Data annotation and analysis method and device, equipment and medium

Abstract

The invention relates to the technical field of big data and the field of financial science and technology, and discloses a data annotation and analysis method, which comprises the following steps: acquiring a data set containing a to-be-annotated field and a business field, creating an annotation task based on the data set, determining an annotation template, annotating the to-be-annotated field based on the annotation template, and generating an annotation result; and checking the annotation result, generating metadata corresponding to the annotation result, and performing data analysis on the annotation result based on the metadata to generate an analysis result. According to the method and the system, the data annotation and analysis functions are integrated in the same system, so that the analysis process after data annotation is simplified, the dependence on an additional system is reduced, the utilization efficiency of a data annotation result is improved, the integrated processing of data analysis is realized, and support is provided for real-time and accurate analysis.

Description

Technical Field

[0001] The present invention relates to the fields of big data technology and financial technology, and in particular to a data labeling and analysis method, device, equipment and storage medium. Background Art

[0002] As the financial industry gradually transforms from informatization to intelligence, financial institutions have put forward higher requirements for the coverage of business intelligence. In the process of realizing intelligence, data labeling and data analysis are indispensable links. Especially in financial scenarios, data labeling affects the recognition accuracy of business models and the reliability of business decisions. However, most of the data labeling systems on the market currently only focus on the labeling process, and lack support for subsequent quality assessment and analysis of the labeling results, resulting in the inability to effectively use the labeled data in subsequent applications.

[0003] Specifically, current annotation systems usually impose strict restrictions on the format of annotation data, so that the annotation results can only be output in a fixed format; after the annotation is completed, if the results need to be analyzed for quality indicators such as accuracy and recall, the annotation data often needs to be imported into an external analysis platform, which will complete the data merging, calculation and analysis operations. Although this method has achieved data analysis to a certain extent, it has also introduced new systems and operating steps, increasing the complexity of data flow.

[0004] In financial institutions, this multi-system separation model places high demands on the overall IT architecture of the enterprise. For small and medium-sized financial enterprises, they lack the resources and technical foundation to integrate the annotation system and analysis platform, so they can only complete the quality assessment of the annotated data through manual statistics, or simply ignore the analysis link. This situation not only affects the value of the annotated data, but also reduces the accuracy of data-driven decision-making, and cannot meet the financial industry's needs for real-time evaluation and efficient application of data annotation results.

[0005] Therefore, existing technologies cannot effectively support financial institutions to directly analyze the annotation results. There is a lack of a unified method that can integrate data annotation and analysis functions, which makes it difficult to improve the work efficiency and result utilization rate of the data annotation process. In the process of promoting the intelligence of the financial industry, there is an urgent need for an integrated annotation and analysis solution to complete data annotation and analysis on the same platform, simplify system processes, reduce data transmission steps, and improve the utilization efficiency of annotation data. Summary of the invention

[0006] The main purpose of the present invention is to provide a data labeling and analysis method, device, equipment and storage medium, aiming to solve the technical problems that the prior art lacks an integrated solution and cannot complete the efficient utilization of data after labeling in the same system, resulting in a complex analysis process and dependence on additional systems.

[0007] To achieve the above objectives, the present invention provides a data annotation and analysis method, comprising:

[0008] Get the data set containing the fields to be annotated and the business fields;

[0009] Based on the data set, create a labeling task, and determine a labeling template according to the labeling task;

[0010] Annotate the field to be annotated according to the annotation task and the annotation template to generate an annotation result;

[0011] Checking the annotation result according to the annotation template;

[0012] When the annotation result passes the inspection, metadata corresponding to the annotation result is generated, the metadata including business field information and annotation task attribute information;

[0013] The annotation result is analyzed based on the metadata to generate an analysis result.

[0014] Furthermore, to achieve the above-mentioned purpose, the present invention provides a data labeling and analysis device, comprising:

[0015] The data management module is used to obtain the data set containing the fields to be annotated and the business fields;

[0016] A task management module, used to create a labeling task based on the data set, and determine a labeling template according to the labeling task;

[0017] A labeling processing module, used to label the to-be-labeled field according to the labeling task and the labeling template, and generate a labeling result;

[0018] A verification module, used for verifying the annotation result according to the annotation template;

[0019] A metadata generation module, used to generate metadata corresponding to the annotation result when the annotation result passes the inspection, the metadata including business field information and annotation task attribute information;

[0020] A data analysis module is used to perform data analysis on the annotation results based on the metadata to generate analysis results.

[0021] Furthermore, to achieve the above-mentioned purpose, the present invention also provides a computer device, which includes a memory, a processor, and a data labeling and analysis program stored in the memory and executable on the processor, and the data labeling and analysis program, when executed by the processor, implements the steps of the data labeling and analysis method described above.

