Business data evaluation method, device, equipment and computer readable storage medium

Abstract

The application provides a business data evaluation method, device and equipment and a computer readable storage medium. The method comprises: obtaining a metadata set generated in a business process; the metadata set represents description information corresponding to technical data and process definition in the business process; performing feature extraction on the metadata set from a business value dimension to obtain a metadata feature set; the business value dimension is a business management and operation state represented by technical metadata and business metadata; performing importance evaluation classification on the metadata feature set using a target multi-classification model to obtain an importance classification result corresponding to the metadata set; wherein the target multi-classification model is obtained by performing importance evaluation classification training on an initial multi-classification model using a feature sample set with labeled classification results; and performing data analysis on the business based on the importance classification result. Through the application, the efficiency and accuracy of business data evaluation can be improved through artificial intelligence technology.

Description

Technical Field

[0001] This application relates to computer technology, and more particularly to a business data evaluation method, apparatus, device, and computer-readable storage medium. Background Technology

[0002] As the data management industry matures and develops, the importance of data assets to enterprises is becoming increasingly prominent. Currently, companies rely on data assets to make smarter and more effective decisions, provide better products and services, reduce costs, and control risks. During business development, companies generate a large number of data tables. With the increasing volume and variety of data in these tables, effective methods are needed to evaluate these massive datasets and extract the more important and valuable data assets for decision-makers to analyze.

[0003] However, current methods either rely on manual importance assessment, resulting in a large workload and low efficiency in business data evaluation, or are limited to simple classification and evaluation of the importance of data assets through established rules or single indicators, thereby reducing the efficiency and accuracy of business data evaluation. Summary of the Invention

[0004] This application provides a business data evaluation method, apparatus, and computer-readable storage medium, which can improve the efficiency and accuracy of business data evaluation through artificial intelligence technology.

[0005] The technical solution of this application embodiment is implemented as follows:

[0006] This application provides a business data evaluation method, including:

[0007] Acquire the metadata generated during the business process; the metadata represents the descriptive information corresponding to the technical data and process definitions during the business process.

[0008] Feature extraction is performed on the metadata dataset from the business value dimension to obtain a metadata feature set; the business value dimension is the business management and operation status represented by the technical metadata and the business metadata.

[0009] The metadata feature set is classified and evaluated for importance using a target multi-classification model to obtain the importance ranking results corresponding to the metadata dataset; wherein...

[0010] The target multi-class classification model is obtained by training the initial multi-class classification model with importance evaluation using a feature sample set with labeled classification results; the feature sample set contains metadata features of multiple sample classes;

[0011] Based on the importance ranking results, data analysis of the business can be performed.

[0012] This application provides a business data evaluation apparatus, including:

[0013] The data acquisition module is used to acquire metadata generated during the business process; the metadata represents the descriptive information corresponding to the technical data and process definitions during the business process.

[0014] The feature extraction module is used to extract features from the metadata dataset from the business value dimension to obtain a metadata feature set; the business value dimension is the business management and operation status represented by the technical metadata and the business metadata.

[0015] The model calculation module is used to evaluate and classify the metadata feature set using a target multi-classification model, obtaining the importance ranking results corresponding to the metadata dataset; wherein...

[0016] The target multi-class classification model is obtained by training an initial multi-class classification model using a feature sample set with labeled classification results, and the feature sample set contains metadata features of multiple sample classes.

[0017] The data analysis module is used to perform data analysis on the business based on the importance classification results.

[0018] In the above-described apparatus, the feature extraction module is further configured to extract technical metadata features from the technical metadata; the technical metadata features represent technical statistical indicators corresponding to business scale and business access volume; extract business metadata features from the business metadata; the business metadata features represent descriptive statistical indicators corresponding to business attributes and business process definitions; and perform at least one feature processing, namely feature normalization, feature transformation, and feature selection, on the technical metadata features and the business metadata features respectively to obtain the metadata feature set.

[0019] In the aforementioned device, the technical metadata includes at least one of the following: data source information, data transformation description information, object definition, data structure definition, rules used for data cleaning and data updating, and mapping from source data to destination data; the business metadata includes at least one of the following: business name, business definition, business description information, and business attributes; the technical metadata features include at least one of the following: number of downstream dependent tasks, number of upstream dependent tasks, access frequency, whether it serves online business, number of online businesses served, number of fields, number of data rows, and storage size; the business metadata features include at least one of the following: business category, business star rating, and whether it involves sensitive data.

[0020] In the above-mentioned device, the target multi-classification model is an ensemble learning model, which includes at least one classification tree. The model calculation module is further configured to use each of the at least one classification tree to perform importance assessment and classification on different metadata features in the metadata feature set, thereby obtaining a sub-classification result corresponding to each classification tree. The sub-classification result belongs to at least one preset importance level. Based on each sub-classification result, a summary vote is performed, and the importance classification result is determined based on the voting result.

[0021] In the aforementioned apparatus, the business data evaluation apparatus further includes a model training module. This model training module is used to obtain an original feature set before using a target multi-classification model to evaluate and classify the metadata feature set to obtain the importance ranking result corresponding to the metadata dataset. The original feature set contains multiple original metadata features with labeled classification results. Feature processing is performed on the original feature set to obtain the feature sample set. The initial multi-classification model is trained using the feature sample set to obtain the target multi-classification model.

[0022] In the above-described apparatus, the model training module is further configured to process the original feature set using at least one of the following methods: feature normalization, feature transformation, and feature selection, thereby obtaining the feature sample set.

[0023] In the above-described apparatus, the model training module is further configured to: select a preset first number of feature samples from the feature sample set as the current training set in the current training iteration; use the initial multi-classification model to perform classification prediction on the current training set to obtain a predicted classification result; iteratively update the initial multi-classification model based on the predicted classification result and the labeled classification result until a preset training objective is reached, then end the training to obtain a trained multi-classification model; select a preset second number of feature samples from the feature sample set as a test set; use the trained multi-classification model to perform classification prediction on the test set to obtain a test classification result; use a preset evaluation method to evaluate the classification effect of the test classification result to obtain an evaluation result; and when the evaluation result is passed, use the trained multi-classification model as the target multi-classification model.

[0024] In the above apparatus, the initial multi-class model includes: an initial classification tree. The model training module is further configured to: when the current training is the first training iteration, use the initial classification tree to perform classification prediction on the current training set to obtain the classification prediction value corresponding to the initial classification tree; when the current training is not the first training iteration, construct the current classification tree corresponding to the current training based on the classification prediction value corresponding to the previous classification tree and the labeled classification result; use the current classification tree to perform classification prediction on the current training set to obtain the classification prediction value corresponding to the current classification tree; continue the above iterative process until a preset training objective is reached; and use at least one classification tree obtained during the training process as the trained multi-class model.

[0025] This application provides an electronic device, including:

[0026] Memory, used to store executable instructions;

[0027] The processor, when executing executable instructions stored in the memory, implements the business data evaluation method provided in the embodiments of this application.

[0028] This application provides a computer-readable storage medium storing executable instructions for inducing a processor to execute and implement the business data evaluation method provided in this application.

