A metadata information generation method, related apparatus, device, and storage medium

Abstract

The application discloses a metadata information generation method, comprising: obtaining existing metadata of a target metadata table, the existing metadata being metadata that already exists in the target metadata table, and the existing metadata comprising K metadata; obtaining K metadata information according to the existing metadata; performing feature processing on the K metadata information to obtain K metadata features; and obtaining target metadata information for the target metadata table based on the K metadata features through a metadata prediction model, wherein the target metadata information is metadata information corresponding to missing metadata, and the missing metadata is metadata that does not exist in the target metadata table. The application also provides related devices, equipment and storage media. The application can save labor costs and improve work efficiency on one hand, and can avoid the influence of subjective factors, thereby improving information accuracy on the other hand.

Description

Technical Field

[0001] This application relates to the field of computers, and in particular to a method, apparatus, device and storage medium for generating metadata information. Background Technology

[0002] As business grows, the importance of metadata on the data side is increasing daily. Metadata is data that describes other data, or structured data used to provide information about a resource. Metadata describes information resources or data objects, and its purpose is to identify resources, evaluate resources, track changes in resources during use, achieve simple and efficient management of large amounts of networked data, and realize effective management of resources in use.

[0003] However, missing metadata information is inevitable. For example, a data table may lack table description information or field information. Therefore, technical personnel usually need to supplement this missing metadata information through the data platform to obtain a data table with more complete metadata information.

[0004] However, supplementing missing metadata often requires a significant investment of time and effort from technical personnel, resulting in high labor costs, low efficiency, and potential disruption to normal business operations. Furthermore, subjective factors can lead to inaccuracies in the manually supplemented metadata. Summary of the Invention

[0005] This application provides a method, related apparatus, device, and storage medium for generating metadata information. On the one hand, it eliminates the need for technicians to spend a significant amount of time and effort manually supplementing missing metadata information, thereby saving labor costs and improving work efficiency. On the other hand, by using models to predict missing metadata information, the influence of subjective factors can be avoided, thus improving information accuracy.

[0006] In view of this, this application provides a method for generating metadata information, including:

[0007] Retrieve existing metadata from the target metadata table. Existing metadata refers to metadata that already exists in the target metadata table, and the existing metadata includes K metadata items, where K is an integer greater than or equal to 1.

[0008] Obtain K metadata information from the existing metadata, where there is a one-to-one correspondence between the metadata information and the metadata;

[0009] Feature processing is performed on K metadata information to obtain K metadata features, where there is a one-to-one correspondence between the metadata features and the metadata information;

[0010] Based on K metadata features, target metadata information for the target metadata table is obtained through a metadata prediction model. The target metadata information is the metadata information corresponding to the missing metadata, and the missing metadata is metadata that does not exist in the target metadata table.

[0011] This application also provides a metadata information generation apparatus, comprising:

[0012] The acquisition module is used to acquire existing metadata of the target metadata table. The existing metadata refers to metadata that already exists in the target metadata table, and the existing metadata includes K metadata items, where K is an integer greater than or equal to 1.

[0013] The acquisition module is also used to acquire K metadata information based on existing metadata, wherein the metadata information has a one-to-one correspondence with the metadata;

[0014] The processing module is used to perform feature processing on K metadata information to obtain K metadata features, wherein the metadata features have a one-to-one correspondence with the metadata information;

[0015] The acquisition module is also used to obtain target metadata information for the target metadata table based on K metadata features through a metadata prediction model. The target metadata information is the metadata information corresponding to the missing metadata, and the missing metadata is metadata that does not exist in the target metadata table.

[0016] In one possible design, in another implementation of another aspect of the embodiments of this application, the K metadata information includes table name text information, wherein the table name text information belongs to a preset language;

[0017] The processing module is specifically used to perform word segmentation on the table name text information to obtain a table name text sequence, wherein the table name text sequence includes M words, where M is an integer greater than or equal to 1;

[0018] Each word in the table name text sequence is preprocessed to obtain M word vector embeddings, where the word vector embeddings have a one-to-one correspondence with the words, and the M word vector embeddings belong to the metadata features of the table name text information;

[0019] The acquisition module is specifically used to obtain the semantic information vector corresponding to the table name text sequence based on the embedding of M word vectors and through the encoder included in the metadata prediction model;

[0020] Based on semantic information vectors, table description information is obtained through the decoder included in the metadata prediction model, where the table description information belongs to the target metadata information.

[0021] In one possible design, in another implementation of another aspect of the embodiments of this application,

[0022] The acquisition module is specifically used to retrieve the table name text information corresponding to the table name from the target metadata table if the existing metadata includes the table name.

[0023] or,

[0024] The acquisition module is specifically used to acquire the associated metadata table based on the data lineage included in the existing metadata if the existing metadata does not include the data table name. The associated metadata table is either the upstream metadata table or the downstream metadata table.

[0025] Retrieve the text information of the table name corresponding to the data table name from the associated metadata table.

[0026] In one possible design, in another implementation of another aspect of the embodiments of this application, the existing metadata includes at least one of the following: number of downstream dependent tasks, number of upstream dependent tasks, access frequency, online business service status, number of online services, number of fields, number of data rows, and storage size.

[0027] The processing module is specifically used to perform data binning on the metadata information of the downstream dependent task number if the existing metadata includes the number of downstream dependent tasks, so as to obtain the metadata features of the downstream dependent task number, wherein the metadata features of the downstream dependent task number belong to K metadata features.

[0028] If the existing metadata includes the number of upstream dependent tasks, then the metadata information of the number of upstream dependent tasks is binned to obtain the metadata features of the number of upstream dependent tasks. The metadata features of the number of upstream dependent tasks belong to K metadata features.

[0029] If the existing metadata includes access frequency, then the access frequency is binned to obtain the metadata features of the access frequency, where the metadata features of the access frequency belong to K metadata features.

[0030] If the existing metadata includes information about online business services, then the online business service information is one-hot encoded to obtain the metadata features of the online business service information. The metadata features of the online business service information belong to K metadata features.

[0031] If the existing metadata includes the number of online services, feature scaling is performed on the number of online services to obtain the metadata features of the number of online services. The metadata features of the number of online services belong to K metadata features.

[0032] If the existing metadata includes the number of fields, perform feature scaling on the number of fields to obtain the metadata feature of the number of fields, where the metadata feature of the number of fields belongs to K metadata features;

[0033] If the existing metadata includes the number of data rows, perform feature scaling on the number of data rows to obtain the metadata features of the number of data rows, where the metadata features of the number of data rows belong to K metadata features;

[0034] If the existing metadata includes storage size, feature scaling is performed on the storage size to obtain the metadata features of the storage size, where the metadata features of the storage size belong to K metadata features.

[0035] In one possible design, in another implementation of another aspect of the embodiments of this application, the existing metadata also includes at least one of business importance level and business sensitivity;

[0036] The processing module is specifically used to perform one-hot encoding on the metadata information corresponding to the business importance level if the existing metadata includes the business importance level, so as to obtain the metadata features of the business importance level. The metadata features of the business importance level belong to K metadata features.

[0037] If the existing metadata includes business-sensitive cases, then the metadata information corresponding to the business-sensitive cases is processed by one-hot encoding to obtain the metadata features of the business-sensitive cases. The metadata features of the business-sensitive cases belong to K metadata features.

[0038] The acquisition module is specifically used to obtain the business type information of the target metadata table based on K metadata features through a metadata prediction model.

[0039] In one possible design, in another implementation of another aspect of the embodiments of this application,

[0040] The acquisition module is specifically used to obtain the business type information of the target metadata table based on K metadata features through a decision tree model. The decision tree model is a metadata prediction model, which is a network based on the relationship between nodes, with each node corresponding to a different metadata feature.

[0041] or,

[0042] The acquisition module is specifically used to obtain the business type information of the target metadata table based on K metadata features through a random forest model. The random forest model is a metadata prediction model, which includes at least two decision trees, each of which is used to predict a classification result.

[0043] or,

[0044] The acquisition module is specifically used to obtain the business type of the target metadata table based on K metadata features through an extensible boosting tree model;

[0045] or,

[0046] The acquisition module is specifically used to obtain the first feature vector based on K metadata features and through the input layer included in the metadata prediction model;

[0047] Based on the first feature vector, a second feature vector is obtained through at least one hidden layer included in the metadata prediction model;

[0048] Based on the second feature vector, the probability distribution is obtained through the output layer of the metadata prediction model;

[0049] The business type information of the target metadata table is determined based on the probability distribution.

[0050] In one possible design, in another implementation of another aspect of the embodiments of this application, the existing metadata also includes at least one of the business type and business sensitivity conditions;

[0051] The processing module is specifically used to perform one-hot encoding on the metadata information corresponding to the business type if the existing metadata includes the business type, so as to obtain the metadata features of the business type. The metadata features of the business type belong to K metadata features.

[0052] If the existing metadata includes business-sensitive cases, then the metadata information corresponding to the business-sensitive cases is processed by one-hot encoding to obtain the metadata features of the business-sensitive cases. The metadata features of the business-sensitive cases belong to K metadata features.

[0053] The acquisition module is specifically used to obtain business importance level information for a target metadata table based on K metadata features and through a metadata prediction model.

[0054] In one possible design, in another implementation of another aspect of the embodiments of this application,

[0055] The acquisition module is specifically used to obtain business importance level information for the target metadata table based on K metadata features through a decision tree model. The decision tree model is a metadata prediction model, which is a network based on the relationship between nodes, with each node corresponding to a different metadata feature.

[0056] or,

[0057] The acquisition module is specifically used to acquire business importance level information for the target metadata table based on K metadata features through a random forest model. The random forest model is a metadata prediction model, which includes at least two decision trees, each of which is used to predict a classification result.

[0058] or,

[0059] The acquisition module is specifically used to obtain the business importance level of a target metadata table based on K metadata features through an extensible boosting tree model;

[0060] or,

[0061] The acquisition module is specifically used to obtain the first feature vector based on K metadata features and through the input layer included in the metadata prediction model;

[0062] Based on the first feature vector, a second feature vector is obtained through at least one hidden layer included in the metadata prediction model;

[0063] Based on the second feature vector, the probability distribution is obtained through the output layer of the metadata prediction model;

[0064] The business importance level information of the target metadata table is determined based on the probability distribution.

[0065] In one possible design, in another implementation of another aspect of the embodiments of this application, the existing metadata also includes at least one of the business type and the business importance level;

[0066] The processing module is specifically used to perform one-hot encoding on the metadata information corresponding to the business type if the existing metadata includes the business type, so as to obtain the metadata features of the business type. The metadata features of the business type belong to K metadata features.

[0067] If the existing metadata includes the business importance level, then the metadata information corresponding to the business importance level is one-hot encoded to obtain the metadata features of the business importance level. The metadata features of the business importance level belong to K metadata features.

[0068] The acquisition module is specifically used to obtain business-sensitive information about the target metadata table based on K metadata features and through a metadata prediction model.

[0069] In one possible design, in another implementation of another aspect of the embodiments of this application,

[0070] The acquisition module is specifically used to acquire business-sensitive information about the target metadata table based on K metadata features through a decision tree model. The decision tree model is a metadata prediction model, which is a network based on the relationship between nodes, with each node corresponding to a different metadata feature.

[0071] or,

[0072] The acquisition module is specifically used to acquire business-sensitive information about the target metadata table based on K metadata features using a random forest model. The random forest model is a metadata prediction model, which includes at least two decision trees, each of which is used to predict a classification result.

[0073] or,

[0074] The acquisition module is specifically used to obtain business-sensitive information about the target metadata table based on K metadata features through an extensible boosting tree model.

[0075] or,

[0076] The acquisition module is specifically used to obtain the first feature vector based on K metadata features and through the input layer included in the metadata prediction model;

[0077] Based on the first feature vector, a second feature vector is obtained through at least one hidden layer included in the metadata prediction model;

[0078] Based on the second feature vector, the probability distribution is obtained through the output layer of the metadata prediction model;

[0079] The business sensitivity information of the target metadata table is determined based on the probability distribution.

[0080] In one possible design, in another implementation of another aspect of the embodiments of this application, the metadata information generation device further includes a generation module and a recording module;

[0081] The generation module is used to generate K metadata records based on the K metadata features and the K metadata information after the acquisition module obtains the target metadata information for the target metadata table through the metadata prediction model.

[0082] The generation module is also used to generate target metadata records based on the missing metadata and target metadata information;

[0083] The recording module is used to record K metadata records and target metadata records to the blockchain for synchronous storage. The blockchain is used to provide metadata information of the data table for at least two node devices.