[0022] Furthermore, to achieve the above objectives, the present invention also provides a computer-readable storage medium, on which a data labeling and analysis program is stored, and when the data labeling and analysis program is executed by a processor, the steps of the data labeling and analysis method described above are implemented.

[0023] Beneficial effects: The present invention relates to the fields of big data technology and financial technology, and discloses a data annotation and analysis method, which obtains a data set containing fields to be annotated and business fields, creates an annotation task based on the data set and determines an annotation template, annotates the fields to be annotated based on the annotation template, and generates annotation results; verifies the annotation results, generates metadata corresponding to the annotation results, and performs data analysis on the annotation results based on the metadata to generate analysis results. The present invention simplifies the analysis process after data annotation by integrating data annotation and analysis functions in the same system, reduces dependence on additional systems, improves the utilization efficiency of data annotation results, realizes integrated processing of data analysis, and provides support for real-time and accurate analysis. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] The present invention will be further described below with reference to the accompanying drawings and embodiments, in which:

[0025] Figure 1 A schematic diagram of an application environment of a data annotation and analysis method according to an embodiment of the present invention;

[0026] Figure 2 It is a flow chart of an embodiment of a data labeling and analysis method of the present invention;

[0027] Figure 3 A schematic diagram of functional modules of a preferred embodiment of the data labeling and analysis device of the present invention;

[0028] Figure 4 A schematic diagram of the structure of a computer device in one embodiment of the present invention;

[0029] Figure 5 FIG. 4 is another schematic diagram of the structure of a computer device in one embodiment of the present invention. DETAILED DESCRIPTION

[0030] It should be understood that the specific embodiments described herein are only used to explain the present invention, and are not used to limit the present invention.

[0031] The data annotation and analysis method provided by the embodiment of the present invention can be applied in Figure 1In an application environment, the user terminal communicates with the server terminal through a network. The server terminal can obtain a data set including fields to be annotated and business fields through the user terminal, create an annotation task based on the data set and determine an annotation template, annotate the fields to be annotated based on the annotation template, and generate an annotation result; verify the annotation result, generate metadata corresponding to the annotation result, perform data analysis on the annotation result based on the metadata, and generate an analysis result. The present invention simplifies the analysis process after data annotation by integrating data annotation and analysis functions in the same system, reduces dependence on additional systems, improves the utilization efficiency of data annotation results, realizes integrated processing of data analysis, and provides support for real-time and accurate analysis. Among them, the user terminal can be but is not limited to various personal computers, laptops, smart phones, tablet computers and portable wearable devices. The server terminal can be implemented with an independent server or a server cluster consisting of multiple servers. The present invention is described in detail below through specific embodiments.

[0032] See also Figure 2 , Figure 2 This is a flow chart of an embodiment of a data annotation and analysis method provided by the present invention. It should be noted that although a logical sequence is shown in the flow chart, in some cases, the steps shown or described may be performed in a different order than that shown here.

[0033] like Figure 2 As shown, the data annotation and analysis method proposed in the present invention includes the following steps:

[0034] S10, obtaining a data set including fields to be annotated and business fields;

[0035] In this embodiment, the data set is a collection of fields to be annotated and business fields. The fields to be annotated are the content that the system will annotate (such as classification labels, risk levels, etc.), while the business fields are auxiliary information that describes the data context (such as customer type, timestamp, transaction amount, etc.) for subsequent analysis. The structural specifications of the data set are predefined in the system design, including the standards for the fields to be annotated and the business fields. For example, the system classifies and describes the field content, and identifies and classifies the fields when the data is uploaded to ensure that the fields in the data set meet the standards for the fields to be annotated or business fields.

[0036] Data sets can come from different data sources, such as local files, external databases, or third-party data services. They can be in structured formats, such as CSV, JSON, or database export formats, to facilitate the system to identify and read the fields to be annotated and business fields. The system allows data sets to be obtained from multiple sources by setting up a data set import module. Different data formats (such as CSV, JSON) are parsed into a unified internal structure after import, ensuring that data sets meet the system's recognition and parsing requirements regardless of their source.

[0037] After acquiring the data set, the system classifies the fields according to the field name or field predefined rules, marks the fields that meet the annotation requirements as fields to be annotated, and marks the descriptive or auxiliary information as business fields. The system automatically identifies the field type using field name matching or predefined templates. Field classification can be achieved through a mapping table. For example, the system maps specific keywords (such as "label" and "category") as fields to be annotated, and identifies fields such as "customer type" and "time" as business fields to ensure that the content of the data set meets the annotation and analysis requirements.