[0029] The embodiments of this application have the following beneficial effects:

[0030] By utilizing a target multi-classification model to assess the importance of metadata generated from business processes, the assessment speed is improved, thereby increasing the efficiency of business data assessment. Furthermore, since the target multi-classification model is obtained by training an initial multi-classification model using a feature sample set containing multiple metadata features, it can comprehensively and systematically assess metadata datasets based on assessment metrics of multiple metadata features, thus improving the accuracy of business data assessment. Attached Figure Description

[0031] Figure 1 This is an optional structural diagram of the business data evaluation system architecture provided in the embodiments of this application;

[0032] Figure 2 This is an optional structural diagram of the business data evaluation device provided in the embodiments of this application;

[0033] Figure 3 This is an optional flowchart illustrating the business data evaluation method provided in the embodiments of this application;

[0034] Figure 4This is an optional flowchart illustrating the business data evaluation method provided in the embodiments of this application;

[0035] Figure 5 This is an optional flowchart illustrating the business data evaluation method provided in the embodiments of this application;

[0036] Figure 6 This is a schematic diagram illustrating at least one classification tree provided in an embodiment of this application performing classification prediction on a test set;

[0037] Figure 7 This is an optional flowchart illustrating the business data evaluation method provided in the embodiments of this application;

[0038] Figure 8 This is an optional flowchart illustrating the business data evaluation method provided in the embodiments of this application;

[0039] Figure 9 This is an optional flowchart illustrating the practical process of the business data evaluation method provided in the embodiments of this application. Detailed Implementation

[0040] To make the objectives, technical solutions, and advantages of this application clearer, the application will be further described in detail below with reference to the accompanying drawings. The described embodiments should not be regarded as limitations on this application. All other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0041] In the following description, references are made to “some embodiments,” which describe a subset of all possible embodiments. However, it is understood that “some embodiments” may be the same subset or different subsets of all possible embodiments and may be combined with each other without conflict.

[0042] In the following description, the terms "first, second, third" are used merely to distinguish similar objects and do not represent a specific ordering of objects. It is understood that "first, second, third" may be interchanged in a specific order or sequence where permitted, so that the embodiments of this application described herein can be implemented in an order other than that illustrated or described herein.

[0043] In the implementation of this application, the collection and processing of relevant data should strictly comply with the requirements of relevant laws and regulations, obtain the informed consent or separate consent of the personal information subject, and carry out subsequent data use and processing within the scope of laws and regulations and the authorization of the personal information subject.

[0044] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used herein is for the purpose of describing embodiments of this application only and is not intended to limit this application.

[0045] Before providing a further detailed description of the embodiments of this application, the nouns and terms involved in the embodiments of this application will be explained, and the nouns and terms involved in the embodiments of this application shall be interpreted as follows.

[0046] 1) Metadata: Metadata is data that describes the relationships between data. In data application systems, metadata generally refers to data that describes data concepts, relationships between data, and data processing rules. Domain semantics and knowledge also fall under the category of metadata.

[0047] Metadata, also known as data about data, refers to any data used to help identify, describe, and locate online electronic resources. Metadata can be used to describe the content characteristics and other features of the data itself, aiming to enhance the discovery, identification, development, organization, and evaluation of online information resources, as well as the selection, location, and retrieval of relevant information resources, tracking changes in resources during use, and achieving the integration, effective management, and long-term preservation of information resources.

[0048] 2) Metadata Management: As information describing data, metadata helps businesses better understand and apply data through its collection, storage, retrieval, sharing, and openness capabilities. It is also the cornerstone of data asset management, and its types include data dictionaries, data lineage, business metrics, and operational metrics.

[0049] 3) Data Asset Management: This refers to a set of business functions for planning, controlling, and providing data and information assets. It includes developing, implementing, and monitoring plans, policies, programs, projects, processes, methods, and procedures related to data, thereby controlling, protecting, delivering, and enhancing the value of data assets. Through a combination of processes, systems, and technologies, it provides strong support to business teams in terms of data delivery quality, efficiency, cost, and security, enabling data to be rapidly transformed into assets and contributing to business success.

[0050] 4) Data Asset Rating and Tiering: Businesses generate a large number of data tables during their development, but not all data is an asset. Therefore, an evaluation system is needed to comprehensively rate and tier data assets, classifying them into three levels of importance, such as high, medium, and low.

[0051] 5) Metadata Features: Metadata features are specific representations built upon the original metadata. They are individual, measurable attributes, typically represented by columns in the metadata set. For a general two-dimensional metadata dataset, each metadata element is represented by a row, and each metadata feature by a column, with a specific value for each metadata element.

[0052] 6) A database, simply put, can be viewed as an electronic filing cabinet—a place to store electronic files, where users can perform operations such as adding, querying, updating, and deleting data. A "database" is a collection of data stored together in a certain way, capable of being shared by multiple users, with minimal redundancy, and independent of application programs.

[0053] 7) A data warehouse (DW) is a strategic collection of data that supports all types of data in decision-making processes at all levels of an enterprise. It is a single data store created for analytical reporting and decision support purposes. It provides guidance for business process improvement, monitoring of time, cost, quality, and control for enterprises that require business intelligence. A data warehouse is a structured data environment that serves as a data source for decision support systems (DSS) and online analytical applications. Data warehouses study and solve problems related to extracting information from databases. The characteristics of a data warehouse include subject-oriented, integrated, stable, and time-varying nature.

[0054] 8) A Database Management System (DBMS) is a computer software system designed to manage databases, generally possessing basic functions such as storage, retrieval, security, and backup. DBMSs can be classified according to the database model they support, such as relational or XML (Extensible Markup Language); or according to the type of computer they support, such as server clusters or mobile phones; or according to the query language used, such as SQL (Structured Query Language) or XQuery; or according to performance priorities, such as maximum scale or highest operating speed; or other classification methods. Regardless of the classification method used, some DBMSs can cross categories, for example, simultaneously supporting multiple query languages.

[0055] 9) Artificial Intelligence (AI) is the theory, methods, technology, and application systems that use digital computers or machines controlled by digital computers to simulate, extend, and expand human intelligence, perceive the environment, acquire knowledge, and use that knowledge to achieve optimal results. In other words, AI is a comprehensive technology within computer science that attempts to understand the essence of intelligence and produce a new kind of intelligent machine that can react in a way similar to human intelligence. AI studies the design principles and implementation methods of various intelligent machines, enabling them to possess the functions of perception, reasoning, and decision-making.

[0056] Artificial intelligence (AI) is a comprehensive discipline encompassing a wide range of fields, including both hardware and software technologies. Fundamental AI technologies generally include sensors, dedicated AI chips, cloud computing, distributed storage, big data processing, operating / interactive systems, and mechatronics. AI software technologies primarily include computer vision, speech processing, natural language processing, and machine learning / deep learning.

[0057] 10) Machine Learning (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. Machine learning is the core of artificial intelligence and the fundamental way to endow computers with intelligence; its applications span all areas of artificial intelligence. Machine learning and deep learning typically include techniques such as artificial neural networks, belief networks, reinforcement learning, transfer learning, inductive learning, and instructional learning.