[0084] In one possible design, in another implementation of another aspect of the embodiments of this application, the metadata information generation device further includes a display module;

[0085] The display module is used to obtain the target metadata information of the target metadata table based on K metadata features through a metadata prediction model, and then display the target metadata information of the target metadata table.

[0086] or,

[0087] The target metadata information of the target metadata table is sent to the terminal device so that the terminal device can display the target metadata information of the target metadata table.

[0088] This application also provides a computer device, including: a memory, a processor, and a bus system;

[0089] The memory is used to store programs;

[0090] The processor is used to execute programs in memory, and the processor is used to execute the methods mentioned above according to the instructions in the program code;

[0091] Bus systems are used to connect memory and processor to enable communication between them.

[0092] Another aspect of this application provides a computer-readable storage medium storing instructions that, when executed on a computer, cause the computer to perform the methods described above.

[0093] Another aspect of this application provides a computer program product or computer program including computer instructions stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium and executes the computer instructions, causing the computer device to perform the methods provided in the above aspects.

[0094] As can be seen from the above technical solutions, the embodiments of this application have the following advantages:

[0095] This application provides a method for generating metadata information. First, existing metadata of a target metadata table is obtained. This existing metadata refers to metadata that already exists in the target metadata table. Then, K metadata information pieces are obtained based on the existing metadata. These K metadata information pieces are then subjected to feature processing to obtain K metadata features. Finally, based on these K metadata features, target metadata information for the target metadata table can be obtained through a metadata prediction model. Here, target metadata information refers to the metadata information corresponding to missing metadata, which is metadata that does not exist in the target metadata table. Through this method, based on existing metadata information, a trained model can predict missing metadata information, thereby achieving refined operation and management of data assets. On the one hand, it eliminates the need for technical personnel to spend a significant amount of time and effort manually supplementing missing metadata information, thus saving labor costs and improving work efficiency. On the other hand, predicting missing metadata information through a model avoids the influence of subjective factors, thereby improving information accuracy. Attached Figure Description

[0096] Figure 1 This is a schematic diagram of the architecture of the metadata management system in an embodiment of this application;

[0097] Figure 2 This is a schematic diagram of the target metadata information generation process in an embodiment of this application;

[0098] Figure 3 This is a flowchart illustrating the metadata information generation method in an embodiment of this application.

[0099] Figure 4 This is a schematic diagram illustrating the acquisition of table description information based on a metadata prediction model in an embodiment of this application;

[0100] Figure 5 This is a schematic diagram illustrating the acquisition of table name text information based on data lineage in an embodiment of this application;

[0101] Figure 6 This is a schematic diagram illustrating the use of a confusion matrix to evaluate a metadata prediction model in an embodiment of this application.

[0102] Figure 7 This is a schematic diagram of an optional structure of a distributed system applied to a blockchain system in an embodiment of this application;

[0103] Figure 8 This is an optional schematic diagram of the block structure in an embodiment of this application;

[0104] Figure 9 This is a schematic diagram of the operation interface for automatically supplementing target metadata information in the embodiments of this application;

[0105] Figure 10 This is a schematic diagram of a metadata information generation device in an embodiment of this application;

[0106] Figure 11 This is a schematic diagram of the structure of a terminal device in an embodiment of this application;

[0107] Figure 12 This is a schematic diagram of the server structure in an embodiment of this application. Detailed Implementation

[0108] This application provides a method, related apparatus, device, and storage medium for generating metadata information. On the one hand, it eliminates the need for technicians to spend a significant amount of time and effort manually supplementing missing metadata information, thereby saving labor costs and improving work efficiency. On the other hand, by using models to predict missing metadata information, the influence of subjective factors can be avoided, thus improving information accuracy.

[0109] The terms “first,” “second,” “third,” “fourth,” etc. (if present) in the specification, claims, and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a particular order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this application described herein can be implemented, for example, in orders other than those illustrated or described herein. Furthermore, the terms “comprising” and “corresponding to,” and any variations thereof, are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0110] Metadata is data used to describe data. It can be understood as data with a broader scope than data in the general sense. It not only represents information such as the type, name, and value of data, but it can also provide contextual descriptions of the data, such as the domain to which the data belongs, the range of values, the relationships between data, business rules, and the source of the data. In data application systems, metadata generally refers to data that describes data concepts, relationships between data, and data processing rules. Among these, domain semantics and knowledge also fall under the scope of metadata.

[0111] Metadata is beneficial for inventorying data assets and analyzing the relationships between data. Building a metadata management platform lays the foundation for subsequent work, enabling centralized, automated, and version-based management, thereby maximizing the value of data. Metadata management, as information describing data, helps businesses better understand and apply data through its capabilities of collection, storage, retrieval, sharing, and openness. It is also the cornerstone of data asset management, encompassing data dictionaries, data lineage, business metrics, and operational metrics. However, considering that the lack of some metadata information may limit the management of data assets, this application provides a method for generating metadata information. This method can predict missing metadata information based on the collection of existing metadata information, thereby helping businesses better manage data assets, quickly locate business areas and important core data assets, and ultimately improve data efficiency.

[0112] To better address the problem of missing metadata information, this application proposes a method for generating metadata information, which is applied to... Figure 1 The metadata management system shown in the figure includes a server and terminal devices, with the metadata management platform deployed on the terminal devices. The server involved in this application can be a standalone physical server, a server cluster or distributed system composed of multiple physical servers, or a cloud server providing basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, content delivery networks (CDNs), and big data and artificial intelligence platforms. The terminal devices can be smartphones, tablets, laptops, PDAs, personal computers, smart TVs, smartwatches, etc., but are not limited to these. The terminal devices and servers can be directly or indirectly connected via wired or wireless communication, which is not limited in this application. The number of servers and terminal devices is also not limited. The two metadata management systems will be described below.

[0113] I. The metadata management system includes terminal devices;

[0114] First, the terminal device obtains the existing metadata of the target metadata table. Then, the terminal device obtains K metadata information based on the existing metadata. Subsequently, it performs feature processing on the K metadata information to obtain K metadata features. The trained metadata prediction model is then called to process the K metadata features to obtain the target metadata information for the target metadata table. This target metadata information is the metadata information corresponding to the missing metadata.

[0115] II. The metadata management system includes terminal devices and servers;

[0116] First, the terminal device triggers an instruction to supplement information in the target metadata table. The server then retrieves the existing metadata from the target metadata table. Based on this existing metadata, the server obtains K metadata pieces of information. These K metadata pieces are then processed to obtain K metadata features. The server then uses a pre-trained metadata prediction model to process these K metadata features, resulting in the target metadata information for the target metadata table. This target metadata information corresponds to the missing metadata. Finally, the target metadata information can be fed back to the terminal device.

[0117] Understandably, a metadata management platform deployed on terminal devices enables centralized, automated, and version control. Centralized management ensures data consistency across enterprise application systems. Automated management allows data to be reflected and updated automatically through the platform, clearly displaying data relationships to improve communication and configuration efficiency and data quality. Version control integrates the metadata management platform with version application development and design, enhancing data reliability and authority.

[0118] This application utilizes the concepts of Machine Learning (ML) to predict missing metadata information using existing metadata. ML is a multidisciplinary field involving probability theory, statistics, approximation theory, convex analysis, and algorithm complexity theory. It specifically studies how computers can simulate or implement human learning behavior to acquire new knowledge or skills and reorganize existing knowledge structures to continuously improve their performance. ML is the core of artificial intelligence and the fundamental way to endow computers with intelligence; its applications span all areas of artificial intelligence. ML and deep learning typically include techniques such as artificial neural networks, belief networks, reinforcement learning, transfer learning, inductive learning, and instructional learning.

[0119] Based on this, the following will combine Figure 2 This describes the process of generating target metadata information. Please refer to [link / reference]. Figure 2 , Figure 2 This is a schematic diagram illustrating the target metadata information generation process in an embodiment of this application. As shown in the figure, assuming there is a data table with three categories of metadata, each category containing one metadata information, feature processing is performed on these metadata information, which is then input into a trained metadata prediction model. The metadata prediction model outputs the target metadata information for a specific missing metadata. It is understood that in practical applications, the missing metadata to be predicted can be selected according to requirements, and a corresponding metadata prediction model can be used for the prediction of each missing metadata.

[0120] Based on the above introduction, the method for generating metadata information in this application will be described below. Please refer to [link / reference]. Figure 3 One embodiment of the metadata information generation method in this application includes:

[0121] 101. Obtain the existing metadata of the target metadata table, where the existing metadata is metadata that already exists in the target metadata table, and the existing metadata includes K metadata, where K is an integer greater than or equal to 1;

[0122] In this embodiment, the metadata information generation device obtains the existing metadata of the target metadata table, wherein the target metadata table has K metadata, and these K metadata that have already appeared in the target metadata table are the existing metadata.

[0123] Specifically, metadata is generally divided into technical metadata, business metadata, and management metadata. Technical metadata refers to technical statistical indicators generated during data development. It is data used by data warehouse designers and managers for the development and daily management of the data warehouse. It mainly includes data source information, descriptions of data transformations, definitions of objects and data structures within the data warehouse, rules used for data cleaning and updates, and mappings from source data to destination data. Common technical metadata includes data lineage, fan-in count, fan-out count, field names, field lengths, and database table structures. Business metadata refers to information such as business names, definitions, and descriptions that represent various attributes and concepts in the enterprise environment. To a certain extent, the business context behind all data can be considered business metadata. Common business metadata includes business domains, business definitions, business descriptions, importance levels, and sensitivity levels. Management metadata describes concepts, relationships, and rules related to the management domain within the data system. Common management metadata mainly includes data owners, data quality accountability, and data security levels.

[0124] It should be noted that the metadata information generation device can be deployed on terminal devices, servers, or a metadata management system consisting of terminal devices and servers; no limitation is made here.

[0125] 102. Obtain K metadata information based on existing metadata, where each metadata information has a one-to-one correspondence with the existing metadata.

[0126] In this embodiment, the metadata information generation device acquires K metadata information from existing metadata. The existing metadata includes K metadata items. Therefore, by acquiring the metadata information corresponding to each metadata item, K metadata information items are obtained. These K metadata information items constitute "facts." In one case, all K metadata information items may originate from the target metadata table; in another case, some may originate from the target metadata data packet, and the other part may originate from the associated metadata table of the target metadata table. It is understood that facts are relative to predictions, i.e., truly determined metadata information. For example, the metadata is "fan-in number," and its corresponding metadata information is "50."

[0127] 103. Perform feature processing on K metadata information to obtain K metadata features, where the metadata features have a one-to-one correspondence with the metadata information;

[0128] In this embodiment, the metadata information generation device performs feature processing on each of the K metadata information to obtain the metadata features corresponding to each metadata information.

[0129] Specifically, metadata information can be processed based on feature engineering. Feature engineering includes, but is not limited to, feature normalization (e.g., normalizing the metadata information of technical metadata and business metadata, typically using normal distribution normalization) and feature transformation processing (e.g., regularization, discretization, and one-hot encoding of the metadata information of technical metadata and business metadata).

[0130] 104. Based on K metadata features, obtain target metadata information for the target metadata table through a metadata prediction model. The target metadata information is the metadata information corresponding to the missing metadata, and the missing metadata is metadata that does not exist in the target metadata table.

[0131] In this embodiment, the metadata information generation device invokes a metadata prediction model to predict information from the K metadata features obtained after feature processing, thereby obtaining target metadata information for the target metadata table. Information prediction refers to predicting missing metadata information based on existing metadata information using the metadata prediction model, thereby collecting complete metadata information. For example, if the missing metadata is "business importance level," the corresponding target metadata information is "very important."

[0132] This application provides a method for generating metadata information. Using the aforementioned method, based on existing metadata information, a trained model can predict missing metadata information, thereby enabling refined operation and management of data assets. On the one hand, it eliminates the need for technical personnel to spend significant time and effort manually supplementing missing metadata information, thus saving labor costs and improving work efficiency. On the other hand, predicting missing metadata information through a model avoids the influence of subjective factors, thereby improving information accuracy.

[0133] Optionally, in the above Figure 3 Based on the corresponding embodiments, in another optional embodiment provided by this application, the K metadata information includes table name text information, wherein the table name text information belongs to a preset language;

[0134] Feature processing is performed on K metadata information to obtain K metadata features, which may specifically include:

[0135] The table name text information is segmented to obtain a table name text sequence, where the table name text sequence includes M words, where M is an integer greater than or equal to 1;

[0136] Each word in the table name text sequence is preprocessed to obtain M word vector embeddings, where the word vector embeddings have a one-to-one correspondence with the words, and the M word vector embeddings belong to the metadata features of the table name text information;

[0137] Based on K metadata features, target metadata information is obtained through a metadata prediction model, which may include:

[0138] Based on the embedding of M word vectors, the semantic information vector corresponding to the table name text sequence is obtained through the encoder included in the metadata prediction model.