[0038] Example: Assume that a bank needs to annotate and analyze customer loan records. The bank hopes to identify potential high-risk loans by annotating risk factors in loan records, and analyze loan situations of different customer types by business fields.

[0039] The customer loan records of a bank are organized into a data set, which contains multiple fields, such as "loan type", "repayment method", "risk assessment", etc. Here, "risk assessment" is the content that needs to be annotated, which is used to determine the risk level of the loan, so it is defined as a field to be annotated. Fields such as "customer type", "loan amount", and "loan date" describe the background information of the loan and belong to business fields. They will be used to analyze the loan performance and risk distribution of different customers.

[0040] The bank can import this data set directly from its internal database into the system. For example, the bank's IT department connects to its loan management database through this system, selects a table containing customer loan records, and imports the table as a data set into the annotation system. The system reads the table content and identifies the field "risk assessment" as a field to be annotated, while "customer type", "loan amount", etc. are used as business fields for subsequent analysis.

[0041] When importing data, the system automatically parses the field content according to predefined rules, marks "risk assessment" as a field to be marked, and marks and stores other descriptive fields as business fields. This automatic parsing process ensures that the fields can be correctly classified after data import, avoiding the tedious steps of manual screening and improving data processing efficiency.

[0042] By predefining and classifying the fields to be annotated and business fields in the data set, the system can obtain data sets from different sources and in multiple formats. The unified field classification and identification standards simplify the data preparation process, ensure that the format and content of the data set meet the requirements of annotation and analysis, and improve the efficiency of data set utilization and the accuracy of data analysis.

[0043] S20, creating a labeling task based on the data set, and determining a labeling template according to the labeling task;

[0044] In this embodiment, the system creates basic information of the labeling task based on the imported data set. The basic information of the labeling task includes the task name, task description, and task priority, which are used to describe the purpose, content, and processing priority of the task. The system can automatically generate the task name or allow the user to enter a custom task description. The task priority can be preset by the system or manually set by the user. After generating the basic information of the task, the system associates this information with the data set so that it can be viewed and managed in the task management interface.

[0045] After the annotation task is created, the system selects a suitable annotation template based on the requirements of the task. The annotation template defines the field format and required items in the annotation process to ensure that the annotation task meets specific needs and format requirements. The system can filter templates that match the task from the annotation template library based on the field type of the task or the task description information. For example, the system automatically selects a template containing the corresponding field from the template library based on the characteristics of the field to be annotated in the data set (such as field name, data type, etc.); if there are multiple templates that meet the conditions, the system allows the user to manually confirm or automatically select the best matching template.

[0046] After determining the annotation template, the system configures the template's mandatory fields and optional fields for the annotation task. Required fields are items that must be completed during the annotation process, while optional fields are items that provide auxiliary information for annotation, which can improve annotation efficiency. The system sets the display and filling rules for required and optional fields according to the template definition. For example, for a risk assessment task in financial data, the required fields of the template may be "risk level", "risk category", etc., while the optional fields may include auxiliary information such as "remarks" or "labels". After the system is configured, it ensures that the required fields are mandatory in the annotation interface, while the optional fields can be filled in freely according to needs.

[0047] Example description: In the financial field, suppose an insurance company needs to perform risk annotation on customer policies to identify high-risk policies. After the insurance company uploads the policy data set to the system, the system automatically creates a labeling task and associates the data set. According to the requirements of the labeling task, the system selects the "Insurance Risk Assessment" template from the template library, which contains the fields commonly used in insurance labeling tasks. The required fields of the template include "Risk Score" and "Risk Category", while the optional fields include auxiliary items such as "Remarks" or "Annotator's Suggestions".

[0048] The system automatically associates the annotation template with the task and displays the required and optional fields for the user in the task interface. After the annotator enters the task interface, the system ensures that the "risk score" and "risk category" fields are required, while the "remarks" field is optional. Through this integrated annotation process, insurance companies can ensure the accuracy and consistency of data when annotating insurance policies, while improving annotation efficiency and providing a reliable data foundation for subsequent risk assessment.

[0049] By associating data sets and automatically selecting or configuring annotation templates when creating annotation tasks, the system can achieve standardization and modularization of the annotation process when creating tasks, simplifying user operations and improving the accuracy and consistency of annotation tasks. Through the allocation of required fields and optional fields, the system clarifies the key data items and auxiliary data items of the task in the annotation task, so that the annotator can ensure the integrity of the data when performing the task and can flexibly fill in the auxiliary information, thereby improving the annotation efficiency.