[0058] The solutions provided in this application relate to technologies for data asset rating based on artificial intelligence, which are specifically illustrated through the following embodiments.

[0059] This application provides a business data evaluation method, apparatus, device, and computer-readable storage medium, which can improve the efficiency and accuracy of business data evaluation through artificial intelligence technology. The following describes exemplary applications of the electronic devices provided in this application. These electronic devices can be implemented as various types of user terminals such as laptops, tablets, desktop computers, set-top boxes, and mobile devices (e.g., mobile phones, portable music players, personal digital assistants, dedicated messaging devices, portable gaming devices), or as servers. The following describes exemplary applications when the device is implemented as a terminal.

[0060] See Figure 1 , Figure 1This is an optional architecture diagram of the business data evaluation system 100 provided in the embodiments of this application. The terminal connects to the server 200 through the network 300, which can be a wide area network or a local area network, or a combination of the two.

[0061] Terminal 400 is used to acquire metadata datasets generated during business processes. The metadata datasets represent the descriptive information corresponding to technical data and process definitions during business processes. Features are extracted from the metadata datasets from the business value dimension to obtain a metadata feature set. The business value dimension represents the business management and operational status as depicted by the technical and business metadata. A target multi-classification model is used to assess the importance of the metadata feature set, resulting in an importance grading result for the metadata dataset. The target multi-classification model is obtained by training an initial multi-classification model using a feature sample set with labeled classification results. The feature sample set contains metadata features for multiple sample classes. Data analysis of the business is performed based on the importance grading results, and the classification results are displayed on the graphical interface 400-1. Server 200 stores the metadata datasets generated during business processes in database 500 and retrieves the metadata datasets from database 500 when terminal 400 acquires the metadata datasets, providing them to terminal 400.

[0062] In some embodiments, server 200 may 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, CDN, and big data and artificial intelligence platforms. Terminal 400 may be a smartphone, tablet computer, laptop computer, desktop computer, smart speaker, smartwatch, etc., but is not limited thereto. Terminals and servers can be directly or indirectly connected via wired or wireless communication, which is not limited in this embodiment.

[0063] See Figure 2 , Figure 2 This is a schematic diagram of the structure of the terminal 400 provided in the embodiment of this application. Figure 2 The terminal 400 shown includes at least one processor 410, a memory 450, at least one network interface 420, and a user interface 430. The various components in the terminal 400 are coupled together via a bus system 440. It is understood that the bus system 440 is used to implement communication between these components. In addition to a data bus, the bus system 440 also includes a power bus, a control bus, and a status signal bus. However, for clarity, ... Figure 2 The general labeled all buses as Bus System 440.

[0064] Processor 410 can be an integrated circuit chip with signal processing capabilities, such as a general-purpose processor, a digital signal processor (DSP), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. Among them, the general-purpose processor can be a microprocessor or any conventional processor, etc.

[0065] User interface 430 includes one or more output devices 431 that enable the presentation of media content, including one or more speakers and / or one or more visual displays. User interface 430 also includes one or more input devices 432, including user interface components that facilitate user input, such as a keyboard, mouse, microphone, touch screen display, camera, other input buttons and controls.

[0066] The memory 450 may be removable, non-removable, or a combination thereof. Exemplary hardware devices include solid-state storage, hard disk drives, optical disk drives, etc. The memory 450 may optionally include one or more storage devices physically located away from the processor 410.

[0067] The memory 450 may include volatile memory or non-volatile memory, or both. The non-volatile memory may be read-only memory (ROM), and the volatile memory may be random access memory (RAM). The memory 450 described in this application embodiment is intended to include any suitable type of memory.

[0068] In some embodiments, memory 450 is capable of storing data to support various operations, examples of which include programs, modules, and data structures or subsets or supersets thereof, as illustrated below.

[0069] Operating system 451 includes system programs for handling various basic system services and performing hardware-related tasks, such as the framework layer, core library layer, driver layer, etc., for implementing various basic business functions and handling hardware-based tasks;

[0070] The network communication module 452 is used to reach other computing devices via one or more (wired or wireless) network interfaces 420, exemplary network interfaces 420 including: Bluetooth, WiFi, and Universal Serial Bus (USB), etc.

[0071] Presentation module 453 is configured to enable the presentation of information (e.g., a user interface for operating peripheral devices and displaying content and information) via one or more output devices 431 associated with user interface 430 (e.g., a display screen, a speaker, etc.).

[0072] The input processing module 454 is used to detect and translate one or more user inputs or interactions from one or more input devices 432.

[0073] In some embodiments, the business data evaluation apparatus provided in this application can be implemented in software. Figure 2 A business data evaluation device 455 stored in memory 450 is shown. It can be software in the form of programs and plug-ins, including the following software modules: data acquisition module 4551, feature extraction module 4552, model calculation module 4553, and data analysis module 4554. These modules are logically connected and can therefore be arbitrarily combined or further split according to the functions they implement.

[0074] The functions of each module will be explained below.

[0075] In other embodiments, the apparatus provided in this application can be implemented in hardware. As an example, the apparatus provided in this application can be a processor in the form of a hardware decoding processor, which is programmed to execute the business data evaluation method provided in this application. For example, the processor in the form of a hardware decoding processor can be one or more application-specific integrated circuits (ASICs), DSPs, programmable logic devices (PLDs), complex programmable logic devices (CPLDs), field-programmable gate arrays (FPGAs), or other electronic components.

[0076] The business data evaluation method provided in this application will be described in conjunction with exemplary applications and implementations of the terminals provided in the embodiments of this application.

[0077] See Figure 3 , Figure 3 This is an optional flowchart illustrating the business data evaluation method provided in this application embodiment, which will be combined with... Figure 3 The steps shown are explained.

[0078] S101. Obtain the metadata generated during the business process; the metadata represents the descriptive information corresponding to the technical data and process definitions during the business process.

[0079] In this embodiment of the application, the business data evaluation device acquires the metadata generated during the business process and evaluates the data generated during the business process based on the metadata.

[0080] In this embodiment, the metadata dataset represents the descriptive information corresponding to technical data and process definitions during business processes. The metadata dataset can consist of technical metadata and business metadata. Technical metadata refers to technical statistical indicators generated during data development, which are data used by data warehouse designers and administrators for the development and daily management of the data warehouse. In some embodiments, the scope of technical metadata may include: 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. Examples include data lineage, fan-in count, fan-out count, field names, field lengths, and database table structures. Business metadata refers to various attributes and concepts in the enterprise environment represented by business names, definitions, and descriptions. In some embodiments, the business context behind the data generated during business processes can be considered as business metadata, such as business names, business definitions, and business descriptions.

[0081] In this embodiment, the business data evaluation device can obtain business metadata from manually filled data tables and technical metadata from the configuration information of data warehouse scheduling tasks, thereby obtaining a metadata dataset. The specific selection is made according to the actual situation, and this embodiment does not limit it.

[0082] In this embodiment, the data structure of metadata can be a two-dimensional vector consisting of metadata items and corresponding metadata content. The metadata content refers to specific data, and the metadata items are metrics describing the metadata.

[0083] In some embodiments, the metadata collected by the business data evaluation device can be presented in a table, with each row of data in the table corresponding to a technical metadata or business metadata.