[0139] Based on semantic information vectors, table description information is obtained through the decoder included in the metadata prediction model, where the table description information belongs to the target metadata information.

[0140] This embodiment introduces a method for generating table description information using a metadata prediction model. Assuming the target metadata information is table description information, the following will combine... Figure 4 This paper introduces a method for obtaining table description information based on a metadata prediction model.

[0141] Specifically, please refer to Figure 4 , Figure 4This is a schematic diagram illustrating the acquisition of table description information based on a metadata prediction model in this embodiment of the application. As shown in the figure, the table name text information is assumed to be "Employee Information Registration Table," meaning the preset language can be English. Based on this, the table name text information can be segmented to obtain a table name text sequence, such as "Employee / Information / Registration / Table," where " / " represents the separator. Next, each word in the table name text sequence is preprocessed to obtain M word vector embeddings. For example, word vectorization (word-to-vector, word2vec) can be used to generate the word vector embedding corresponding to each word. Then, the M word vector embeddings are input to the encoder included in the metadata prediction model, and after encoding, a semantic information vector is generated. The semantic information vector is input to the decoder included in the metadata prediction model, and after decoding, a translation result is generated, i.e., the table description information, such as "Employee Information Registration Table."

[0142] It should be noted that the encoder included in the metadata prediction model may be a recurrent neural network (RNN), a long short-term memory network (LSTM), a gated recurrent unit (GRU), or other networks, which are not limited here.

[0143] It should be noted that the preset language can be English, Japanese, French, German or other languages. This is only for illustration and should not be construed as a limitation of this application.

[0144] Secondly, in this embodiment of the application, a method for generating table description information using a metadata prediction model is provided. In this method, a text translation model is used as a metadata prediction model to directly translate the existing table name text information, thereby inferring the table description information, thus improving the efficiency and accuracy of supplementing metadata information.

[0145] Optionally, in the above Figure 3 Based on the corresponding embodiments, in another optional embodiment provided by this application, obtaining K metadata information based on existing metadata may specifically include:

[0146] If the existing metadata includes the table name, then retrieve the table name text information corresponding to the table name from the target metadata table;

[0147] or,

[0148] Retrieve K pieces of metadata information based on existing metadata, which may specifically include:

[0149] If the existing metadata does not include the data table name, then the associated metadata table is obtained based on the data lineage included in the existing metadata. The associated metadata table is either the upstream metadata table or the downstream metadata table.

[0150] Retrieve the text information of the table name corresponding to the data table name from the associated metadata table.

[0151] This embodiment introduces a method for obtaining table name text information based on the associated metadata table. As can be seen from the previous embodiments, when predicting table description information, it is necessary to obtain the table name text information corresponding to the data table name. The following will illustrate the method for obtaining table name text information with examples.

[0152] Method 1: Directly retrieve the table name text information from the target metadata table;

[0153] Specifically, assuming that the target metadata table contains a table name among the K metadata information, then we can directly extract the table name text information corresponding to that table name.

[0154] Method 2: Obtain the table name text information from the associated metadata table;

[0155] Specifically, assuming that the table name is not present among the K metadata entries in the target metadata table, we then obtain the data lineage of the target metadata table. Data merging, transformation, and circulation form data lineage relationships, which reveal the relationships between tables and fields. For easier understanding, please refer to [link to relevant documentation]. Figure 5 , Figure 5 This is a schematic diagram of obtaining table name text information based on data lineage in an embodiment of this application. As shown in the figure, the associated metadata table of the target metadata table can be metadata table A, metadata table B, or metadata table F. Metadata table A and metadata table B belong to the upstream data tables of the target metadata table, and metadata table F belongs to the downstream data table of the target metadata table.

[0156] Based on this, any associated metadata table that includes a data table name can be selected, and the table name text information corresponding to the data table name in the associated metadata table can be used as the table name text information of the target metadata table.

[0157] Furthermore, in this embodiment of the application, a method for obtaining table name text information based on associated metadata tables is provided. In the case where there is no data table name in the target metadata table, the data table name of the associated metadata table can be obtained according to the data lineage, and the table name text information corresponding to the data table name can be used as the basis for predicting table description information. On the one hand, this can improve the flexibility of information prediction, and on the other hand, it can help to obtain more accurate prediction results.

[0158] Optionally, in the above Figure 3 Based on the corresponding embodiments, in another optional embodiment provided by the present application, the existing metadata includes at least one of the following: number of downstream dependent tasks, number of upstream dependent tasks, access frequency, online business service status, number of online services, number of fields, number of data rows, and storage size;

[0159] Feature processing is performed on K metadata information to obtain K metadata features, which may specifically include:

[0160] If the existing metadata includes the number of downstream dependent tasks, then the metadata information of the number of downstream dependent tasks is binned to obtain the metadata features of the number of downstream dependent tasks. The metadata features of the number of downstream dependent tasks belong to K metadata features.

[0161] If the existing metadata includes the number of upstream dependent tasks, then the metadata information of the number of upstream dependent tasks is binned to obtain the metadata features of the number of upstream dependent tasks. The metadata features of the number of upstream dependent tasks belong to K metadata features.

[0162] If the existing metadata includes access frequency, then the access frequency is binned to obtain the metadata features of the access frequency, where the metadata features of the access frequency belong to K metadata features.

[0163] If the existing metadata includes information about online business services, then the online business service information is one-hot encoded to obtain the metadata features of the online business service information. The metadata features of the online business service information belong to K metadata features.

[0164] If the existing metadata includes the number of online services, feature scaling is performed on the number of online services to obtain the metadata features of the number of online services. The metadata features of the number of online services belong to K metadata features.

[0165] If the existing metadata includes the number of fields, perform feature scaling on the number of fields to obtain the metadata feature of the number of fields, where the metadata feature of the number of fields belongs to K metadata features;

[0166] If the existing metadata includes the number of data rows, perform feature scaling on the number of data rows to obtain the metadata features of the number of data rows, where the metadata features of the number of data rows belong to K metadata features;

[0167] If the existing metadata includes storage size, feature scaling is performed on the storage size to obtain the metadata features of the storage size, where the metadata features of the storage size belong to K metadata features.

[0168] This embodiment introduces a method for feature processing of metadata information based on feature engineering. Typically, existing metadata includes at least one of the following: number of downstream dependent tasks, number of upstream dependent tasks, access frequency, online business service status, number of online services, number of fields, number of data rows, and storage size. The feature processing methods for these metadata information will be described below.

[0169] 1. Number of downstream dependent tasks (i.e., fan-out number);

[0170] Specifically, the number of downstream dependent tasks refers to the number of lower-level modules directly called by this module; a large fan-out indicates high module complexity. The metadata information of the number of downstream dependent tasks can be represented numerically. Data binning is performed on the metadata information of the number of downstream dependent tasks to obtain its metadata characteristics. Data binning (also known as discrete binning or segmentation) is a data preprocessing technique used to reduce the impact of minor observation errors; it is a method of grouping multiple continuous values ​​into a smaller number of "bins".

[0171] It should be noted that the feature processing method for the number of downstream dependent tasks is not limited to data binning. In practical applications, other reasonable feature processing methods can be selected according to the situation. This is only for illustration and should not be construed as a limitation of this application.

[0172] 2. Number of upstream dependent tasks (i.e., fan-in number);

[0173] Specifically, the number of upstream dependent tasks refers to the number of parent modules that directly call this module. A large fan-in indicates high reusability of the module. The metadata information of the number of upstream dependent tasks can be represented numerically. Binning processing is performed on the metadata information of the number of upstream dependent tasks to obtain the metadata characteristics of the number of upstream dependent tasks.

[0174] It should be noted that the feature processing method for the number of upstream dependent tasks is not limited to data binning. In practical applications, other reasonable feature processing methods can be selected according to the situation. This is only for illustration and should not be construed as a limitation of this application.

[0175] III. Access frequency;

[0176] Specifically, the access frequency can be 1 day, 7 days, or 30 days, etc., meaning the metadata information of the access frequency can be represented as a numerical value. Binning processing is then performed on the metadata information of the access frequency to obtain the metadata characteristics of the access frequency.

[0177] It should be noted that the characteristic processing method for access frequency is not limited to data binning. In practical applications, other reasonable characteristic processing methods can be selected according to the situation. This is only for illustration and should not be construed as a limitation of this application.

[0178] IV. Online business service status;

[0179] Specifically, online business service status refers to whether online business services are provided. The metadata information for online business services can be represented as a result, such as "yes" or "no". One-hot encoding is performed on the metadata information of the online business service status to obtain its metadata features. One-hot encoding is a data preprocessing method that transforms categorical data into features of the same length. For example, if the online business service status is "yes", the metadata feature is represented as (1,0). Conversely, if the online business service status is "no", the metadata feature is represented as (0,1).

[0180] It should be noted that the feature processing method for online business service status is not limited to one-hot encoding. In practical applications, other reasonable feature processing methods can be selected according to the situation. This is only for illustration and should not be construed as a limitation of this application.

[0181] V. Number of online services;

[0182] Specifically, the metadata information of the number of online services can be represented as numerical values. Feature scaling is then applied to this metadata information to obtain the metadata features of the online service quantity. Feature scaling is a method for standardizing the range of independent variables or data features. In data processing, it is also called data standardization and is typically performed during the data preprocessing step. Feature scaling can limit a large range of data to a specified range.

[0183] It should be noted that the feature processing method for the number of online services is not limited to feature scaling. In practical applications, other reasonable feature processing methods can be selected according to the situation, such as data binning. This is only for illustration and should not be construed as a limitation of this application.

[0184] VI. Number of fields;

[0185] Specifically, the metadata information regarding the number of fields can be represented as a numerical value. Feature scaling processing is then applied to this metadata information to obtain the metadata feature representing the number of fields.

[0186] It should be noted that the feature processing method for the number of fields is not limited to feature scaling. In practical applications, other reasonable feature processing methods can be selected according to the situation, such as data binning. This is only for illustration and should not be construed as a limitation of this application.

[0187] VII. Number of data rows;

[0188] Specifically, the metadata information for the number of data rows can be represented as a numerical value. Feature scaling processing is then applied to the metadata information for the number of data rows to obtain the metadata features for that number of data rows.

[0189] It should be noted that the feature processing method for the number of data rows is not limited to feature scaling. In practical applications, other reasonable feature processing methods can be selected according to the situation, such as data binning. This is only for illustration and should not be construed as a limitation of this application.

[0190] 8. Storage size;

[0191] Specifically, the metadata information of storage size can be represented as a numerical value. Feature scaling processing is then applied to the metadata information of storage size to obtain the metadata features of storage size.

[0192] It should be noted that the feature processing method for storage size is not limited to feature scaling. In practical applications, other reasonable feature processing methods can be selected according to the situation, such as data binning. This is only for illustration and should not be construed as a limitation of this application.

[0193] Secondly, in this application embodiment, a method for feature processing of metadata information based on feature engineering is provided. Through the above method, the metadata features obtained after processing can have better interpretability (e.g., by using data binning), and can capture more complex relationships (e.g., by using neural networks), and reduce data redundancy and reduce data dimensionality (e.g., by principal component analysis). Thus, accurate prediction results can be generated using machine learning.

[0194] Optionally, in the above Figure 3 Based on the corresponding embodiments, in another optional embodiment provided by this application, the existing metadata also includes at least one of the following: business importance level and business sensitivity.

[0195] Feature processing is performed on K metadata information to obtain K metadata features, which may specifically include:

[0196] If the existing metadata includes the business importance level, then the metadata information corresponding to the business importance level is one-hot encoded to obtain the metadata features of the business importance level. The metadata features of the business importance level belong to K metadata features.

[0197] If the existing metadata includes business-sensitive cases, then the metadata information corresponding to the business-sensitive cases is processed by one-hot encoding to obtain the metadata features of the business-sensitive cases. The metadata features of the business-sensitive cases belong to K metadata features.

[0198] Based on K metadata features, target metadata information for the target metadata table is obtained through a metadata prediction model, including:

[0199] Based on K metadata features, the business type information of the target metadata table is obtained through a metadata prediction model.

[0200] This embodiment introduces a method for generating business type information based on a metadata prediction model. Existing metadata also includes at least one of business importance level and business sensitivity. Assuming both business importance level and business sensitivity are included, the metadata information corresponding to the business importance level (e.g., "yes") needs to be one-hot encoded to obtain the metadata feature of the business importance level (e.g., (1,0)). The metadata information corresponding to the business sensitivity (e.g., "no") is then one-hot encoded to obtain the metadata feature of the business sensitivity (e.g., (0,1)).