[0050] S30, annotating the field to be annotated according to the annotation task and the annotation template to generate an annotation result;

[0051] In this embodiment, the system displays the fields to be annotated on the annotation interface according to the requirements of the annotation task, and displays the required fields and optional fields in the annotation template to ensure that the annotator knows which fields need to be filled in. The required fields are the core content of the annotation task, and the optional fields are auxiliary information. The system displays the data of the fields to be annotated on the annotation interface, and loads the field content configured by the annotation template onto the interface. According to the template configuration, the system marks the required fields as "required" and sets optional marks for the optional fields so that the annotator can quickly identify them in the interface. The annotation interface can be in the form of a form, supporting multiple input methods such as text boxes and options to accommodate different types of annotation data.

[0052] The system displays the data content of the fields to be annotated one by one. The annotator enters the annotation content one by one according to the annotation template in the interface or selects from the preset options to generate complete annotated data. The system displays the data content one by one in the order of the fields to be annotated in the interface. The annotator can fill in or select the annotation content in the corresponding input box. For text annotation, the system provides a text input box; for classification annotation, the system provides a preset drop-down option or selection box. After the annotator completes the input one by one, the system records the annotation content of each piece of data.

[0053] During the annotation process, the system verifies the annotation content in real time to ensure that the filled-in annotation results meet the format and requirements of the annotation template. This includes checking the text length, data type (such as numbers, text), and the legality of options to prevent invalid data from entering the annotation results. The system is configured with a real-time verification function in the annotation interface. For example, the system sets minimum and maximum character limits for required fields, sets a numerical range for numeric fields, or sets a legal option set for selection fields. When the annotator enters data, the system verifies the content immediately. If there is input that does not conform to the format, a prompt message will be displayed, requiring the annotator to correct it before submitting.

[0054] After the annotation is completed, the system saves the annotation results to the database or a specified storage location. The annotation results include the annotation content of the field to be annotated, the auxiliary field data in the template, and other complete annotation data for subsequent analysis and verification. After the annotation results are completed, the system automatically generates a record, including the field to be annotated, the annotation content, and the field content that matches the template. The results are stored in the database and associated with the identification information of the annotation task. The system can automatically generate a unique identifier based on the database for subsequent calls and queries. After storage is completed, the system updates the status of the annotation task to "completed" or "pending verification".

[0055] The annotation task interface displays the fields to be annotated, configures the required and optional fields, and combines real-time format verification and intelligent annotation methods to improve the accuracy and efficiency of annotation. The generated annotation results not only contain the complete content of the fields to be annotated, but also conform to the specifications of the annotation template, which is convenient for subsequent automatic inspection and analysis. In addition, the real-time verification function of the system can effectively prevent the generation of invalid annotation data and ensure the quality of annotation.

[0056] S40, checking the annotation result according to the annotation template;

[0057] In this embodiment, the system first compares the field content in the annotation result with the field requirements in the annotation template to ensure that the annotation result contains the required fields and optional fields required in the template. This step is to ensure that the annotation result is consistent with the template structure. The system automatically checks the field name and quantity in the annotation result according to the field configuration of the annotation template to ensure that all required fields exist. The system compares the annotation result and the template field list through a field matching algorithm, marks missing or redundant fields as errors, and prompts that the inspection fails.

[0058] The system checks the required fields in the annotation results one by one to ensure that each required field is filled in and not empty to avoid missing fields. The required fields are the core content of the annotation data and must be filled in completely to ensure data quality. The system performs a non-empty check on each required field in the annotation results. If the field content is empty or contains only blank characters, the system will record the data as unqualified and mark the field as missing in the inspection results. At the same time, the system can generate the inspection status of each required field through an automatic verification report.

[0059] The system checks the format of each field in the annotation result one by one according to the requirements for each field format in the annotation template. For example, whether the text field meets the character limit, whether the numeric field is within the specified range, whether the classification field is one of the predefined options, etc. The system configures format validation rules for each field. For text fields, the system checks the number of characters and content type; for numeric fields, the system checks their numeric range; for classification fields, the system compares whether its options are legal. If the field format does not meet the requirements, the system will generate a format error report and mark the error field in the inspection record.

[0060] The system updates the status of the annotation results that pass the inspection to "Inspected" and the status of the annotation results that fail the inspection to "To be modified". This step ensures that the status of the annotation task is updated synchronously in the system for subsequent processing. The system stores the inspection results and error reports in the database and updates the status field of the annotation results. When all fields pass the inspection, the system updates the status to "Inspected"; if there are unqualified fields, the system sets the status to "To be modified" and generates an inspection report for the user to view.