[0084] S102. Extract features from the metadata dataset from the business value dimension to obtain the metadata feature set; the business value dimension is the business management and operation status represented by the technical metadata and business metadata.

[0085] In this embodiment of the application, in order to obtain more valuable information for evaluating business processes from the original metadata set, the business data evaluation device can perform necessary cleaning, preprocessing analysis and feature extraction on the metadata from the business value dimension to obtain a metadata feature set, wherein the metadata feature set represents the attribute features with business value that the metadata can embody.

[0086] In this embodiment of the application, the business value dimension represents the dimension for evaluating the overall value of business data, and can cover multiple dimensions such as popularity, breadth, and profitability.

[0087] S103. Use the target multi-classification model to evaluate and classify the metadata feature set to obtain the importance ranking results corresponding to the metadata dataset; wherein, the target multi-classification model is obtained by training the initial multi-classification model with the feature sample set with labeled classification results; the feature sample set contains metadata features of multiple sample classes.

[0088] In this embodiment, the business data evaluation device can use a target multi-classification model to classify the metadata dataset based on the metadata feature set, classifying multiple metadata items contained in the metadata set into at least one preset importance level, thereby obtaining the importance classification result corresponding to the metadata dataset. Each preset importance level in the importance classification result contains at least one metadata item corresponding to that preset importance level.

[0089] In this embodiment, the target multi-classification model is obtained by training an initial multi-classification model using a feature sample set to assess the importance of various metadata features. It has the ability to classify multiple categories based on the importance of various metadata features. In some embodiments, at least one preset importance level can be a value from 1 to 5, where 1 represents that the business value of the metadata feature is very important; 5 represents that the importance of the metadata feature is the lowest; and 2, 3, and 4 represent important, medium, and ordinary importance levels, respectively.

[0090] S104. Perform data analysis on business operations based on the importance classification results.

[0091] In this embodiment, since the importance classification result represents the different preset importance levels of various metadata contained in the metadata set, when the business data evaluation device obtains the importance classification result corresponding to the metadata dataset, it can extract the metadata that is more important to the business process based on the importance classification result, save it as data assets in the business process, and based on the data assets, know the changing trends of various indicators of the business data, the ranking of the overall value indicators, etc., thereby realizing data analysis of the business process and realizing the effective operation and accumulation of data assets.

[0092] It is understood that in the embodiments of this application, the business data evaluation device uses a target multi-classification model to evaluate the importance of the metadata dataset generated by the business process, thereby improving the evaluation speed and thus improving the efficiency of business data evaluation. Furthermore, since the target multi-classification model is obtained by training the initial multi-classification model using a feature sample set containing multiple metadata features, the target multi-classification model can be used to conduct a comprehensive and systematic evaluation of the metadata dataset based on evaluation indicators of multiple metadata features, thereby improving the accuracy of business data evaluation.

[0093] In some embodiments, see Figure 4 , Figure 4 This is an optional flowchart illustrating the business data evaluation method provided in this application embodiment, based on... Figure 3 The metadata dataset includes: technical metadata and business metadata; technical metadata represents technical data generated and used during business operations; business metadata represents contextual description information generated and defined during business operations. S102 can be implemented by executing S1021-S1023, which will be explained in conjunction with each step.

[0094] S1021. Extract technical metadata features from the technical metadata; the technical metadata features represent the technical statistical indicators corresponding to the business scale and business access volume.

[0095] In this embodiment of the application, the business data evaluation device can extract technical statistical indicators that characterize the business scale and business access volume from the technical metadata as technical metadata features.

[0096] In some embodiments, technical metadata features may include at least one of the following: number of downstream dependent tasks, number of upstream dependent tasks, access frequency, whether it serves online services, number of online services served, number of fields, number of data rows, and storage size.

[0097] S1022. Extract business metadata features from business metadata; business metadata features represent descriptive statistical indicators corresponding to business attributes and business process definitions.

[0098] In this embodiment of the application, the business data evaluation device can extract descriptive statistical indicators that characterize the business attributes and business process definitions from the business metadata, and use them as features of the business metadata.

[0099] In some embodiments, business metadata characteristics may include at least one of the following: the business category to which it belongs, the business star rating, and whether it involves sensitive data.

[0100] S1023. Perform at least one of the following feature processing methods on the technical metadata features and business metadata features: feature normalization, feature transformation, and feature selection, to obtain a metadata feature set.

[0101] In this embodiment of the application, the business data evaluation device can process the technical metadata features and the business metadata features respectively to obtain a metadata feature set.

[0102] In this embodiment, the business data evaluation device can utilize the feature construction method in feature engineering to construct a high-quality metadata feature set from the original technical metadata features and business metadata features, and use a target multi-classification model to classify and predict the metadata feature set after feature processing, so as to improve the stability and accuracy of the classification prediction results.

[0103] In this embodiment, the business data evaluation device can perform feature normalization processing on the technical metadata features and the business metadata features respectively, so as to improve the stability of the feature samples input to the target multi-classification model, thereby ensuring the stability of the prediction results of the target multi-classification model. In some embodiments, the feature normalization method can adopt normal distribution normalization.

[0104] In this embodiment, the business data evaluation device can perform feature transformation on technical metadata features and business metadata features respectively. In some embodiments, the feature transformation method may include: regularization, discretization, one-hot encoding, etc.

[0105] In this embodiment, the business data evaluation device can perform feature selection on technical metadata features and business metadata features respectively, to eliminate irrelevant or redundant features and retain high-value features. In some embodiments, the feature selection method may include: the Extreme Gradient Boosting (XGboost) algorithm.

[0106] It should be noted that, in the embodiments of this application, the method of feature processing by the business data evaluation device can be any one of the feature normalization, feature transformation and feature selection mentioned above, or any combination of several can be used to construct features. The specific selection is based on the actual situation, and the embodiments of this application do not limit it.

[0107] It is understood that, in the embodiments of this application, the business data evaluation device obtains a metadata feature set by performing feature processing on technical meta-features and business meta-features, thereby improving the accuracy of classification and prediction of the metadata feature set using the target multi-classification model.

[0108] In some embodiments, see Figure 5 , Figure 5This is an optional flowchart illustrating the business data evaluation method provided in this application embodiment, based on... Figure 3 or Figure 4 The target multi-class classification model can be an ensemble learning model, wherein the ensemble learning model contains at least one classification tree. S103 can be implemented by executing S1031-S1032, which will be explained in conjunction with each step.

[0109] S1031. Using each of the at least one classification tree, evaluate and classify the importance of different metadata features in the metadata feature set to obtain the sub-classification results corresponding to each classification tree. The sub-classification results belong to at least one preset importance level.

[0110] In this embodiment of the application, the business data evaluation device uses each of the at least one classification tree to classify and predict the metadata feature set, thereby obtaining the correspondence between different metadata in the metadata feature set and at least one preset importance level, which serves as the sub-classification result corresponding to each classification tree.

[0111] S1032. Perform a summary vote based on the results of each subclassification, and determine the importance ranking results based on the voting results.