[0201] Specifically, the K metadata features obtained after feature processing are input into the trained metadata prediction model, thereby outputting the business type information, such as "WeChat Moments business" or "payment business". In the classification task, it is necessary to model the training set and predict manually labeled tags. The metadata prediction model used in this application can be a traditional learning model, such as a tree model or Logistic Regression (LR), or an ensemble learning model, such as a random forest model or an extreme gradient boosting (XGBoost) model, or a deep learning model, such as LSTM or neural networks, etc., without limitation.

[0202] During the training phase, model evaluation can also be performed, which involves validating the model using predictions on the test set data after modeling. Multi-class classification models can be evaluated using metrics such as confusion matrices. (See [link to relevant documentation]). Figure 6 , Figure 6This is a schematic diagram illustrating the use of a confusion matrix to evaluate a metadata prediction model in an embodiment of this application. As shown in the figure, assuming there are 150 sample data points, 50 are predicted as category 1, 50 as category 2, and 50 as category 3. The sum of each row represents the true number of samples in that category, and the sum of each column represents the number of samples predicted as belonging to that category. For example, the first row indicates that 43 samples belonging to category 1 were correctly predicted as category 1, and 2 samples belonging to category 1 were incorrectly predicted as category 2.

[0203] Understandably, the area under the curve (AUC), accuracy, or recall can also be used to evaluate model performance.

[0204] Furthermore, in this embodiment of the application, a method for generating business type information based on a metadata prediction model is provided. By using the above method, existing metadata information can be used, and a trained model can be used to predict the missing business type information, thereby achieving refined operation and management of data assets.

[0205] Optionally, in the above Figure 3 Based on the corresponding embodiments, in another optional embodiment provided by this application, the business type information of the target metadata table is obtained through a metadata prediction model based on K metadata features. Specifically, this may include:

[0206] Based on K metadata features, the business type information of the target metadata table is obtained through a decision tree model. The decision tree model is a metadata prediction model, which is a network based on the relationship between nodes, with each node corresponding to a different metadata feature.

[0207] or,

[0208] Based on K metadata features, a metadata prediction model is used to obtain the business type information of the target metadata table, which may include:

[0209] Based on K metadata features, the business type information of the target metadata table is obtained through a random forest model. The random forest model is a metadata prediction model, which includes at least two decision trees, each of which is used to predict a classification result.

[0210] or,

[0211] Based on K metadata features, a metadata prediction model is used to obtain the business type information of the target metadata table, which may include:

[0212] Based on K metadata features, the business type of the target metadata table is obtained through an extensible boosting tree model;

[0213] or,

[0214] Based on K metadata features, a metadata prediction model is used to obtain the business type information of the target metadata table, which may include:

[0215] Based on K metadata features, the first feature vector is obtained through the input layer included in the metadata prediction model;

[0216] Based on the first feature vector, a second feature vector is obtained through at least one hidden layer included in the metadata prediction model;

[0217] Based on the second feature vector, the probability distribution is obtained through the output layer of the metadata prediction model;

[0218] The business type information of the target metadata table is determined based on the probability distribution.

[0219] This embodiment introduces a method for predicting the business type based on a model. As can be seen from the foregoing embodiments, the metadata prediction model used in this application can be a traditional learning model, an ensemble learning model, or a deep learning model. The types of metadata prediction models used in this application will be described below.

[0220] I. Decision Tree Model;

[0221] Specifically, the decision tree model is used as a metadata prediction model. The decision tree model includes multiple nodes, each node identifies a classification on a metadata feature. After inputting K metadata features into the decision tree model, these K metadata features are classified, and finally the business type information is output.

[0222] II. Random Forest Model;

[0223] Specifically, a random forest model is used as the metadata prediction model. The random forest model consists of multiple decision trees, each with multiple nodes, and each node identifies a classification on a metadata feature. After inputting K metadata features into different decision tree models, the business type information is determined based on the classification results output by all decision trees.

[0224] III. XGBoost Model;

[0225] Specifically, the XGBoost model serves as a metadata prediction model. After inputting K metadata features into different XGBoost models, the XGBoost model can perform initial screening of the K metadata features, removing irrelevant or redundant features, retaining the top few metadata features, and finally outputting the relevant business type information.

[0226] IV. Neural Network Model;

[0227] Specifically, K metadata features are input into the input layer of the metadata prediction model, which outputs a first feature vector. This first feature vector is then input into at least one hidden layer to obtain a second feature vector. Finally, the second feature vector is input into the output layer, where a softmax operation can be used to obtain a probability distribution. The business type information can then be determined based on this probability distribution.

[0228] Furthermore, in this embodiment of the application, a method for predicting the business type information based on a model is provided. Through the above method, a variety of models that can predict the business type information are provided. Among them, the XGBoost model shows a relatively strong classification effect. In the choice between computational cost and final effect, using the XGBoost model as the classification model can have a better classification effect.

[0229] Optionally, in the above Figure 3 Based on the corresponding embodiments, in another optional embodiment provided by this application, the existing metadata also includes at least one of the business type and business sensitivity.

[0230] Feature processing is performed on K metadata information to obtain K metadata features, which may specifically include:

[0231] If the existing metadata includes the business type, then the metadata information corresponding to the business type is one-hot encoded to obtain the metadata features of the business type. The metadata features of the business type belong to K metadata features.

[0232] If the existing metadata includes business-sensitive cases, then the metadata information corresponding to the business-sensitive cases is processed by one-hot encoding to obtain the metadata features of the business-sensitive cases. The metadata features of the business-sensitive cases belong to K metadata features.

[0233] Based on K metadata features, target metadata information for the target metadata table is obtained through a metadata prediction model, which may specifically include:

[0234] Based on K metadata features, a metadata prediction model is used to obtain business importance level information for the target metadata table.

[0235] This embodiment introduces a method for generating business importance level information based on a metadata prediction model. Existing metadata also includes at least one of the business type and business sensitivity. Assuming both the business type and business sensitivity are included, the metadata information corresponding to the business type (e.g., "WeChat Moments business") needs to be one-hot encoded to obtain the metadata features of the business type (e.g., (1,0,0,0)). The metadata information corresponding to the business sensitivity (e.g., "No") is then one-hot encoded to obtain the metadata features of the business sensitivity (e.g., (0,1)).

[0236] Specifically, the K metadata features obtained after feature processing are input into the trained metadata prediction model, thereby outputting business importance level information, such as "important". In the classification task, it is necessary to model the training set and predict manually labeled tags. The metadata prediction model used in this application can be a traditional learning model, an ensemble learning model, or a deep learning model; no limitation is made here. During the training phase, model evaluation can also be performed, i.e., after modeling, predictions are validated on the test set data. Multi-class model evaluation metrics can be used, including but not limited to confusion matrix, AUC, accuracy, or recall, to evaluate model performance.

[0237] Furthermore, this application provides a method for generating business importance level information based on a metadata prediction model. By using the existing metadata information and a trained model, missing business importance level information can be predicted, thereby achieving refined operation and management of data assets.

[0238] Optionally, in the above Figure 3 Based on the corresponding embodiments, in another optional embodiment provided by this application, business importance level information for the target metadata table is obtained through a metadata prediction model based on K metadata features. Specifically, this may include:

[0239] Based on K metadata features, the business importance level information for the target metadata table is obtained through a decision tree model. The decision tree model is a metadata prediction model, which is a network based on the relationship between nodes, with each node corresponding to a different metadata feature.

[0240] or,

[0241] Based on K metadata features, a metadata prediction model is used to obtain business importance level information for a target metadata table, which may specifically include:

[0242] Based on K metadata features, a random forest model is used to obtain business importance level information for the target metadata table. The random forest model is a metadata prediction model, which includes at least two decision trees, each of which is used to predict a classification result.

[0243] or,

[0244] Based on K metadata features, a metadata prediction model is used to obtain business importance level information for a target metadata table, which may specifically include:

[0245] Based on K metadata features, a scalable boosting tree model is used to obtain the business importance level of the target metadata table;

[0246] or,

[0247] Based on K metadata features, a metadata prediction model is used to obtain business importance level information for a target metadata table, which may specifically include:

[0248] Based on K metadata features, the first feature vector is obtained through the input layer included in the metadata prediction model;

[0249] Based on the first feature vector, a second feature vector is obtained through at least one hidden layer included in the metadata prediction model;

[0250] Based on the second feature vector, the probability distribution is obtained through the output layer of the metadata prediction model;

[0251] The business importance level information of the target metadata table is determined based on the probability distribution.

[0252] This embodiment introduces a method for predicting the importance level of business information based on a model. As can be seen from the foregoing embodiments, the metadata prediction model used in this application can be a traditional learning model, an ensemble learning model, or a deep learning model. The type of metadata prediction model used in this application will be described below.

[0253] I. Decision Tree Model;

[0254] Specifically, the decision tree model is used as a metadata prediction model. The decision tree model includes multiple nodes, each node identifies a classification on a metadata feature. After inputting K metadata features into the decision tree model, these K metadata features are classified, and finally the business importance level information is output.

[0255] II. Random Forest Model;

[0256] Specifically, a random forest model is used as the metadata prediction model. The random forest model consists of multiple decision trees, each with multiple nodes, and each node identifies a classification on a metadata feature. After inputting K metadata features into different decision tree models, the business importance level information is determined based on the classification results output by all decision trees.

[0257] III. XGBoost Model;

[0258] Specifically, the XGBoost model serves as a metadata prediction model. After inputting K metadata features into different XGBoost models, the XGBoost model can perform initial screening of the K metadata features, removing irrelevant or redundant features, retaining the top few metadata features, and finally outputting the business importance level information.

[0259] IV. Neural Network Model;

[0260] Specifically, K metadata features are input into the input layer of the metadata prediction model, which outputs a first feature vector. This first feature vector is then input into at least one hidden layer to obtain a second feature vector. Finally, the second feature vector is input into the output layer, where a softmax operation can be used to obtain a probability distribution. The importance level of the business can then be determined based on this probability distribution.

[0261] Furthermore, in this embodiment of the application, a method for predicting business importance level information based on a model is provided. Through the above method, various models that can predict business importance level information are provided. Among them, the XGBoost model shows a relatively strong classification effect. In the choice between computational cost and final effect, using the XGBoost model as the classification model can have a better classification effect.

[0262] Optionally, in the above Figure 3 Based on the corresponding embodiments, in another optional embodiment provided by this application, the existing metadata also includes at least one of the business type and the business importance level;

[0263] Feature processing is performed on K metadata information to obtain K metadata features, which may specifically include:

[0264] If the existing metadata includes the business type, then the metadata information corresponding to the business type is one-hot encoded to obtain the metadata features of the business type. The metadata features of the business type belong to K metadata features.

[0265] If the existing metadata includes the business importance level, then the metadata information corresponding to the business importance level is one-hot encoded to obtain the metadata features of the business importance level. The metadata features of the business importance level belong to K metadata features.

[0266] Based on K metadata features, target metadata information for the target metadata table is obtained through a metadata prediction model, which may specifically include:

[0267] Based on K metadata features, a metadata prediction model is used to obtain information on business sensitivity of the target metadata table.

[0268] This embodiment introduces a method for generating business-sensitive information based on a metadata prediction model. Existing metadata also includes at least one of the business type and business importance level. Assuming both business type and business importance level are included, the metadata information corresponding to the business type (e.g., "WeChat Moments business") needs to be one-hot encoded to obtain the metadata feature of the business importance level (e.g., (1,0,0,0)). The metadata information corresponding to the business importance level (e.g., "is") is then one-hot encoded to obtain the metadata feature of the business importance level (e.g., (1,0)).

[0269] Specifically, the K metadata features obtained after feature processing are input into the trained metadata prediction model, thereby outputting business sensitivity information, for example, "yes". In the classification task, it is necessary to model the training set and predict manually labeled tags. The metadata prediction model used in this application can be a traditional learning model, an ensemble learning model, or a deep learning model; no limitation is made here. During the training phase, model evaluation can also be performed, that is, after modeling, predictions are validated on the test set data. Multi-class model evaluation metrics can be used, including but not limited to confusion matrix, AUC, accuracy, or recall, to evaluate the model performance.

[0270] Furthermore, this application provides a method for generating business-sensitive information based on a metadata prediction model. By using existing metadata information and a trained model, missing business-sensitive information can be predicted, thereby achieving refined operation and management of data assets.

