization module 32 is used to authorize the virtual data table access rights of visiting users based on the set permission table;

[0069] The distributed model construction module 33 is used to construct a distributed model according to the encapsulated algorithm library in combination with the standardized virtual data table;

[0070] The second database establishment module 34 is used to establish a comprehensive processing layer virtual database based on virtualization technology according to the distributed models of each data node;

[0071] The comprehensive model construction module 35 is used to construct a comprehensive model based on the distributed models of each data node.

[0072] In one implementation, the construction of the distributed model is based on each encapsulated algorithm library in combination with the standardized virtual data table for modeling; the construction of the distributed model introduces an existing model, and the existing model has clear inputs and outputs and provides WebService or Restful interfaces.

[0073] For the sake of convenient description, the above - mentioned parts are divided into respective modules (or units) according to functional modules for separate description. Of course, when implementing the present invention, the functions of the respective modules (or units) can be realized in one or more software or hardware.

[0074] After introducing the method and device for building and scheduling the cloud - edge collaborative distributed model according to the exemplary embodiments of the present invention, next, a computing device according to another exemplary embodiment of the present invention will be introduced.

[0075] Those skilled in the art of the relevant technical field can understand that various aspects of the present invention can be implemented as a system, a method, or a program product. Therefore, various aspects of the present invention can be specifically implemented in the following forms, namely: a complete hardware implementation, a complete software implementation (including firmware, microcode, etc.), or an implementation combining hardware and software aspects, which can be collectively referred to as "circuits", "modules", or "systems" here. In some possible embodiments, the computing device according to the present invention may include at least one processor and at least one memory. Among them, the memory stores program code. When the program code is executed by the processor, the processor is caused to execute the steps in the method for building and scheduling the cloud - edge collaborative distributed model according to various exemplary embodiments of the present invention described above in this specification. For example, the processor can execute steps such as Figure 1 step S11 shown in, based on virtualization technology, establish a virtual database for the distributed processing layer according to the physical data tables of each data node; and step S12, based on the set permission table, authorize the virtual data table access rights of visiting users; and step S13, according to the encapsulated algorithm library, construct a distributed model in combination with the standardized virtual data table; step S14, based on virtualization technology, establish a virtual database for the comprehensive processing layer according to the distributed models of each data node; step S15, construct a comprehensive model based on the distributed models of each data node.

[0076] In some possible embodiments, various aspects of the method for building and scheduling the cloud - edge collaborative distributed model provided by the present invention can also be implemented in the form of a program product, which includes program code. When the program product runs on a computer device, the program code is used to cause the computer device to execute the steps in the method for building and scheduling the cloud - edge collaborative distributed model according to various exemplary embodiments of the present invention described above in this specification. For example, the computer device can execute steps such as Figure 1Step S11 shown in [description] is to establish a virtual database for the distributed processing layer based on virtualization technology according to the physical data tables of each data node; step S12 is to authorize the virtual data table access rights of visiting users based on the set permission table; step S13 is to construct a distributed model by combining the encapsulated algorithm library with the standardized virtual data table; step S14 is to establish a virtual database for the comprehensive processing layer based on virtualization technology according to the distributed models of each data node; step S15 is to construct a comprehensive model based on the distributed models of each data node.

[0077] The program product can adopt any combination of one or more readable media. The readable media can be a readable signal medium or a readable storage medium. The readable storage medium can be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination of the above. More specific examples of the readable storage medium (a non-exhaustive list) include: an electrical connection having one or more wires, a portable disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. The program product for the construction and scheduling of the distributed model for cloud-edge collaboration in the embodiments of the present invention can adopt a portable compact disk read-only memory (CD-ROM) and include program code, and can run on a computing device. However, the program product of the present invention is not limited to this. In the present invention, the readable storage medium can be any tangible medium that contains or stores a program, and this program can be used by or in combination with an instruction execution system, apparatus, or device.

[0078] Those skilled in the art should understand that the embodiments of the present invention can be provided as a method, a system, or a computer program product. Therefore, the present invention can take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present invention can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk memories, CD-ROMs, optical memories, etc.) containing computer-usable program code.