[0112] In this embodiment, the business data evaluation device summarizes the sub-classification results corresponding to each classification tree, and based on the summary results, counts the number of votes for each metadata feature belonging to each preset importance level. The preset importance level with the highest number of votes is taken as the classification result of the metadata feature, and finally the importance classification result of the entire metadata dataset is obtained.

[0113] In some embodiments, the business data evaluation device may also preset different weight values ​​for each classification tree in at least one classification tree. In this way, when voting, the sub-classification results corresponding to each classification tree can be multiplied by their respective preset weights before voting, and then the importance classification results can be determined based on the voting results.

[0114] In some embodiments, the target multi-class classification model includes at least one classification tree such as Figure 6 As shown, the metadata feature set contains at least one metadata feature from feature 1 to feature 5, with a preset importance level of 1-5. In the subclassification results of the first tree, the importance prediction for feature 2 is 1, indicating that it is very important; in the subclassification results of the second tree, the importance prediction for feature 2 is 3, indicating that it is of medium importance; in the subclassification results of the third tree, the importance prediction for feature 2 is also 1. Therefore, the business data evaluation device can determine the preset importance level 1 as the classification result of feature 2.

[0115] It is understood that, in this embodiment of the application, the importance classification assessment and prediction of the metadata feature set is performed using an ensemble learning model, saving the workload of manually classifying the importance of metadata and improving the efficiency of business data assessment. Furthermore, the accuracy of importance classification prediction can be improved by utilizing the multiple decision classification trees contained in the ensemble learning model, thereby enhancing the accuracy of business data assessment.

[0116] In some embodiments, see Figure 7 , Figure 7 This is an optional flowchart illustrating the business data evaluation method provided in this application embodiment, based on... Figure 5 Before S102, S001-S003 can also be executed, which will be explained in conjunction with each step.

[0117] S001. Obtain the original feature set; the original feature set contains multiple original metadata features with labeled classification results.

[0118] In this embodiment of the application, the business data evaluation device first obtains the original feature set, wherein the original feature set contains multiple original metadata features extracted from business sample data, and each original metadata feature has a corresponding labeled classification result.

[0119] In this embodiment of the application, the classification result is labeled as belonging to at least one preset importance level.

[0120] S002. Perform feature processing on the original feature set to obtain the feature sample set.

[0121] In this embodiment of the application, in order to improve the generalization ability of the classification model, the business data evaluation device can construct features from the original feature set to obtain a feature sample set, and use the feature sample set to train the initial multi-classification model, thereby improving the training effect of training the initial multi-classification model.

[0122] In this embodiment, the process of feature processing of the original feature set by the business data evaluation device is consistent with that described in S102, and will not be repeated here. The business data evaluation device can also use feature normalization methods to accelerate model convergence during training; it can also use feature transformation methods to increase the diversity of feature samples, thereby improving the training effect and classification accuracy of the model.

[0123] S003. Train the initial multi-classification model using the feature sample set to obtain the target multi-classification model.

[0124] In this embodiment of the application, the business data evaluation device uses the feature sample set obtained by feature construction processing to train the initial multi-classification model to obtain the target multi-classification model.

[0125] In this embodiment, the initial multi-classification model can be various types of network models, such as traditional learning models, ensemble learning models, and deep learning models. The business data evaluation device can input the feature sample set into the initial multi-classification model, use the initial multi-classification model to perform importance classification prediction on the feature sample set, obtain the classification prediction result, compare the classification prediction result with the labeled classification result corresponding to the feature sample set to obtain the training error, and iteratively update the initial multi-classification model based on the training error. Through a preset objective optimization function, the training error is gradually reduced in each subsequent training until the training conditions are met, thus obtaining the target multi-classification model.

[0126] In some embodiments, business data evaluation may also divide the feature sample set into a training set and a test set to train the initial multi-classification model. See also Figure 8 , Figure 8 This is an optional flowchart illustrating a business data evaluation method provided in an embodiment of this application, based on... Figure 7 S003 can be achieved by executing S201-S205, which will be explained in detail below:

[0127] S201. In the current training, a preset first number of feature samples are selected from the feature sample set as the current training set.

[0128] In this embodiment of the application, for the current training, the business data evaluation device selects a preset first number of feature samples from the feature sample set as the current training set.

[0129] In this embodiment of the application, for each training of the initial multi-classification model, the business data evaluation device selects the current training set from the feature sample set. That is to say, the current training set used for each training of the initial multi-classification model is different.

[0130] In some embodiments, the business data evaluation device may select the current training set by random selection or by other means, depending on the actual situation. This application embodiment does not limit the selection.

[0131] In some embodiments, the preset first quantity can be 70% of the total number of feature samples in the feature sample set, or it can be preset to other values ​​according to the needs of actual application. The specific selection is based on the actual situation, and the embodiments of this application do not limit it.

[0132] S202. Use the initial multi-classification model to perform classification prediction on the current training set, obtain the predicted classification result, and iteratively update the initial multi-classification model based on the predicted classification result and the labeled classification result until the preset training objective is reached, then end the training and obtain the trained multi-classification model.

[0133] In this embodiment, the business data evaluation device uses an initial multi-classification model to perform classification prediction on the current training set for the current training iteration, obtains the predicted classification result corresponding to the current training iteration, and adjusts and updates the model parameters of the initial multi-classification model based on the error between the predicted classification result and the labeled classification result of the current training iteration. The adjusted multi-classification model is then used for the next round of training, and the above iterative process is repeated. In each model parameter update, the error between the predicted classification result and the labeled classification result is optimized until the preset training objective is reached. If the preset number of training iterations is reached, or the error is less than the preset threshold, the training ends, and the trained multi-classification model is obtained.

[0134] In some embodiments, the initial multi-classification model includes an initial classification tree, which can be implemented by executing S2021-S2023, and will be described in conjunction with each step.

[0135] S2021. When the current training is the first training, the initial classification tree is used to perform classification prediction on the current training set to obtain the classification prediction value corresponding to the initial classification tree.

[0136] In this embodiment of the application, the business data evaluation device can use an initial classification tree as an initial multi-classification model to perform classification prediction on the current training set.

[0137] In some embodiments, the initial classification tree may be a classification and regression tree (CART) or other types of decision classification trees. The specific choice depends on the actual situation and is not limited in the embodiments of this application.

[0138] In this embodiment of the application, when the current training is the first training, the business data evaluation device uses the default tree structure and leaf node parameter values ​​as the initial classification tree, and uses the initial classification tree to perform classification prediction on the current training set to obtain the classification prediction value corresponding to the initial classification tree.

[0139] S2022. When the current training is not the first training, construct the current classification tree corresponding to this training based on the classification prediction value and labeled classification result of the previous classification tree.

[0140] In this embodiment, when the current training is not the first training iteration, the business data evaluation device obtains the training error of the previous training based on the classification, testing, and labeling results corresponding to the previous classification tree. The business data evaluation device then constructs the structure of the classification tree corresponding to the current training and the parameters corresponding to each leaf node based on the training error of the previous training, using this as the current classification tree. This allows the device to fit the training error of the previous training through the current classification tree, enabling the training error to gradually decrease and converge during the training process.