[0271] Optionally, in the above Figure 3 Based on the corresponding embodiments, in another optional embodiment provided by this application, business sensitivity information for the target metadata table is obtained through a metadata prediction model based on K metadata features. Specifically, this may include:

[0272] Based on K metadata features, business sensitivity information for the target metadata table is obtained through a decision tree model. The decision tree model is a metadata prediction model, which is a network based on the relationship between nodes, with each node corresponding to a different metadata feature.

[0273] or,

[0274] Based on K metadata features, a metadata prediction model is used to obtain business sensitivity information for the target metadata table, which may include:

[0275] Based on K metadata features, a random forest model is used to obtain business-sensitive information for the target metadata table. The random forest model is a metadata prediction model, which includes at least two decision trees, each of which is used to predict a classification result.

[0276] or,

[0277] Based on K metadata features, a metadata prediction model is used to obtain business sensitivity information for the target metadata table, which may include:

[0278] Based on K metadata features, a scalable boosting tree model is used to obtain business-sensitive information for the target metadata table;

[0279] or,

[0280] Based on K metadata features, a metadata prediction model is used to obtain business sensitivity information for the target metadata table, which may include:

[0281] Based on K metadata features, the first feature vector is obtained through the input layer included in the metadata prediction model;

[0282] Based on the first feature vector, a second feature vector is obtained through at least one hidden layer included in the metadata prediction model;

[0283] Based on the second feature vector, the probability distribution is obtained through the output layer of the metadata prediction model;

[0284] The business sensitivity information of the target metadata table is determined based on the probability distribution.

[0285] This embodiment introduces a method for predicting business-sensitive information based on a model. As can be seen from the foregoing embodiments, the metadata prediction model used in this application can be a traditional learning model, an ensemble learning model, or a deep learning model. The types of metadata prediction models used in this application will be described below.

[0286] I. Decision Tree Model;

[0287] Specifically, the decision tree model is used as a metadata prediction model. The decision tree model includes multiple nodes, each node identifies a classification on a metadata feature. After inputting K metadata features into the decision tree model, these K metadata features are classified, and finally, business sensitivity information is output.

[0288] II. Random Forest Model;

[0289] Specifically, a random forest model is used as the metadata prediction model. The random forest model consists of multiple decision trees, each with multiple nodes, and each node identifies a classification on a metadata feature. After inputting K metadata features into different decision tree models, business sensitivity information is determined based on the classification results output by all decision trees.

[0290] III. XGBoost Model;

[0291] Specifically, the XGBoost model serves as a metadata prediction model. After inputting K metadata features into different XGBoost models, the XGBoost model can perform initial screening of the K metadata features, removing irrelevant or redundant features, retaining the top few metadata features, and finally outputting information on business sensitivity.

[0292] IV. Neural Network Model;

[0293] Specifically, K metadata features are input into the input layer of the metadata prediction model, which outputs a first feature vector. This first feature vector is then input into at least one hidden layer to obtain a second feature vector. Finally, the second feature vector is input into the output layer, where a softmax operation can be used to obtain a probability distribution. The business sensitivity information can then be determined based on this probability distribution.

[0294] Furthermore, in this embodiment of the application, a method for predicting business-sensitive information based on a model is provided. Through the above method, various models that can predict business-sensitive information are provided. Among them, the XGBoost model demonstrates a relatively strong classification effect. In the choice between computational cost and final effect, using the XGBoost model as the classification model can achieve better classification results.

[0295] Optionally, in the above Figure 3 Based on the corresponding embodiments, in another optional embodiment provided by this application, after obtaining the target metadata information for the target metadata table through a metadata prediction model based on K metadata features, it may further include:

[0296] Generate K metadata records based on K metadata elements and K metadata information;

[0297] Generate a target metadata record based on the missing metadata and the target metadata information;

[0298] K metadata records and target metadata records are synchronously stored on the blockchain, whereby the blockchain is used to provide metadata information for the data table to at least two node devices.

[0299] This embodiment introduces a method for storing metadata information of a data table based on blockchain technology. After generating the target metadata information, metadata records can be generated for the metadata and its corresponding metadata information. Similarly, target metadata records can be generated for missing metadata and its corresponding target metadata information. All K metadata records and the target metadata are recorded in the block for synchronous storage.

[0300] Specifically, the above process can be implemented in a distributed system. The system involved in this application can be a distributed system formed by connecting a client 202 and multiple nodes 201 (any form of computing device in the network, such as a server or terminal) through network communication. Taking a blockchain system as an example of a distributed system, see [link to relevant documentation]. Figure 7 , Figure 7 This is an optional structural diagram illustrating the application of a distributed system to a blockchain system in this embodiment of the application. It consists of multiple nodes (any form of computing device connected to the network, such as servers or user terminals) and clients, forming a peer-to-peer (P2P) network. The P2P protocol is an application layer protocol running on top of the Transmission Control Protocol (TCP). In the distributed system, any machine, such as a server or terminal, can join and become a node. A node includes a hardware layer, a middleware layer, an operating system layer, and an application layer.

[0301] See Figure 7 The functions of each node in the blockchain system shown include:

[0302] 1) Routing: A basic function of nodes used to support communication between nodes.

[0303] In addition to routing capabilities, nodes can also have the following functions:

[0304] 2) Applications are deployed in the blockchain to implement specific business needs. They record data related to the implementation of functions to form record data, carry digital signatures in the record data to indicate the source of the task data, and send the record data to other nodes in the blockchain system. When other nodes successfully verify the source and integrity of the record data, they add the record data to a temporary block.

[0305] For example, the business logic implemented by the application includes:

[0306] 2.1) A wallet is used to provide the function of conducting electronic currency transactions, including initiating transactions (i.e., sending the transaction record of the current transaction to other nodes in the blockchain system; after other nodes successfully verify the transaction, they store the transaction record data in the temporary block of the blockchain as a response to acknowledge the validity of the transaction; of course, the wallet also supports querying the remaining electronic currency in the electronic currency address;

[0307] 2.2) Shared ledger, used to provide functions such as storage, query and modification of ledger data. It sends the record data of the operation on the ledger data to other nodes in the blockchain system. After the other nodes verify the validity, as a response to acknowledge the validity of the ledger data, they store the record data in a temporary block. They can also send confirmation to the node that initiated the operation.

[0308] 2.3) Smart contracts are computerized protocols that can execute the terms of a contract. They are implemented through code deployed on a shared ledger that executes when certain conditions are met. Based on actual business needs, the code is used to complete automated transactions, such as querying the logistics status of goods purchased by a buyer and transferring the buyer's electronic money to the merchant's address after the buyer signs for the goods. Of course, smart contracts are not limited to executing contracts for transactions; they can also execute contracts for processing received information.

[0309] 3) A blockchain consists of a series of blocks that are sequentially generated. Once a new block is added to the blockchain, it will not be removed. The blocks contain the data submitted by the nodes in the blockchain system.

[0310] Please see Figure 8 , Figure 8 This is an optional schematic diagram of the block structure in this embodiment. Each block includes the hash value of the transaction records stored in this block (the hash value of this block) and the hash value of the previous block. The blocks are connected through their hash values ​​to form a blockchain. Additionally, the block may include information such as a timestamp when it was generated. A blockchain is essentially a decentralized database, a chain of data blocks linked together using cryptographic methods. Each data block contains relevant information used to verify the validity of the information (anti-counterfeiting) and to generate the next block.

[0311] Secondly, in this embodiment of the application, a method for storing metadata information of a data table based on blockchain technology is provided. Through the above method, the actual metadata information and the predicted metadata information can also be processed on the chain, which makes it convenient for relevant personnel or institutions to view them, and can effectively prevent the tampering of these metadata information.

[0312] Optionally, in the above Figure 3 Based on the corresponding embodiments, in another optional embodiment provided by this application, after obtaining the target metadata information for the target metadata table through a metadata prediction model based on K metadata features, it may further include:

[0313] Displays the target metadata information in the target metadata table;

[0314] or,

[0315] The target metadata information of the target metadata table is sent to the terminal device so that the terminal device can display the target metadata information of the target metadata table.

[0316] This embodiment describes a method for automatically completing and displaying the content of a target metadata table. For relevant personnel, an information completion command can be triggered through the metadata management platform to select the target metadata information to be supplemented.

[0317] Specifically, for ease of understanding, please refer to Figure 9 , Figure 9 This is a schematic diagram of the operation interface for automatically supplementing target metadata information in the embodiments of this application, such as... Figure 9 As shown in Figure (A), the metadata management platform displays the target metadata table (i.e., metadata table A). If a user wants to understand the missing metadata, an information supplementation instruction is triggered. Assuming an information supplementation instruction is triggered for the "related business type," the user will be redirected to... Figure 9 The interface shown in Figure (B) displays the business type information corresponding to "Business Type" (e.g., financial business). If a request to supplement information regarding "Business Importance Level" is triggered, the user will be redirected to the interface shown in Figure (B). Figure 9 The interface shown in Figure (C) displays the business importance level information corresponding to the "Business Importance Level" (e.g., highest). Assuming a supplementary information command for "Business Sensitive Situation" is triggered, the user will be redirected to the interface shown in Figure (C). Figure 9 The interface shown in Figure (D) displays the business sensitivity information (e.g., sensitive) corresponding to "Business Sensitive Situation".

[0318] Secondly, in this embodiment of the application, a method is provided to automatically complete and display the content of the target metadata table. Through the above method, relevant personnel can also select the metadata to be supplemented according to their needs. After prediction, the predicted metadata data is provided to the relevant parties, thereby improving the flexibility and operability of the solution.

[0319] The metadata information generation apparatus in this application is described in detail below. Please refer to [link / reference]. Figure 10 , Figure 10 This is a schematic diagram of one embodiment of the metadata information generation device in this application. The metadata information generation device 30 includes:

[0320] The acquisition module 301 is used to acquire the existing metadata of the target metadata table, wherein the existing metadata is metadata that already exists in the target metadata table, and the existing metadata includes K metadata, where K is an integer greater than or equal to 1;

[0321] The acquisition module 301 is also used to acquire K metadata information based on existing metadata, wherein the metadata information has a one-to-one correspondence with the metadata;

[0322] The processing module 302 is used to perform feature processing on K metadata information to obtain K metadata features, wherein the metadata features have a one-to-one correspondence with the metadata information;

[0323] The acquisition module 301 is also used to acquire target metadata information for the target metadata table based on K metadata features through a metadata prediction model, wherein the target metadata information is the metadata information corresponding to the missing metadata, and the missing metadata is metadata that does not exist in the target metadata table.

[0324] Optionally, in the above Figure 10 Based on the corresponding embodiments, in another embodiment of the metadata information generation device 30 provided in this application, the K metadata information includes table name text information, wherein the table name text information belongs to a preset language;

[0325] The processing module 302 is specifically used to perform word segmentation on the table name text information to obtain a table name text sequence, wherein the table name text sequence includes M words, where M is an integer greater than or equal to 1;

[0326] Each word in the table name text sequence is preprocessed to obtain M word vector embeddings, where the word vector embeddings have a one-to-one correspondence with the words, and the M word vector embeddings belong to the metadata features of the table name text information;

[0327] The acquisition module 301 is specifically used to acquire the semantic information vector corresponding to the table name text sequence based on the embedding of M word vectors and through the encoder included in the metadata prediction model;

[0328] Based on semantic information vectors, table description information is obtained through the decoder included in the metadata prediction model, where the table description information belongs to the target metadata information.

[0329] Optionally, in the above Figure 10 Based on the corresponding embodiments, in another embodiment of the metadata information generation device 30 provided in this application,

[0330] The acquisition module 301 is specifically used to obtain the table name text information corresponding to the data table name from the target metadata table if the existing metadata includes the data table name;

[0331] or,

[0332] The acquisition module 301 is specifically used to acquire the associated metadata table based on the data lineage included in the existing metadata if the existing metadata does not include the data table name. The associated metadata table is either the upstream metadata table or the downstream metadata table.

[0333] Retrieve the text information of the table name corresponding to the data table name from the associated metadata table.

[0334] Optionally, in the above Figure 10 Based on the corresponding embodiments, in another embodiment of the metadata information generation device 30 provided in this application, the existing metadata includes at least one of the following: number of downstream dependent tasks, number of upstream dependent tasks, access frequency, online business service status, number of online services, number of fields, number of data rows, and storage size.

[0335] The processing module 302 is specifically used to perform data binning on the metadata information of the downstream dependent task number if the existing metadata includes the number of downstream dependent tasks, to obtain the metadata features of the downstream dependent task number, wherein the metadata features of the downstream dependent task number belong to K metadata features.

[0336] If the existing metadata includes the number of upstream dependent tasks, then the metadata information of the number of upstream dependent tasks is binned to obtain the metadata features of the number of upstream dependent tasks. The metadata features of the number of upstream dependent tasks belong to K metadata features.