[0061] The accuracy and compliance of the annotation results are ensured by performing multi-dimensional checks on the annotation results and the requirements of the annotation template, including field existence, required field completeness, format consistency, etc. The system automatically performs field checks and format verification, effectively reducing errors caused by manual inspections and improving the efficiency and accuracy of inspections. Through the automatic storage and status update of inspection results, the status of the annotation results is dynamically tracked in the system, providing a high-quality data foundation for subsequent analysis.

[0062] S50, when the annotation result passes the inspection, generating metadata corresponding to the annotation result, the metadata including business field information and annotation task attribute information;

[0063] In this embodiment, the system extracts business field information from the data set and integrates the business field data into the metadata. Business field information is a background description of the annotation results, providing the context and associated information of the annotation results, such as customer type, transaction date, location, etc., to support subsequent analysis tasks. After the annotation results pass the inspection, the system automatically extracts the business field information related to the annotation results from the original data set or the associated information of the annotation task. The system matches and identifies business fields by field name or field position, and integrates their values ​​into the metadata. After the extraction is completed, the system establishes a mapping relationship between the business field and the annotation results to ensure data relevance.

[0064] The annotation task attribute information is descriptive data of the annotation task, usually including the name of the task, creation time, processing priority and other information. This information provides task-level context for the annotation results in the metadata, so as to trace the generation environment of the annotation results and the annotation task settings. The system retrieves the task attribute information from the annotation task management module and extracts the content of the task name, task ID, task priority and other fields into the metadata. The system can associate the annotation results with the task ID and attach the task attribute information to the annotation result entry in the metadata to ensure that the annotation results contain complete context information when stored and analyzed.

[0065] The system integrates the annotation results, business field information, and annotation task attribute information into a metadata entry, ensuring that the metadata contains complete annotation results, task background, and business information for subsequent data analysis and traceability. The system integrates the annotation results, business field information, and annotation task attribute information into a unified metadata entry. The metadata entry structure can be stored in JSON, XML, or database table record formats, and the specific structure depends on the system data storage architecture. The integrated metadata entry is associated with the ID or other identifier of the annotation task for subsequent analysis and call.

[0066] The system stores the generated metadata entries in the database or designated storage module to facilitate subsequent data analysis and task management. The storage of metadata keeps the annotation results consistent with the business fields and task attribute information, ensuring the integrity and traceability of the data. The system stores the integrated metadata entries in the database or other data storage modules. The metadata entries can be associated with meta-information such as task ID, annotation time, and data set identifier to achieve effective data management. The system sets the metadata entry status to "generated" or "pending analysis" so that it can be directly retrieved and applied in the analysis module.

[0067] By generating corresponding metadata after the annotation results pass the inspection, the integrated management of annotation results, business fields and task attribute information is realized. Metadata not only contains the annotation results themselves, but also covers the attributes and business field information of the annotation tasks, so that the analysis of the annotation results has more comprehensive background support. The generation and storage of metadata simplifies the management process of annotation results, ensures data integrity and consistency, and provides accurate and rich background information for subsequent data analysis and decision-making.

[0068] S60: Perform data analysis on the annotation result based on the metadata to generate an analysis result.

[0069] In this embodiment, the system determines the dimensions of data analysis based on the business field information and the annotation task attribute information in the metadata. The analysis dimensions provide a basis for grouping and filtering for data analysis, such as grouping by time, customer type, or risk category. The system reads the business field and task attribute information in the metadata, and determines the analysis dimensions based on preset analysis rules or user selections. For example, the system can set analysis dimensions grouped by "time period" and "customer type", or allow users to select "risk category" as a custom dimension for hierarchical analysis.

[0070] The system groups the annotation results according to the analysis dimension, so that the annotated data can be compared under the same analysis dimension. Through group analysis, the statistical characteristics of the annotation results can be obtained in different dimensions, providing support for indicators such as analysis accuracy and risk distribution. The system groups the annotation results in the metadata according to the determined analysis dimensions (such as time period, customer type, etc.), and uses data query or aggregation functions to group the annotated data of the same dimension. The grouped data is temporarily stored in the system and associated with the ID of the analysis task for further calculation.

[0071] The system performs comparative analysis on the grouped annotation results. The comparative analysis reveals the statistical characteristics of the annotation data by calculating data analysis indicators such as accuracy and recall in each group, and provides key analysis results required by the business. The system applies data analysis calculations to each group. For example, the calculation methods of accuracy and recall are:

[0072] Accuracy = (number of correct annotations / total number of annotations) × 100%

[0073] Recall rate = (number of correct annotations / number of required annotations) × 100%

[0074] The system summarizes each set of calculated indicators and forms preliminary analysis data, which is saved in the system's temporary analysis data set.