[0141] S2023. Use the current classification tree to perform classification prediction on the current training set, and obtain the classification prediction value corresponding to the current classification tree. Continue the above iterative process until the preset training target is reached. Then, use at least one classification tree obtained during the training process as the trained multi-class model.

[0142] In this embodiment, the business data evaluation device uses the current classification tree constructed during the current training to perform classification prediction on the randomly selected current training set, obtaining the classification prediction value corresponding to the current classification tree as the training result, and determines whether the current training has achieved the preset training objective. If the current training has not achieved the preset training objective, the business data evaluation device enters the next round of training, taking the next round of training as the current training, and continues to construct the current classification tree corresponding to the current training based on the training error of the previous training. This iterative process continues until the preset training objective is achieved, at which point the training ends, and at least one classification tree is obtained. The business data evaluation device uses at least one classification tree as the trained multi-classification model.

[0143] S203. Select a preset second number of test sets from the feature sample set.

[0144] In this embodiment of the application, the business data evaluation device can randomly select a preset second number of test sets from the feature sample set. For example, 30% of the feature samples in the feature sample set can be selected as the test set.

[0145] S204. Use the trained multi-classification model to perform classification prediction on the test set to obtain the test classification results.

[0146] In this embodiment of the application, in order to verify the classification effect of the trained multi-classification model, the business data evaluation device uses the trained multi-classification model to perform classification prediction on the test set and obtain test classification results, so as to verify whether the classification effect meets the requirements through the test classification results.

[0147] S205. Evaluate the classification effect of the test classification results and obtain the evaluation results. When the evaluation results meet the preset evaluation threshold, the trained multi-classification model is used as the target multi-classification model.

[0148] In this embodiment of the application, the business data evaluation device uses a preset model performance evaluation method to evaluate the classification performance of the test classification results based on indicators such as accuracy, recall, and Receiver Operating Characteristic (ROC) curve, and obtains the evaluation result. When the evaluation result is passed, the trained multi-classification model is used as the target multi-classification model.

[0149] In some embodiments, when the business data evaluation device evaluates the classification effect of the test classification results based on the accuracy evaluation metric, the business data evaluation device can compare the test classification results with the labeled classification results to obtain the classification accuracy. The classification test accuracy is the ratio of the number of feature samples with correct classification results to the total number of feature samples in the test set. When the classification accuracy is greater than a preset accuracy threshold, it indicates that the classification prediction results of the trained multi-classification model meet the requirements of importance evaluation classification, and the business data evaluation device uses the trained multi-classification model as the target multi-classification model.

[0150] It is understood that in the embodiments of this application, the feature sample set used for training has undergone feature construction processing, which can improve the model accuracy of the target multi-classification model obtained by training. Furthermore, since the target multi-classification model is obtained by training the initial multi-classification model using a feature sample set containing multiple metadata features, the target multi-classification model can be used to comprehensively and systematically evaluate the metadata dataset based on the evaluation indicators of multiple metadata features, thereby further improving the accuracy of business data evaluation.

[0151] The following will describe an exemplary application of the embodiments of this application in a real-world application scenario.

[0152] In this embodiment of the application, for financial business scenarios, it can be as follows: Figure 9As shown, the business data evaluation method proposed in this application is implemented through data preparation, data processing, feature construction, model tuning, and model deployment stages. In the data preparation stage, the business data evaluation device can acquire business data tables filled out by data warehouse designers and managers as business metadata, and automatically read technical data tables from data warehouse configuration information and operation records as technical metadata, thus obtaining 10,000 tables as a metadata dataset. Next, in the data preparation stage, 5% (no less than 500 tables) can be randomly sampled from the metadata set and manually categorized for importance. The importance level of the data assets is manually determined based on business experience. For example, the preset classification of importance levels can be defined as: 1-Very Important; 2-Important; 3-Medium; 4-Ordinary; 5-Unimportant. To unify the manual judgment standards of different annotators, 100 tables can be randomly selected from the 500 tables, and blind labeling can be performed by two or more annotators. When the consistency rate of labeling by all annotators is greater than or equal to 90%, the labeling result is considered satisfactory, and the process proceeds to feature construction. During feature construction, the business data evaluation device can use the technical metadata features and business metadata features from 500 tables as the original feature samples in the original feature sample set, and the manually labeled classification results as the labeled classification results for each original feature sample in the original feature sample set. Furthermore, in the feature construction stage, the business data evaluation device can construct a data asset feature set by combining the technical metadata and business metadata features in the original feature sample set using feature construction methods from feature engineering. During model training in the model tuning stage, the business data evaluation device divides the data asset dataset into training and testing sets. Using the training set, it trains at least one candidate model, such as traditional learning models like tree models (decision trees) and logistic regression (LR) models; ensemble learning models like XGBoost and random forest models; and deep learning models like Long Short-Term Memory (LSTM) models and neural network models, to select the target multi-class classification model with the best classification performance. The business data evaluation device trains each of the above candidate models. After the training of each candidate model is completed, the business data evaluation device enters the model evaluation process, uses the test set to predict and verify the training results of each candidate model, and uses multi-classification model evaluation indicators to evaluate the model effect of at least one candidate model. Finally, based on multiple experiments in the embodiments of this application, in terms of the classification task of the importance of data assets and data characteristics, the XGBoost model in the ensemble learning model shows a relatively strong classification effect. In the choice between computational cost and final effect, the embodiments of this application finally adopt the XGBoost model as the target multi-classification model.During the model deployment phase, the business data assessment device can input all unlabeled data tables and metadata features from subsequent business processes into the XGBoost model to predict the current importance level of all data assets and complete the classification task of data asset rating.

[0153] It is understood that using the business data evaluation method in the embodiments of this application for multi-class task prediction modeling can achieve the effect of data asset rating and stratification, obtain a refined classification of data asset rating, and thus help product and business teams to implement differentiated data asset monitoring and processing for data assets of different categories, improve the level of refined operation, improve the level of data asset management, solve the problem of data quality and value assessment in the data development process, and help the business quickly realize data assetization.

[0154] The following description continues to illustrate the exemplary structure of the business data evaluation device 455 provided in the embodiments of this application as a software module. In some embodiments, such as Figure 2 As shown, the software modules stored in the service data evaluation device 455 in the memory 450 may include:

[0155] Data acquisition module 4551 is used to acquire metadata generated during the business process; the metadata represents the descriptive information corresponding to the technical data and process definition during the business process;

[0156] Feature extraction module 4552 is used to extract features from the metadata dataset from the business value dimension to obtain a metadata feature set; the business value dimension is the business management and operation status represented by the technical metadata and the business metadata.

[0157] The model calculation module 4553 is used to perform importance assessment and classification on the metadata feature set using a target multi-classification model, and obtain the importance ranking results corresponding to the metadata dataset; wherein...

[0158] The target multi-class classification model is obtained by training an initial multi-class classification model using a feature sample set with labeled classification results, and the feature sample set contains metadata features of multiple sample classes.