[0337] If the existing metadata includes access frequency, then the access frequency is binned to obtain the metadata features of the access frequency, where the metadata features of the access frequency belong to K metadata features.

[0338] If the existing metadata includes information about online business services, then the online business service information is one-hot encoded to obtain the metadata features of the online business service information. The metadata features of the online business service information belong to K metadata features.

[0339] If the existing metadata includes the number of online services, feature scaling is performed on the number of online services to obtain the metadata features of the number of online services. The metadata features of the number of online services belong to K metadata features.

[0340] If the existing metadata includes the number of fields, perform feature scaling on the number of fields to obtain the metadata feature of the number of fields, where the metadata feature of the number of fields belongs to K metadata features;

[0341] If the existing metadata includes the number of data rows, perform feature scaling on the number of data rows to obtain the metadata features of the number of data rows, where the metadata features of the number of data rows belong to K metadata features;

[0342] If the existing metadata includes storage size, feature scaling is performed on the storage size to obtain the metadata features of the storage size, where the metadata features of the storage size belong to K metadata features.

[0343] Optionally, in the above Figure 10 Based on the corresponding embodiments, in another embodiment of the metadata information generation device 30 provided in this application, the existing metadata also includes at least one of business importance level and business sensitivity.

[0344] The processing module 302 is specifically used to perform one-hot encoding on the metadata information corresponding to the business importance level if the existing metadata includes the business importance level, so as to obtain the metadata features of the business importance level, wherein the metadata features of the business importance level belong to K metadata features.

[0345] If the existing metadata includes business-sensitive cases, then the metadata information corresponding to the business-sensitive cases is processed by one-hot encoding to obtain the metadata features of the business-sensitive cases. The metadata features of the business-sensitive cases belong to K metadata features.

[0346] The acquisition module 301 is specifically used to obtain the business type information of the target metadata table based on K metadata features through a metadata prediction model.

[0347] Optionally, in the above Figure 10 Based on the corresponding embodiments, in another embodiment of the metadata information generation device 30 provided in this application,

[0348] The acquisition module 301 is specifically used to acquire the business type information of the target metadata table based on K metadata features through a decision tree model. The decision tree model is a metadata prediction model, which is a network based on the relationship between nodes, with each node corresponding to a different metadata feature.

[0349] or,

[0350] The acquisition module 301 is specifically used to acquire the business type information of the target metadata table based on K metadata features through a random forest model. The random forest model is a metadata prediction model, which includes at least two decision trees, each of which is used to predict a classification result.

[0351] or,

[0352] The acquisition module 301 is specifically used to obtain the business type of the target metadata table based on K metadata features through an extensible boosting tree model;

[0353] or,

[0354] The acquisition module 301 is specifically used to obtain the first feature vector based on K metadata features and through the input layer included in the metadata prediction model;

[0355] Based on the first feature vector, a second feature vector is obtained through at least one hidden layer included in the metadata prediction model;

[0356] Based on the second feature vector, the probability distribution is obtained through the output layer of the metadata prediction model;

[0357] The business type information of the target metadata table is determined based on the probability distribution.

[0358] Optionally, in the above Figure 10 Based on the corresponding embodiments, in another embodiment of the metadata information generation device 30 provided in this application, the existing metadata also includes at least one of the business type and business sensitivity.

[0359] The processing module 302 is specifically used to perform one-hot encoding on the metadata information corresponding to the business type if the existing metadata includes the business type, so as to obtain the metadata features of the business type, wherein the metadata features of the business type belong to K metadata features.

[0360] If the existing metadata includes business-sensitive cases, then the metadata information corresponding to the business-sensitive cases is processed by one-hot encoding to obtain the metadata features of the business-sensitive cases. The metadata features of the business-sensitive cases belong to K metadata features.

[0361] The acquisition module 301 is specifically used to acquire business importance level information for the target metadata table based on K metadata features and through a metadata prediction model.

[0362] Optionally, in the above Figure 10 Based on the corresponding embodiments, in another embodiment of the metadata information generation device 30 provided in this application,

[0363] The acquisition module 301 is specifically used to acquire business importance level information for the target metadata table based on K metadata features through a decision tree model. The decision tree model is a metadata prediction model, which is a network based on the relationship between nodes, with each node corresponding to a different metadata feature.

[0364] or,

[0365] The acquisition module 301 is specifically used to acquire business importance level information for the target metadata table based on K metadata features through a random forest model. The random forest model is a metadata prediction model, which includes at least two decision trees, each of which is used to predict a classification result.

[0366] or,

[0367] The acquisition module 301 is specifically used to obtain the business importance level of the target metadata table based on K metadata features through an extensible boosting tree model;

[0368] or,

[0369] The acquisition module 301 is specifically used to obtain the first feature vector based on K metadata features and through the input layer included in the metadata prediction model;

[0370] Based on the first feature vector, a second feature vector is obtained through at least one hidden layer included in the metadata prediction model;

[0371] Based on the second feature vector, the probability distribution is obtained through the output layer of the metadata prediction model;

[0372] The business importance level information of the target metadata table is determined based on the probability distribution.

[0373] Optionally, in the above Figure 10 Based on the corresponding embodiments, in another embodiment of the metadata information generation device 30 provided in this application, the existing metadata also includes at least one of the business type and the business importance level;

[0374] The processing module 302 is specifically used to perform one-hot encoding on the metadata information corresponding to the business type if the existing metadata includes the business type, so as to obtain the metadata features of the business type, wherein the metadata features of the business type belong to K metadata features.

[0375] If the existing metadata includes the business importance level, then the metadata information corresponding to the business importance level is one-hot encoded to obtain the metadata features of the business importance level. The metadata features of the business importance level belong to K metadata features.

[0376] The acquisition module 301 is specifically used to obtain business sensitivity information for the target metadata table based on K metadata features and through a metadata prediction model.

[0377] Optionally, in the above Figure 10 Based on the corresponding embodiments, in another embodiment of the metadata information generation device 30 provided in this application,

[0378] The acquisition module 301 is specifically used to acquire business sensitivity information for the target metadata table based on K metadata features through a decision tree model. The decision tree model is a metadata prediction model, which is a network based on the relationship between nodes, with each node corresponding to a different metadata feature.

[0379] or,

[0380] The acquisition module 301 is specifically used to acquire business-sensitive information about the target metadata table based on K metadata features through a random forest model. The random forest model is a metadata prediction model, which includes at least two decision trees, each of which is used to predict a classification result.

[0381] or,

[0382] The acquisition module 301 is specifically used to acquire business-sensitive information about the target metadata table based on K metadata features and through an extensible boosting tree model.

[0383] or,

[0384] The acquisition module 301 is specifically used to obtain the first feature vector based on K metadata features and through the input layer included in the metadata prediction model;

[0385] Based on the first feature vector, a second feature vector is obtained through at least one hidden layer included in the metadata prediction model;

[0386] Based on the second feature vector, the probability distribution is obtained through the output layer of the metadata prediction model;

[0387] The business sensitivity information of the target metadata table is determined based on the probability distribution.

[0388] Optionally, in the above Figure 10 Based on the corresponding embodiments, in another embodiment of the metadata information generation device 30 provided in this application, the metadata information generation device 30 further includes a generation module 303 and a recording module 304;

[0389] The generation module 303 is used to generate K metadata records based on the K metadata features and the K metadata information after the acquisition module 301 obtains the target metadata information for the target metadata table through the metadata prediction model.

[0390] The generation module 303 is also used to generate a target metadata record based on the missing metadata and the target metadata information;

[0391] The recording module 304 is used to record K metadata records and target metadata to the blockchain for synchronous storage. The blockchain is used to provide metadata information of the data table for at least two node devices.

[0392] Optionally, in the above Figure 10 Based on the corresponding embodiments, in another embodiment of the metadata information generation device 30 provided in this application, the metadata information generation device 30 further includes a display module 305;

[0393] The display module 305 is used to display the target metadata information of the target metadata table after the acquisition module 301 obtains the target metadata information of the target metadata table based on K metadata features and through the metadata prediction model.

[0394] or,

[0395] The target metadata information of the target metadata table is sent to the terminal device so that the terminal device can display the target metadata information of the target metadata table.

[0396] This application also provides another metadata information generation device, which can be deployed on a terminal device, such as... Figure 11 As shown, for ease of explanation, only the parts related to the embodiments of this application are shown. For specific technical details not disclosed, please refer to the method section of the embodiments of this application. In the embodiments of this application, a smartphone is used as an example for illustration:

[0397] Figure 11 This is a block diagram illustrating a portion of the structure of a smartphone related to the terminal device provided in the embodiments of this application. (Reference) Figure 11The smartphone includes components such as a radio frequency (RF) circuit 410, a memory 420, an input unit 430, a display unit 440, a sensor 450, an audio circuit 460, a wireless fidelity (WiFi) module 470, a processor 480, and a power supply 490. Those skilled in the art will understand that... Figure 4 The smartphone structure shown does not constitute a limitation on smartphones and may include more or fewer components than shown, or combine certain components, or have different component arrangements.

[0398] The following is combined with Figure 11 A detailed introduction to the various components of a smartphone:

[0399] RF circuit 410 can be used for receiving and transmitting signals during information transmission or calls. Specifically, it receives downlink information from the base station and processes it with processor 480; additionally, it transmits uplink data to the base station. Typically, RF circuit 410 includes, but is not limited to, an antenna, at least one amplifier, a transceiver, a coupler, a low-noise amplifier (LNA), a duplexer, etc. Furthermore, RF circuit 410 can also communicate wirelessly with networks and other devices. The aforementioned wireless communication can use any communication standard or protocol, including but not limited to Global System for Mobile Communication (GSM), General Packet Radio Service (GPRS), Code Division Multiple Access (CDMA), Wideband Code Division Multiple Access (WCDMA), Long Term Evolution (LTE), email, Short Message Service (SMS), etc.

[0400] The memory 420 can be used to store software programs and modules. The processor 480 executes various functions and data processing of the smartphone by running the software programs and modules stored in the memory 420. The memory 420 may mainly include a program storage area and a data storage area. The program storage area may store the operating system, applications required for at least one function (such as sound playback function, image playback function, etc.), etc.; the data storage area may store data created according to the use of the smartphone (such as audio data, phonebook, etc.). In addition, the memory 420 may include high-speed random access memory, and may also include non-volatile memory, such as at least one disk storage device, flash memory device, or other volatile solid-state storage device.

[0401] The input unit 430 can be used to receive input numerical or character information, and to generate key signal inputs related to user settings and function control of the smartphone. Specifically, the input unit 430 may include a touch panel 431 and other input devices 432. The touch panel 431, also known as a touch screen, can collect touch operations performed by the user on or near it (such as operations performed by the user using a finger, stylus, or any suitable object or accessory on or near the touch panel 431), and drive the corresponding connected devices according to a pre-set program. Optionally, the touch panel 431 may include two parts: a touch detection device and a touch controller. The touch detection device detects the user's touch position and the signal generated by the touch operation, and transmits the signal to the touch controller; the touch controller receives touch information from the touch detection device, converts it into touch point coordinates, and sends it to the processor 480, and can also receive and execute commands sent by the processor 480. In addition, the touch panel 431 can be implemented using various types such as resistive, capacitive, infrared, and surface acoustic wave. In addition to the touch panel 431, the input unit 430 may also include other input devices 432. Specifically, other input devices 432 may include, but are not limited to, one or more of the following: physical keyboard, function keys (such as volume control buttons, power buttons, etc.), trackball, mouse, joystick, etc.

[0402] Display unit 440 can be used to display information input by the user or information provided to the user, as well as various menus of the smartphone. Display unit 440 may include display panel 441, optionally configured as a liquid crystal display (LCD), organic light-emitting diode (OLED), or similar form. Further, touch panel 431 may cover display panel 441. When touch panel 431 detects a touch operation on or near it, it transmits the information to processor 480 to determine the type of touch event. Subsequently, processor 480 provides corresponding visual output on display panel 441 based on the type of touch event. Although in Figure 4 In this embodiment, the touch panel 431 and the display panel 441 are two separate components to realize the input and output functions of the smartphone. However, in some embodiments, the touch panel 431 and the display panel 441 can be integrated to realize the input and output functions of the smartphone.