[0075] The system generates a final analysis report based on the results of the comparative analysis. The report contains analysis indicators of the grouped data (such as accuracy, recall, risk distribution, etc.). These analysis results are provided to users for data evaluation and decision-making. The system generates a report file or data table with the analysis results in CSV, JSON, etc., which is convenient for export or display on the front-end interface. The analysis results are stored in the database and associated with the annotation task ID and analysis task ID for subsequent calls. The system can be configured to automatically generate analysis reports or support users to manually view the results.

[0076] By determining the analysis dimension based on the business fields in the metadata and the attribute information of the annotation task, the annotation results are grouped, and indicators such as accuracy and recall are calculated to generate detailed analysis results. The system integrates the automatic grouping and analysis of the annotation data, so that the system can quickly generate data analysis reports without manual intervention, improving the efficiency and accuracy of data analysis. The analysis results provide a direct and reliable basis for business decision-making, and can flexibly adjust the analysis dimensions when business needs change, which is suitable for a variety of business analysis scenarios.

[0077] The present invention relates to the fields of big data technology and financial technology, and discloses a data annotation and analysis method, which obtains a data set containing fields to be annotated and business fields, creates an annotation task based on the data set and determines an annotation template, annotates the fields to be annotated based on the annotation template, and generates an annotation result; verifies the annotation result, generates metadata corresponding to the annotation result, and performs data analysis on the annotation result based on the metadata to generate an analysis result. The present invention simplifies the analysis process after data annotation by integrating data annotation and analysis functions in the same system, reduces dependence on additional systems, improves the utilization efficiency of data annotation results, realizes integrated processing of data analysis, and provides support for real-time and accurate analysis.

[0078] In one embodiment, the above S10 includes:

[0079] S101, receiving a data set upload request and verifying the upload authority of the requesting account;

[0080] S102, when the verification is passed, the data set upload operation is performed;

[0081] S103, when the upload operation of the data set is completed, the data set is parsed, the fields to be marked and the business fields are identified from the data set, and a unique identifier of the data set is generated.

[0082] In this embodiment, the system receives a dataset upload request submitted by a user and verifies the upload permissions of the requesting account to ensure that the upload operation is limited to authorized users. This verification mechanism ensures the security of the data and the stability of the system, and prevents unauthorized data from entering the annotation process. The system provides a front-end upload interface or API interface for dataset upload. After the upload request is submitted, the system retrieves the permission information of the user account from the user permission database to verify whether the user has the upload permission. The system implements permission verification through role-based permission control (RBAC) or access control list (ACL)-based methods. If the verification fails, the system will return an error prompt and terminate the upload process.

[0083] When the user passes the permission verification, the system executes the data set upload operation and transfers the data set file to the server or storage module. The upload operation ensures that the data set is stored completely and securely, and prepares the data for subsequent parsing and annotation tasks. The system receives and stores the uploaded files. The upload operation supports large file transfers. If the file is large, the block upload technology can be used to ensure the reliability and speed of the upload. After the file is uploaded, the system generates a success status and temporarily stores the data set in the specified storage path (such as a database or file system). The upload success status will be returned to the user, indicating that the upload operation has been completed.

[0084] After the upload operation is completed, the system parses the data set file to identify the fields to be annotated and the business fields. Data set parsing classifies the data set content by analyzing the file structure and field content to ensure that the annotations and business fields are correctly processed in subsequent processes. The system performs structured parsing based on the format of the uploaded file (such as CSV, JSON, etc.). For example, for CSV files, the system can match and parse fields by column header names or preset field rules; for JSON files, the system parses the field hierarchy. The system uses matching rules or configuration files to identify the fields to be annotated and the business fields, and stores the data content in a structured form after parsing for subsequent annotation operations.

[0085] For ease of management and tracking, the system generates a unique identifier (ID) for each uploaded dataset. The unique identifier makes the dataset unique in the system, ensuring that each dataset can be accurately located and called when multiple uploads or multiple datasets are managed in parallel. The system uses a hash algorithm or a database auto-increment ID to generate a unique identifier (such as a UUID or MD5 hash value). The unique identifier is associated with the dataset metadata to ensure that a specific dataset can be accurately retrieved during annotation tasks and data analysis. The generated unique identifier is stored in the database and associated with the dataset content so that it can be called and tracked by all relevant modules in the system.

[0086] This embodiment implements secure management of data upload by receiving a data set upload request and verifying user permissions. The data set is parsed and the fields to be annotated and the business fields are identified to ensure that the content structure of the data set is complete and meets the annotation requirements. A unique identifier is generated to facilitate the tracking and management of the data set, realizing an integrated process of data set upload, parsing and management, laying a data foundation for subsequent annotation and analysis.