[0159] The data analysis module 4554 is used to perform data analysis on the business based on the importance classification results. In some embodiments, the feature extraction module 4552 is further used to extract technical metadata features from the technical metadata; the technical metadata features represent technical statistical indicators corresponding to business scale and business access volume; extract business metadata features from the business metadata; the business metadata features represent descriptive statistical indicators corresponding to business attributes and business process definitions; and perform at least one feature processing method among feature normalization, feature transformation, and feature selection on the technical metadata features and the business metadata features respectively to obtain the metadata feature set.

[0160] In some embodiments, the technical metadata includes at least one of: data source information, data transformation description information, object definition, data structure definition, rules used for data cleaning and data updating, and mapping from source data to destination data; the business metadata includes at least one of: business name, business definition, business description information, and business attributes; the technical metadata features include at least one of: number of downstream dependent tasks, number of upstream dependent tasks, access frequency, whether it serves online business, number of online businesses served, number of fields, number of data rows, and storage size; the business metadata features include at least one of: business category, business star rating, and whether it involves sensitive data.

[0161] In some embodiments, the target multi-classification model is an ensemble learning model, which includes at least one classification tree. The model calculation module 4553 is further configured to use each of the at least one classification tree to perform importance assessment and classification on different metadata features in the metadata feature set, thereby obtaining a sub-classification result corresponding to each classification tree. The sub-classification result belongs to at least one preset importance level. Based on each sub-classification result, a summary vote is performed, and the importance classification result is determined based on the voting result.

[0162] In some embodiments, the business data evaluation device further includes a model training module, which is used to obtain an original feature set before using a target multi-classification model to perform importance evaluation and classification on the metadata feature set and obtain the importance ranking result corresponding to the metadata dataset; the original feature set contains multiple original metadata features with labeled classification results; the original feature set is subjected to feature processing to obtain the feature sample set; and the initial multi-classification model is trained using the feature sample set to obtain the target multi-classification model.

[0163] In some embodiments, the model training module is further configured to process the original feature set using at least one of the following methods: feature normalization, feature transformation, and feature selection, thereby obtaining the feature sample set.

[0164] In some embodiments, the model training module is further configured to: select a preset first number of feature samples from the feature sample set as the current training set in the current training iteration; use the initial multi-classification model to perform classification prediction on the current training set to obtain a predicted classification result; and iteratively update the initial multi-classification model based on the predicted classification result and the labeled classification result until a preset training objective is reached, at which point training ends to obtain a trained multi-classification model; select a preset second number of feature samples from the feature sample set as a test set; use the trained multi-classification model to perform classification prediction on the test set to obtain a test classification result; use a preset evaluation method to evaluate the classification effect of the test classification result to obtain an evaluation result; and when the evaluation result is passed, use the trained multi-classification model as the target multi-classification model.

[0165] In some embodiments, the initial multi-class model includes: an initial classification tree. The model training module is further configured to: when the current training is the first training iteration, use the initial classification tree to perform classification prediction on the current training set to obtain the classification prediction value corresponding to the initial classification tree; when the current training is not the first training iteration, construct the current classification tree corresponding to the current training iteration based on the classification prediction value corresponding to the previous classification tree and the labeled classification result; use the current classification tree to perform classification prediction on the current training set to obtain the classification prediction value corresponding to the current classification tree; continue the above iterative process until a preset training objective is reached; and use at least one classification tree obtained during the training process as the trained multi-class model.

[0166] It should be noted that the description of the above device embodiments is similar to the description of the above method embodiments, and has similar beneficial effects. For technical details not disclosed in the device embodiments of this application, please refer to the description of the method embodiments of this application for understanding.

[0167] This application provides a computer program product or computer program that includes 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 business data evaluation method described in this application.

[0168] This application provides a computer-readable storage medium storing executable instructions. When these executable instructions are executed by a processor, they cause the processor to perform the method provided in this application, for example... Figure 3-5 , Figure 7 and Figure 8 The method shown in the figure.

[0169] In some embodiments, the computer-readable storage medium may be a memory such as FRAM, ROM, PROM, EPROM, EEPROM, flash memory, magnetic surface memory, optical disk, or CD-ROM; or it may be a variety of devices including one or any combination of the above-mentioned memories.

[0170] In some embodiments, executable instructions may take the form of a program, software, software module, script, or code, written in any form of programming language (including compiled or interpreted languages, or declarative or procedural languages), and may be deployed in any form, including as a standalone program or as a module, component, subroutine, or other unit suitable for use in a computing environment.

[0171] As an example, executable instructions may, but do not necessarily, correspond to files in a file system. They may be stored as part of a file that holds other programs or data, for example, in one or more scripts in a Hyper Text Markup Language (HTML) document, in a single file dedicated to the program in question, or in multiple collaborating files (e.g., a file that stores one or more modules, subroutines, or code sections).

[0172] As an example, executable instructions can be deployed to execute on a single computing device, or on multiple computing devices located in one location, or on multiple computing devices distributed across multiple locations and interconnected via a communication network.

[0173] In summary, through the embodiments of this application, the business data evaluation device utilizes a target multi-classification model to evaluate the importance of metadata generated by the business process, thereby improving the evaluation speed and efficiency of business data evaluation. Furthermore, since the target multi-classification model is obtained by training an initial multi-classification model using a feature sample set containing multiple metadata features, the target multi-classification model can be used to conduct a comprehensive and systematic evaluation of the metadata dataset based on evaluation indicators of multiple metadata features, thereby improving the accuracy of business data evaluation.

[0174] The above description is merely an embodiment of this application and is not intended to limit the scope of protection of this application. Any modifications, equivalent substitutions, and improvements made within the spirit and scope of this application are included within the scope of protection of this application.

Claims

1. A method of evaluating service data, characterized by, The method comprises the following steps: obtaining a metadata set generated in a business process; the metadata set represents description information corresponding to technical data and process definition in the business process; the metadata set comprises technical metadata and business metadata; the technical metadata represents technical data generated and used in the business running process; and the business metadata represents context description information generated and defined in the business running process; extracting technical metadata features from the technical metadata from a business value dimension; the technical metadata features represent technical statistical indicators corresponding to business scale and business access volume; extracting business metadata features from the business metadata from a business value dimension; the business metadata features represent descriptive statistical indicators corresponding to business attributes and business process definition; respectively performing at least one feature processing of feature normalization, feature transformation and feature selection on the technical metadata features and the business metadata features to obtain a metadata feature set; the business value dimension is a business management and operation state represented by the technical metadata and the business metadata; using a target multi-classification model to perform importance evaluation classification on the metadata feature set to obtain an importance grading result corresponding to the metadata set; wherein, the target multi-classification model is obtained by performing importance evaluation classification training on an initial multi-classification model using a feature sample set with labeled classification results; the feature sample set contains metadata features of multiple sample classes; based on the importance grading result, data analysis of the business is realized.

2. The method of claim 1, wherein, The technical metadata comprises at least one of data source information, data conversion description information, object definition, data structure definition, rules for data cleaning and data updating, and mapping from source data to destination data; The business metadata comprises at least one of business name, business definition, business description information, and business attributes; The technical metadata features comprise at least one of the number of downstream dependent tasks, the number of upstream dependent tasks, access frequency, whether to serve online business, the number of online business services, the number of fields, the number of data rows, and storage size; The business metadata features comprise at least one of the classification of the business, the star rating of the business, and whether sensitive data is involved.