[0079] The present invention is described with reference to the flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to the embodiments of the present invention. It should be understood that each flow and / or block in the flowcharts and / or block diagrams, and the combination of flows and / or blocks in the flowcharts and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing devices to generate instructions for implementing the processes Figure 1One or more processes and / or blocks Figure 1 Means for specifying functions in one or more blocks

[0080] These computer program instructions can also be stored in a computer-readable memory that directs a computer or other programmable data processing device to operate in a particular manner, such that the instructions stored in the computer-readable memory produce a manufacture including an instruction means that implements the functions specified in the process Figure 1 One or more processes and / or blocks Figure 1 In one or more blocks

[0081] These computer program instructions can also be loaded onto a computer or other programmable data processing device, such that a series of operational steps are performed on the computer or other programmable device to produce a computer-implemented process, whereby the instructions executed on the computer or other programmable device provide for implementing the steps specified in the process Figure 1 One or more processes and / or blocks Figure 1 In one or more blocks. Although the preferred embodiments of the present invention have been described, additional changes and modifications can be made by those skilled in the art once they learn of the basic inventive concept. Therefore, the appended claims are intended to be construed to include the preferred embodiments as well as all changes and modifications falling within the scope of the present invention

[0082] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalent technologies, the present invention is also intended to include these modifications and variations

Claims

1. A distributed model construction and scheduling method for cloud-edge collaboration, characterized in that: include: Based on virtualization technology, a distributed processing layer virtual database is established according to the physical data tables of each data node; Authorize the visiting user's virtual data table access rights based on the set permission table; Build a distributed model based on the encapsulated algorithm library and standardized virtual data tables; Based on virtualization technology, a comprehensive processing layer virtual database is established according to the distributed model of each data node; Build a comprehensive model based on the distributed model of each data node.

2. The method according to claim 1, characterized in that Based on virtualization technology, a distributed processing layer virtual database is established according to the physical data tables of each data node, including: Based on virtualization technology, the physical data table of each data node is mapped to a virtual data table. The virtual data table is a virtual database table after the physical data of the node is standardized according to the agreed data structure standard. Data virtualization technology is used to shield the technical differences of data source, and the model can be calculated and quickly transplanted across data source nodes through data structure standardization.

3. The method according to claim 2, characterized in that Based on the set permission table, authorize the visiting user's virtual data table access rights, including: The distributed model accesses the data source of each node through data users, and controls the access rights to the source data by authorizing the data users. If the data user can obtain the access rights through the set permission table, the access rights are granted, and the access rights to the source data are authorized, and the data source is mapped to the virtual table.

4. The method according to claim 3, characterized in that The construction of the distributed model is based on modeling of each packaged algorithm library combined with a standardized virtual data table; the construction of the distributed model introduces an existing model, and the existing model has clear input and output and provides a WebService or Restful interface.

5. The method according to claim 4, characterized in that The comprehensive processing layer virtual database includes a comprehensive processing layer virtual data table, a comprehensive model is constructed in the form of a data table, and calculations are performed using the joint query calculation capability of the SQL language.

6. The method according to claim 5, characterized in that The comprehensive model provides comprehensive services to applications through WebService or RestFul interface; the comprehensive model includes a comprehensive processing layer virtual database, which caches execution results and accumulates model result data according to model requirements.

7. A distributed model construction and scheduling device for cloud-edge collaboration, characterized in that: include: The first database establishment module is used to establish a distributed processing layer virtual database based on the physical data table of each data node based on virtualization technology; The authorization module is used to authorize the virtual data table access rights of the visiting user based on the set permission table; A distributed model building module is used to build a distributed model based on the encapsulated algorithm library and the standardized virtual data table; The second database establishment module is used to establish a comprehensive processing layer virtual database based on the virtualization technology and the distributed model of each data node; The comprehensive model building module is used to build a comprehensive model based on the distributed model of each data node.

8. The device according to claim 7, characterized in that The construction of the distributed model is based on modeling of each packaged algorithm library combined with a standardized virtual data table; the construction of the distributed model introduces an existing model, and the existing model has clear input and output and provides a WebService or Restful interface.

9. A computing device, characterized in that: The method comprises at least one processor and at least one memory, wherein the memory stores a computer program, and the processor is used to read the computer program in the memory and execute the method according to any one of claims 1 to 6.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer-executable instructions, and the computer-executable instructions are used to enable a computer to execute the method of any one of claims 1 to 6.

Images