[0087] In one embodiment, the above S20 includes:

[0088] S201, generating basic information of a labeling task based on the unique identifier of the data set, and adding the to-be-labeled field and the business field to the labeling task;

[0089] S202, setting a task name, task description, task priority and task processing parameters in the labeling task, and assigning the labeling task to a preset resource group;

[0090] S203, selecting a labeling template that matches the type and structure of the field to be labeled from a labeling template library according to the requirements of the labeling task;

[0091] S204: configure the required fields and optional fields of the annotation template.

[0092] In this embodiment, the system creates a new labeling task based on the unique identifier of the uploaded data set and generates basic information for the labeling task. The basic information usually includes the unique ID of the labeling task, the task creation time, the data set association, etc., so that the labeling task can be identified and managed in the system. The system reads the unique identifier of the data set and generates a unique task ID and basic information for the labeling task through the task management module. The system associates the unique identifier with the task information and stores it so that the basic information and data set source of the labeling task can be displayed in the task management interface, which is convenient for subsequent task retrieval and management.

[0093] The system adds the fields to be annotated and the business fields in the data set to the annotation task. The fields to be annotated are the main content of the annotation task, and the business fields provide the necessary context information for the annotation task for reference during the annotation and analysis process. The system extracts the fields to be annotated (such as risk level, classification label, etc.) and business fields (such as customer type, transaction time, etc.) from the data set, and imports these field information into the data structure of the annotation task. The system associates the fields to be annotated and the business fields to the task ID of the annotation task to ensure that the task content is complete and the classification is clear, which is convenient for subsequent annotation and data analysis.

[0094] The system sets the name, description, priority, and other task processing parameters for the annotation task, so that the annotator can understand the task goal, urgency, and processing method when performing the task. Task priority and processing parameters are used for automatic task allocation and task process control. The system sets the task name and description through user configuration or system defaults. Task priority can be set according to business needs (such as high, medium, and low), while task processing parameters define the execution rules of the annotation task (such as allowed annotation types, quality inspection requirements, etc.). The system saves these attributes to the annotation task record and automatically assigns tasks according to priority and processing parameters during task scheduling.

[0095] The system selects a suitable template from the annotation template library according to the requirements of the annotation task. The selection of the template is based on the type and structure of the field to be annotated, so as to configure the appropriate annotation rules and field display mode. The system matches the appropriate template according to the type of field to be annotated (such as text, numeric, classification, etc.) and field structure of the annotation task. For example, if the field to be annotated contains "risk level" as a classification item, the system automatically selects a template containing the classification label from the template library. The system associates the selected template with the annotation task through the template ID so that the appropriate fields and annotation rules can be called in the annotation interface.

[0096] After the annotation template is determined, the system automatically sets the required fields and optional fields according to the configuration of the template. Required fields are items that must be completed during the annotation process, while optional fields provide auxiliary information or additional instructions. The system automatically reads the field settings of the template through the template configuration file or template library, marks the required fields as "required", and sets the optional fields as optional. For example, for a loan risk annotation task, the template may stipulate that "risk level" is a required item, while "remarks" is an optional item. The system displays the required and optional items in the annotation interface according to the template configuration, and provides the annotator with corresponding annotation filling instructions.

[0097] This embodiment simplifies and standardizes the configuration process of annotation tasks by generating basic information, adding data fields, setting task attributes, and matching annotation templates when creating annotation tasks. By automatically selecting templates and configuring required and optional fields, the system achieves rapid initialization and modular settings of annotation tasks, reduces the complexity of manual operations, and improves the accuracy and efficiency of annotation tasks.

[0098] In one embodiment, the above S30 includes:

[0099] S301, based on the annotation template, displaying the field to be annotated and the corresponding required fields and optional fields in the annotation interface;

[0100] S302, displaying the data content of the field to be annotated one by one in the annotation interface, and generating an input box or selection item matching the annotation template;

[0101] S303, annotating the field to be annotated based on the input box or the selection item to generate an annotation result;

[0102] S304: Store the annotation results in a database, and set the status of the annotated fields to be annotated as annotated.

[0103] In this embodiment, the system displays the detailed information of the fields to be annotated and the required and optional fields in the annotation interface according to the configuration of the annotation template. Displaying the required and optional fields ensures that the annotator is clear about the filling requirements of the task. The system reads the settings of the required and optional fields from the annotation template, and marks the required fields as "required" in the annotation interface to emphasize their importance. The optional fields are used as auxiliary information for the annotator to fill in freely. The interface layout uses a form or card to display the field content, helping the annotator quickly identify the fields that need to be completed.