3. The method according to any of claims 1-2, characterized in that, The target multi-classification model is an ensemble learning model, and the ensemble learning model comprises at least one classification tree; the target multi-classification model is used to perform importance evaluation classification on the metadata feature set to obtain an importance grading result corresponding to the metadata set, comprising: using each classification tree in the at least one classification tree to perform importance evaluation classification on different metadata features in the metadata feature set respectively to obtain a sub-classification result corresponding to each classification tree, wherein the sub-classification result belongs to at least one preset importance level; based on each sub-classification result, voting is performed, and the importance grading result is determined based on the voting result.

4. The method of claim 3, wherein, Before the step of using the target multi-classification model to perform importance evaluation classification on the metadata feature set to obtain an importance grading result corresponding to the metadata set, the method further comprises: obtaining an original feature set; the original feature set comprising a plurality of original metadata features with labeled classification results; performing feature processing on the original feature set to obtain a feature sample set; training the initial multi-classification model through the feature sample set to obtain a target multi-classification model.

5. The method of claim 4, wherein, The training of the initial multi-classification model through the feature sample set to obtain a target multi-classification model comprises: In the current training, a preset first number of feature samples are selected from the feature sample set as a current training set; using the initial multi-classification model to perform classification prediction on the current training set to obtain a predicted classification result, and iteratively updating the initial multi-classification model based on the predicted classification result and the labeled classification result until a preset training target is reached, ending the training, and obtaining a trained multi-classification model; selecting a preset second number of feature samples from the feature sample set as a test set; using the trained multi-classification model to perform classification prediction on the test set to obtain a test classification result; performing classification effect evaluation on the test classification result to obtain an evaluation result, and when the evaluation result meets a preset evaluation threshold, regarding the trained multi-classification model as the target multi-classification model.

6. The method of claim 5, wherein, The initial multi-classification model comprises an initial classification tree, and the using of the initial multi-classification model to perform classification prediction on the current training set to obtain a predicted classification result, and the iteratively updating of the initial multi-classification model based on the predicted classification result and the labeled classification result until a preset training target is reached, ending the training, and obtaining a trained multi-classification model comprises: when the current training is the first training, using the initial classification tree to perform classification prediction on the current training set to obtain a classification prediction value corresponding to the initial classification tree; when the current training is not the first training, constructing a current classification tree corresponding to the current training based on a classification prediction value corresponding to a previous classification tree and the labeled classification result; using the current classification tree to perform classification prediction on the current training set to obtain a classification prediction value corresponding to the current classification tree, and continuing the iteration process until a preset training target is reached, and regarding at least one classification tree obtained in the training process as the trained multi-classification model.

7. A service data evaluation apparatus characterized by comprising: comprises: a data acquisition module configured to obtain a metadata set generated in a business process; the metadata set represents description information corresponding to technical data and process definitions in the business process; the metadata set comprises technical metadata and business metadata; the technical metadata represents technical data generated and used in a business operation process; and the business metadata represents context description information generated and defined in the business operation process; The feature extraction module is configured to extract technical metadata features from the technical metadata from a business value dimension; the technical metadata features represent technical statistical indicators corresponding to business scale and business access volume; extract business metadata features from the business metadata from the business value dimension; the business metadata features represent descriptive statistical indicators corresponding to business attributes and business process definitions; at least one of feature normalization, feature transformation, and feature selection is performed on the technical metadata features and the business metadata features respectively to obtain a metadata feature set; the business value dimension is a business management and operation state represented by the technical metadata and the business metadata; The model calculation module is configured to perform importance evaluation classification on the metadata feature set using a target multi-classification model to obtain an importance classification result corresponding to the metadata set; wherein, The target multi-classification model is obtained by performing importance evaluation classification training on an initial multi-classification model using a feature sample set with labeled classification results; the feature sample set includes metadata features of multiple sample classes; The data analysis module is configured to perform data analysis on the business based on the importance classification result.

8. The apparatus of claim 7, wherein, The technical metadata includes at least one of data source information, data conversion description information, object definition, data structure definition, rules for data cleaning and data updating, and mapping from source data to destination data. The business metadata includes at least one of business name, business definition, business description information, and business attributes. The technical metadata features include at least one of the number of downstream dependent tasks, the number of upstream dependent tasks, access frequency, whether to serve online business, the number of online business services, the number of fields, the number of data rows, and storage size. The business metadata features include at least one of the classification of the business, the star level of the business, and whether sensitive data is involved.

9. The device of any of claims 7-8, wherein, The target multi-classification model is an ensemble learning model, and the ensemble learning model includes at least one classification tree, The model calculation module is further configured to perform importance evaluation classification on different metadata features in the metadata feature set using each classification tree in the at least one classification tree to obtain a sub-classification result corresponding to each classification tree, and the sub-classification result belongs to at least one preset importance level. Each sub-classification result is summarized and voted, and the importance classification result is determined based on the voting result.

10. The apparatus of claim 9, wherein, The device further includes: The model training module is configured to obtain an original feature set before performing importance evaluation classification on the metadata feature set using a target multi-classification model to obtain an importance classification result corresponding to the metadata set; the original feature set includes multiple original metadata features with labeled classification results; the original feature set is processed to obtain the feature sample set; and the initial multi-classification model is trained using the feature sample set to obtain the target multi-classification model.

11. The device of claim 10, wherein, The model training module is further configured to, in the current training, select a preset first number of feature samples from the feature sample set as a current training set; perform classification prediction on the current training set using the initial multi-classification model to obtain a predicted classification result; and perform iterative updating on the initial multi-classification model based on the predicted classification result and the labeled classification result until a preset training target is reached, and the training is ended to obtain a trained multi-classification model; and select a preset second number of feature samples from the feature sample set as a test set; perform classification prediction on the test set using the trained multi-classification model to obtain a test classification result; perform classification effect evaluation on the test classification result using a preset evaluation method to obtain an evaluation result, and when the evaluation result is passed, the trained multi-classification model is used as the target multi-classification model.

12. The apparatus of claim 11, wherein, The initial multi-classification model comprises an initial classification tree, The model training module is further configured to, when the current training is the first training, perform classification prediction on the current training set using the initial classification tree to obtain a classification prediction value corresponding to the initial classification tree; when the current training is the non-first training, construct a current classification tree corresponding to the current training based on a classification prediction value corresponding to a previous classification tree and the labeled classification result; perform classification prediction on the current training set using the current classification tree to obtain a classification prediction value corresponding to the current classification tree, and continue the above iterative process until a preset training target is reached, and at least one classification tree obtained in the training process is used as the trained multi-classification model.

13. An electronic device, comprising: comprising: a memory configured to store executable instructions; a processor configured to execute the executable instructions stored in the memory to implement the method in any one of claims 1 to 6.

14. A computer-readable storage medium, characterized in that, executable instructions stored in the memory, and configured to be executed by a processor to implement the method in any one of claims 1 to 6.

15. A computer program product comprising computer instructions, characterized in that, The computer instructions are executed by the processor to implement the method in any one of claims 1 to 6.

Images