[0403] The smartphone may also include at least one sensor 450, such as a light sensor, a motion sensor, and other sensors. Specifically, the light sensor may include an ambient light sensor and a proximity sensor, wherein the ambient light sensor can adjust the brightness of the display panel 441 according to the ambient light level, and the proximity sensor can turn off the display panel 441 and / or the backlight when the smartphone is moved to the ear. As a type of motion sensor, an accelerometer sensor can detect the magnitude of acceleration in various directions (generally three axes), and can detect the magnitude and direction of gravity when stationary. It can be used for applications that recognize the smartphone's posture (such as landscape / portrait switching, related games, magnetometer posture calibration), vibration recognition-related functions (such as pedometer, tapping), etc. Other sensors that may be configured in the smartphone, such as gyroscopes, barometers, hygrometers, thermometers, and infrared sensors, will not be described in detail here.

[0404] Audio circuit 460, speaker 461, and microphone 462 provide an audio interface between the user and the smartphone. Audio circuit 460 converts received audio data into electrical signals and transmits them to speaker 461, where speaker 461 converts them into sound signals for output. On the other hand, microphone 462 converts collected sound signals into electrical signals, which are received by audio circuit 460, converted into audio data, and then processed by processor 480 before being transmitted via RF circuit 410 to, for example, another smartphone, or the audio data can be output to memory 420 for further processing.

[0405] WiFi is a short-range wireless transmission technology. Smartphones, through their WiFi modules (470), can help users send and receive emails, browse web pages, and access streaming media, providing wireless broadband internet access. Although Figure 11 WiFi module 470 is shown, but it is understood that it is not an essential component of a smartphone and can be omitted as needed without changing the nature of the invention.

[0406] The processor 480 is the control center of the smartphone, connecting various parts of the smartphone through various interfaces and lines. It performs various functions and processes data by running or executing software programs and / or modules stored in the memory 420, and by calling data stored in the memory 420, thereby providing overall monitoring of the smartphone. Optionally, the processor 480 may include one or more processing units; optionally, the processor 480 may integrate an application processor and a modem processor, wherein the application processor mainly handles the operating system, user interface, and applications, and the modem processor mainly handles wireless communication. It is understood that the aforementioned modem processor may also not be integrated into the processor 480.

[0407] The smartphone also includes a power supply 490 (such as a battery) that powers various components. Optionally, the power supply can be logically connected to the processor 480 through a power management system, thereby enabling functions such as charging, discharging, and power consumption management through the power management system.

[0408] Although not shown, smartphones may also include a camera, Bluetooth module, etc., which will not be described in detail here.

[0409] The steps performed by the terminal device in the above embodiments can be based on this Figure 11 The terminal device structure is shown.

[0410] This application also provides another metadata information generation device, which can be deployed on a server. Figure 12This is a schematic diagram of a server structure provided in an embodiment of this application. The server 500 can vary significantly due to different configurations or performance. It may include one or more central processing units (CPUs) 522 (e.g., one or more processors) and memory 532, and one or more storage media 530 (e.g., one or more mass storage devices) for storing application programs 542 or data 544. The memory 532 and storage media 530 can be temporary or persistent storage. The program stored in the storage media 530 may include one or more modules (not shown in the diagram), each module may include a series of instruction operations on the server. Furthermore, the CPU 522 may be configured to communicate with the storage media 530 and execute the series of instruction operations in the storage media 530 on the server 500.

[0411] Server 500 may also include one or more power supplies 526, one or more wired or wireless network interfaces 550, one or more input / output interfaces 558, and / or one or more operating systems 541, such as Windows Server. TM Mac OS X TM Unix TM Linux TM FreeBSD TM etc.

[0412] The steps performed by the server in the above embodiments can be based on this Figure 12 The server structure shown.

[0413] This application also provides a computer-readable storage medium storing a computer program that, when run on a computer, causes the computer to perform the methods described in the foregoing embodiments.

[0414] This application also provides a computer program product including a program, which, when run on a computer, causes the computer to perform the methods described in the foregoing embodiments.

[0415] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0416] In the several embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and 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 an indirect coupling or communication connection between apparatuses or units through some interfaces, and may be electrical, mechanical, or other forms.

[0417] 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.

[0418] Furthermore, the functional units in the various embodiments of this application 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. The integrated unit can be implemented in hardware or as a software functional unit.

[0419] If the integrated unit 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 application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. 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.

[0420] The above-described embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application.

Claims

1. A method for generating metadata information, characterized in that, include: Obtain existing metadata of the target metadata table, wherein the existing metadata is metadata that already exists in the target metadata table, and the existing metadata includes K metadata, where K is an integer greater than or equal to 1, and the existing metadata includes at least one of the following: number of downstream dependent tasks, number of upstream dependent tasks, access frequency, online business service status, number of online services, number of fields, number of data rows, and storage size. K metadata information are obtained based on the existing metadata, wherein the metadata information has a one-to-one correspondence with the metadata; The K metadata information is subjected to feature processing to obtain K metadata features, wherein the metadata features have a one-to-one correspondence with the metadata information; Based on the K metadata features, target metadata information for the target metadata table is obtained through a metadata prediction model, wherein the target metadata information is the metadata information corresponding to the missing metadata, and the missing metadata is metadata that does not exist in the target metadata table; The step of performing feature processing on the K metadata information to obtain K metadata features includes: If the existing metadata includes the number of downstream dependent tasks, then the metadata information of the number of downstream dependent tasks is binned to obtain the metadata features of the number of downstream dependent tasks, wherein the metadata features of the number of downstream dependent tasks belong to the K metadata features. If the existing metadata includes the number of upstream dependent tasks, then the metadata information of the number of upstream dependent tasks is binned to obtain the metadata features of the number of upstream dependent tasks, wherein the metadata features of the number of upstream dependent tasks belong to the K metadata features. If the existing metadata includes the access frequency, then the access frequency is binned to obtain the metadata features of the access frequency, wherein the metadata features of the access frequency belong to the K metadata features. If the existing metadata includes the online business service information, then the online business service information is subjected to one-hot encoding to obtain the metadata features of the online business service information, wherein the metadata features of the online business service information belong to the K metadata features; If the existing metadata includes the number of online services, feature scaling processing is performed on the number of online services to obtain the metadata features of the number of online services, wherein the metadata features of the number of online services belong to the K metadata features; If the existing metadata includes the number of fields, feature scaling is performed on the number of fields to obtain the metadata feature of the number of fields, wherein the metadata feature of the number of fields belongs to the K metadata features; If the existing metadata includes the number of data rows, feature scaling is performed on the number of data rows to obtain the metadata features of the number of data rows, wherein the metadata features of the number of data rows belong to the K metadata features; If the existing metadata includes the storage size, feature scaling is performed on the storage size to obtain the metadata features of the storage size, wherein the metadata features of the storage size belong to the K metadata features.

2. The generation method according to claim 1, characterized in that, The K metadata information includes table name text information, wherein the table name text information belongs to a preset language; The step of performing feature processing on the K metadata information to obtain K metadata features includes: The table name text information is segmented to obtain a table name text sequence, wherein the table name text sequence includes M words, where M is an integer greater than or equal to 1; Each word in the table name text sequence is preprocessed to obtain M word vector embeddings, wherein the word vector embeddings have a one-to-one correspondence with the words, and the M word vector embeddings belong to the metadata features of the table name text information; The step of obtaining target metadata information based on the K metadata features through a metadata prediction model includes: Based on the embedding of the M word vectors, the semantic information vector corresponding to the table name text sequence is obtained through the encoder included in the metadata prediction model; Based on the semantic information vector, table description information is obtained through the decoder included in the metadata prediction model, wherein the table description information belongs to the target metadata information.

3. The generation method according to claim 2, characterized in that, The step of obtaining K metadata information based on the existing metadata includes: If the existing metadata includes a table name, then the table name text information corresponding to the table name is obtained from the target metadata table; or, The step of obtaining K metadata information based on the existing metadata includes: If the existing metadata does not include the data table name, then the associated metadata table is obtained according to the data lineage included in the existing metadata, wherein the associated metadata table is an upstream metadata table or a downstream metadata table; Obtain the table name text information corresponding to the table name from the associated metadata table.

4. The generation method according to claim 1, characterized in that, The existing metadata also includes at least one of the following: business importance level and business sensitivity. The step of performing feature processing on the K metadata information to obtain K metadata features includes: If the existing metadata includes the business importance level, then the metadata information corresponding to the business importance level is subjected to one-hot encoding to obtain the metadata features of the business importance level, wherein the metadata features of the business importance level belong to the K metadata features; If the existing metadata includes the business-sensitive case, then the metadata information corresponding to the business-sensitive case is subjected to one-hot encoding to obtain the metadata features of the business-sensitive case, wherein the metadata features of the business-sensitive case belong to the K metadata features; The step of obtaining target metadata information for the target metadata table based on the K metadata features through a metadata prediction model includes: Based on the K metadata features, the business type information of the target metadata table is obtained through the metadata prediction model.

5. The generation method according to claim 4, characterized in that, The step of obtaining the business type information of the target metadata table based on the K metadata features and through the metadata prediction model includes: Based on the K metadata features, the business type information of the target metadata table is obtained through a decision tree model. The decision tree model belongs to the metadata prediction model. The decision tree model is a network constructed according to the relationship between nodes, and each node corresponds to a different metadata feature. or, The step of obtaining the business type information of the target metadata table based on the K metadata features and through the metadata prediction model includes: Based on the K metadata features, the business type information of the target metadata table is obtained through a random forest model. The random forest model is a metadata prediction model, and the random forest model includes at least two decision trees, each of which is used to predict a classification result. or, The step of obtaining the business type information of the target metadata table based on the K metadata features and through the metadata prediction model includes: Based on the K metadata features, the business type of the target metadata table is obtained through an extensible boosting tree model; or, The step of obtaining the business type information of the target metadata table based on the K metadata features and through the metadata prediction model includes: Based on the K metadata features, a first feature vector is obtained through the input layer included in the metadata prediction model; Based on the first feature vector, a second feature vector is obtained through at least one hidden layer included in the metadata prediction model; Based on the second feature vector, the probability distribution is obtained through the output layer included in the metadata prediction model; The business type information of the target metadata table is determined based on the probability distribution.

6. The generation method according to claim 1, characterized in that, The existing metadata also includes at least one of the business type and business sensitivity conditions; The step of performing feature processing on the K metadata information to obtain K metadata features includes: If the existing metadata includes the business type to which it belongs, then the metadata information corresponding to the business type to which it belongs is one-hot encoded to obtain the metadata feature of the business type to which it belongs, wherein the metadata feature of the business type to which it belongs belongs to the K metadata features; If the existing metadata includes the business-sensitive case, then the metadata information corresponding to the business-sensitive case is subjected to one-hot encoding to obtain the metadata features of the business-sensitive case, wherein the metadata features of the business-sensitive case belong to the K metadata features; The step of obtaining target metadata information for the target metadata table based on the K metadata features through a metadata prediction model includes: Based on the K metadata features, the business importance level information for the target metadata table is obtained through the metadata prediction model.

7. The generation method according to claim 6, characterized in that, The step of obtaining business importance level information for the target metadata table based on the K metadata features and through the metadata prediction model includes: Based on the K metadata features, business importance level information for the target metadata table is obtained through a decision tree model. The decision tree model belongs to the metadata prediction model. The decision tree model is a network constructed according to the relationship between nodes, and each node corresponds to a different metadata feature. or, The step of obtaining business importance level information for the target metadata table based on the K metadata features and through the metadata prediction model includes: Based on the K metadata features, business importance level information for the target metadata table is obtained through a random forest model. The random forest model is a metadata prediction model, and the random forest model includes at least two decision trees, each of which is used to predict a classification result. or, The step of obtaining business importance level information for the target metadata table based on the K metadata features and through the metadata prediction model includes: Based on the K metadata features, the business importance level of the target metadata table is obtained through an extensible boosting tree model; or, The step of obtaining business importance level information for the target metadata table based on the K metadata features and through the metadata prediction model includes: Based on the K metadata features, a first feature vector is obtained through the input layer included in the metadata prediction model; Based on the first feature vector, a second feature vector is obtained through at least one hidden layer included in the metadata prediction model; Based on the second feature vector, the probability distribution is obtained through the output layer included in the metadata prediction model; The business importance level information of the target metadata table is determined based on the probability distribution.

8. The generation method according to claim 1, characterized in that, The existing metadata also includes at least one of the business type and the business importance level; The step of performing feature processing on the K metadata information to obtain K metadata features includes: If the existing metadata includes the business type to which it belongs, then the metadata information corresponding to the business type to which it belongs is one-hot encoded to obtain the metadata feature of the business type to which it belongs, wherein the metadata feature of the business type to which it belongs belongs to the K metadata features; If the existing metadata includes the business importance level, then the metadata information corresponding to the business importance level is subjected to one-hot encoding to obtain the metadata features of the business importance level, wherein the metadata features of the business importance level belong to the K metadata features; The step of obtaining target metadata information for the target metadata table based on the K metadata features through a metadata prediction model includes: Based on the K metadata features, business sensitivity information for the target metadata table is obtained through the metadata prediction model.

9. The generation method according to claim 8, characterized in that, The step of obtaining business sensitivity information for the target metadata table based on the K metadata features and through the metadata prediction model includes: Based on the K metadata features, business sensitivity information for the target metadata table is obtained through a decision tree model. The decision tree model is a metadata prediction model, which is a network based on the relationship between nodes, with each node corresponding to a different metadata feature. or, The step of obtaining business sensitivity information for the target metadata table based on the K metadata features and through the metadata prediction model includes: Based on the K metadata features, business sensitivity information for the target metadata table is obtained through a random forest model. The random forest model is a metadata prediction model, and the random forest model includes at least two decision trees, each of which is used to predict a classification result. or, The step of obtaining business sensitivity information for the target metadata table based on the K metadata features and through the metadata prediction model includes: Based on the K metadata features, business sensitivity information for the target metadata table is obtained through an extensible boosting tree model; or, The step of obtaining business sensitivity information for the target metadata table based on the K metadata features and through the metadata prediction model includes: Based on the K metadata features, a first feature vector is obtained through the input layer included in the metadata prediction model; Based on the first feature vector, a second feature vector is obtained through at least one hidden layer included in the metadata prediction model; Based on the second feature vector, the probability distribution is obtained through the output layer included in the metadata prediction model; The business sensitivity information of the target metadata table is determined based on the probability distribution.

10. The generation method according to any one of claims 1 to 9, characterized in that, After obtaining the target metadata information for the target metadata table based on the K metadata features through a metadata prediction model, the method further includes: Based on the K metadata and the K metadata information, generate K metadata records; Based on the missing metadata and the target metadata information, a target metadata record is generated; The K metadata records and the target metadata records are synchronously stored on the blockchain, wherein the blockchain is used to provide metadata information of the data table for at least two node devices.

11. The generation method according to any one of claims 1 to 9, characterized in that, After obtaining the target metadata information for the target metadata table based on the K metadata features through a metadata prediction model, the method further includes: Display the target metadata information in the target metadata table; or, The target metadata information of the target metadata table is sent to the terminal device so that the terminal device displays the target metadata information of the target metadata table.

12. A metadata information generation device, characterized in that, include: The acquisition module is used to acquire existing metadata of the target metadata table. The existing metadata is metadata that already exists in the target metadata table, and the existing metadata includes K metadata, where K is an integer greater than or equal to 1. The existing metadata includes at least one of the following: number of downstream dependent tasks, number of upstream dependent tasks, access frequency, online business service status, number of online services, number of fields, number of data rows, and storage size. The acquisition module is further configured to acquire K metadata information based on the existing metadata, wherein the metadata information has a one-to-one correspondence with the metadata; The processing module is used to perform feature processing on the K metadata information to obtain K metadata features, wherein the metadata features have a one-to-one correspondence with the metadata information; The acquisition module is further configured to acquire target metadata information for the target metadata table based on the K metadata features through a metadata prediction model, wherein the target metadata information is the metadata information corresponding to the missing metadata, and the missing metadata is metadata that does not exist in the target metadata table; Specifically, the processing module is used for: If the existing metadata includes the number of downstream dependent tasks, then the metadata information of the number of downstream dependent tasks is binned to obtain the metadata features of the number of downstream dependent tasks, wherein the metadata features of the number of downstream dependent tasks belong to the K metadata features. If the existing metadata includes the number of upstream dependent tasks, then the metadata information of the number of upstream dependent tasks is binned to obtain the metadata features of the number of upstream dependent tasks, wherein the metadata features of the number of upstream dependent tasks belong to the K metadata features. If the existing metadata includes the access frequency, then the access frequency is binned to obtain the metadata features of the access frequency, wherein the metadata features of the access frequency belong to the K metadata features. If the existing metadata includes the online business service information, then the online business service information is subjected to one-hot encoding to obtain the metadata features of the online business service information, wherein the metadata features of the online business service information belong to the K metadata features; If the existing metadata includes the number of online services, feature scaling processing is performed on the number of online services to obtain the metadata features of the number of online services, wherein the metadata features of the number of online services belong to the K metadata features; If the existing metadata includes the number of fields, feature scaling is performed on the number of fields to obtain the metadata feature of the number of fields, wherein the metadata feature of the number of fields belongs to the K metadata features; If the existing metadata includes the number of data rows, feature scaling is performed on the number of data rows to obtain the metadata features of the number of data rows, wherein the metadata features of the number of data rows belong to the K metadata features; If the existing metadata includes the storage size, feature scaling is performed on the storage size to obtain the metadata features of the storage size, wherein the metadata features of the storage size belong to the K metadata features.

13. The apparatus according to claim 12, characterized in that, The K metadata information includes table name text information, wherein the table name text information belongs to a preset language; The processing unit is specifically used for: The table name text information is segmented to obtain a table name text sequence, wherein the table name text sequence includes M words, where M is an integer greater than or equal to 1; Each word in the table name text sequence is preprocessed to obtain M word vector embeddings, wherein the word vector embeddings have a one-to-one correspondence with the words, and the M word vector embeddings belong to the metadata features of the table name text information; The acquisition module is specifically used for: Based on the embedding of the M word vectors, the semantic information vector corresponding to the table name text sequence is obtained through the encoder included in the metadata prediction model; Based on the semantic information vector, table description information is obtained through the decoder included in the metadata prediction model, wherein the table description information belongs to the target metadata information.

14. The apparatus according to claim 13, characterized in that, The acquisition module is specifically used for: If the existing metadata includes a table name, then the table name text information corresponding to the table name is obtained from the target metadata table; or, The step of obtaining K metadata information based on the existing metadata includes: If the existing metadata does not include the data table name, then the associated metadata table is obtained according to the data lineage included in the existing metadata, wherein the associated metadata table is an upstream metadata table or a downstream metadata table; Obtain the table name text information corresponding to the table name from the associated metadata table.

15. The apparatus according to claim 12, characterized in that, The existing metadata also includes at least one of the following: business importance level and business sensitivity. The processing module is specifically used for: If the existing metadata includes the business importance level, then the metadata information corresponding to the business importance level is subjected to one-hot encoding to obtain the metadata features of the business importance level, wherein the metadata features of the business importance level belong to the K metadata features; If the existing metadata includes the business-sensitive case, then the metadata information corresponding to the business-sensitive case is subjected to one-hot encoding to obtain the metadata features of the business-sensitive case, wherein the metadata features of the business-sensitive case belong to the K metadata features; The acquisition module is specifically used for: Based on the K metadata features, the business type information of the target metadata table is obtained through the metadata prediction model.

16. The apparatus according to claim 15, characterized in that, The acquisition module is specifically used to acquire the business type information of the target metadata table based on the K metadata features through a decision tree model. The decision tree model belongs to the metadata prediction model and is a network constructed according to the relationship between nodes, with each node corresponding to a different metadata feature. or, The acquisition module is specifically used to acquire the business type information of the target metadata table based on the K metadata features through a random forest model. The random forest model belongs to the metadata prediction model and includes at least two decision trees, each of which is used to predict a classification result. or, The acquisition module is specifically used to obtain the business type of the target metadata table based on the K metadata features through an extensible boosting tree model; or, The acquisition module is specifically used for: Based on the K metadata features, a first feature vector is obtained through the input layer included in the metadata prediction model; Based on the first feature vector, a second feature vector is obtained through at least one hidden layer included in the metadata prediction model; Based on the second feature vector, the probability distribution is obtained through the output layer included in the metadata prediction model; The business type information of the target metadata table is determined based on the probability distribution.

17. The apparatus according to claim 12, characterized in that, The existing metadata also includes at least one of the business type and business sensitivity conditions; The processing module is specifically used for: If the existing metadata includes the business type to which it belongs, then the metadata information corresponding to the business type to which it belongs is one-hot encoded to obtain the metadata feature of the business type to which it belongs, wherein the metadata feature of the business type to which it belongs belongs to the K metadata features; If the existing metadata includes the business-sensitive case, then the metadata information corresponding to the business-sensitive case is subjected to one-hot encoding to obtain the metadata features of the business-sensitive case, wherein the metadata features of the business-sensitive case belong to the K metadata features; The acquisition module is specifically used for: Based on the K metadata features, the business importance level information for the target metadata table is obtained through the metadata prediction model.

18. The apparatus according to claim 17, characterized in that, The acquisition module is specifically used to acquire business importance level information for the target metadata table based on the K metadata features through a decision tree model. The decision tree model belongs to the metadata prediction model, and the decision tree model is a network constructed according to the relationship between nodes, with each node corresponding to different metadata features. or, The acquisition module is specifically used to acquire business importance level information for the target metadata table based on the K metadata features through a random forest model. The random forest model is a metadata prediction model, and the random forest model includes at least two decision trees, each of which is used to predict a classification result. or, The acquisition module is specifically used to obtain the business importance level of the target metadata table based on the K metadata features through an extensible boosting tree model; or, The acquisition module is specifically used for: Based on the K metadata features, a first feature vector is obtained through the input layer included in the metadata prediction model; Based on the first feature vector, a second feature vector is obtained through at least one hidden layer included in the metadata prediction model; Based on the second feature vector, the probability distribution is obtained through the output layer included in the metadata prediction model; The business importance level information of the target metadata table is determined based on the probability distribution.

19. The apparatus according to claim 12, characterized in that, The existing metadata also includes at least one of the business type and the business importance level; The processing module is specifically used for: If the existing metadata includes the business type to which it belongs, then the metadata information corresponding to the business type to which it belongs is one-hot encoded to obtain the metadata feature of the business type to which it belongs, wherein the metadata feature of the business type to which it belongs belongs to the K metadata features; If the existing metadata includes the business importance level, then the metadata information corresponding to the business importance level is subjected to one-hot encoding to obtain the metadata features of the business importance level, wherein the metadata features of the business importance level belong to the K metadata features; The acquisition module is specifically used for: Based on the K metadata features, business sensitivity information for the target metadata table is obtained through the metadata prediction model.

20. The apparatus according to claim 19, characterized in that, The acquisition module is specifically used to acquire business sensitivity information for the target metadata table based on the K metadata features through a decision tree model. The decision tree model belongs to the metadata prediction model and is a network constructed according to the relationship between nodes, with each node corresponding to a different metadata feature. or, The acquisition module is specifically used to acquire business sensitivity information for the target metadata table based on the K metadata features through a random forest model. The random forest model is a metadata prediction model, and the random forest model includes at least two decision trees, each of which is used to predict a classification result. or, The acquisition module is specifically used to acquire business sensitivity information for the target metadata table based on the K metadata features through an extensible boosting tree model; or, The acquisition module is specifically used for: Based on the K metadata features, a first feature vector is obtained through the input layer included in the metadata prediction model; Based on the first feature vector, a second feature vector is obtained through at least one hidden layer included in the metadata prediction model; Based on the second feature vector, the probability distribution is obtained through the output layer included in the metadata prediction model; The business sensitivity information of the target metadata table is determined based on the probability distribution.

21. The apparatus according to any one of claims 12 to 20, characterized in that, The device also includes a generation module and a recording module; The generation module is used to generate K metadata records based on the K metadata features and the K metadata information after the acquisition module obtains the target metadata information for the target metadata table through a metadata prediction model. The generation module is further configured to generate a target metadata record based on the missing metadata and the target metadata information; The recording module is used to record the K metadata records and the target metadata to the blockchain for synchronous storage, wherein the blockchain is used to provide metadata information of the data table for at least two node devices.

22. The apparatus according to any one of claims 12 to 20, characterized in that, The device also includes a display module; The display module is used to obtain target metadata information for the target metadata table based on the K metadata features through a metadata prediction model, and then display the target metadata information of the target metadata table. or, The target metadata information of the target metadata table is sent to the terminal device so that the terminal device displays the target metadata information of the target metadata table.

23. A computer device, characterized in that, include: Memory, processor, and bus system; The memory is used to store programs; The processor is configured to execute a program in the memory, and the processor is configured to execute the generation method according to any one of claims 1 to 11 according to the instructions in the program code; The bus system is used to connect the memory and the processor to enable communication between the memory and the processor.

24. A computer-readable storage medium comprising instructions, when executed on a computer, causing the computer to perform the generation method as described in any one of claims 1 to 11.

25. A computer program product, characterized in that, The computer program product includes computer instructions, and the execution of the computer instructions by the computer device causes the computer device to perform the generation method as described in any one of claims 1 to 11.

Images