[0104] The system displays the specific data content of the fields to be annotated one by one, ensuring that the annotator can view and process each piece of data in turn. At the same time, the system generates appropriate input boxes or selection items according to the field type of the annotation template so that the annotator can accurately enter or select the annotation content. The system displays the data content of the fields to be annotated in the annotation interface. For text fields, the system generates a text input box; for classification or selection type fields, the system generates a drop-down option or radio button. In this way, the annotator can accurately select or enter the content, making the annotation process more operational and accurate.

[0105] The annotator annotates each field to be annotated through the input box or selection items to complete the data annotation operation. The annotation result is the core data output of the annotation process, which includes the annotator's filling and selection results for the fields to be annotated. The annotator checks the data content of the fields to be annotated one by one, and fills in or selects the annotation content in the corresponding input box or selection items. The system automatically records the content filled in or selected by the annotator and regards it as part of the annotation result. After completing the annotation of each data, the system saves the annotation content to the memory or temporary buffer area to ensure data integrity.

[0106] After the labeling process is completed, the system stores the labeling results in the database and updates the status of the labeled fields to "labeled". This operation ensures that the labeled data is stored persistently and records the labeling status for subsequent inspection and analysis. After the labeling is completed, the system writes all the labeling results to the database in batches. Each record contains meta-information such as the content of the field to be labeled, the data filled in by the labeler, and the labeling time. At the same time, the system updates the status field of the field to be labeled to "labeled" and records the task progress in the task management module for subsequent management and viewing.

[0107] This embodiment optimizes the execution process of the annotation task by displaying the fields to be annotated, generating input boxes or options, recording the annotation results, and updating the annotation status in the annotation interface. Displaying the fields to be annotated and the required and optional fields configured in the template effectively improves the operability of the annotation interface and reduces the errors made by the annotators. Input boxes and options are automatically generated according to the field type to ensure the structuring and accuracy of the data. The annotation result storage and status update function improves the data integrity and consistency of the system, and provides a reliable data foundation for subsequent data verification and analysis.

[0108] In one embodiment, the above S40 includes:

[0109] S401, checking whether the required fields in the annotation result have been fully filled in, and checking whether the optional fields in the annotation result comply with the format requirements in the annotation template;

[0110] S402, performing compliance check on the value range, data type and option range of the annotation result according to the compliance requirements in the annotation template;

[0111] S403, updating the status of the annotation results that have passed the inspection to "inspected", and updating the status of the annotation results that have not passed the inspection to "to be modified".

[0112] In this embodiment, the system checks the required fields in the annotation results one by one to ensure that all required fields are filled in and the content is not empty. This step ensures the integrity of the annotation data and avoids missing data. The system traverses the required fields in the annotation result record and checks whether each field has valid data. If the field is empty or contains an invalid value (such as a space or placeholder), the system will record the annotation result as unqualified. After the inspection is completed, the system generates a field integrity report and marks the records that need to be modified.

[0113] The system performs format verification on the optional field content according to the annotation template to ensure that the optional field content conforms to the specified format or data type. This process ensures that the information in the optional field conforms to the annotation template requirements in terms of format, which is convenient for subsequent use and data consistency maintenance. The system reads the format requirements of the optional field from the annotation template, such as text length limit, numeric type, or specific character format. Then the optional field content of each annotation result is checked, such as verifying the text length, character type, etc., to ensure that the optional field format conforms to the template requirements. If the field format does not match, the system generates a format error report and marks the error field.

[0114] The system performs compliance checks on the numeric fields, data types, and option fields in the annotation results to ensure that the field content meets the numeric range and option range requirements in the annotation template. This check ensures the accuracy and consistency of the annotation data and prevents invalid data from entering the database. The system reads the range, data type, and option requirements of the numeric field from the annotation template and verifies one by one whether the content of the annotation results meets the requirements of the template. For example, the system checks whether the numeric field is within the specified range and verifies whether the content of the option field is one of the allowed options. After all compliance checks are completed, the system records the fields that do not meet the requirements and generates a compliance inspection report.

[0115] The system updates the status of the annotation results that pass the inspection to "Inspected" and the status of the annotation results that fail the inspection to "To be modified". This operation ensures that the status of the annotation results is synchronized in the system, which is convenient for subsequent modification and management. The system sets the status field of all annotation results that pass the inspection to "Inspected" and stores this status in the database. For the annotation results that fail the inspection, the system updates the status field to "To be modified" and generates an inspection report for subsequent review and processing of unqualified data.

[0116] This embodiment implements multi-level quality control of the annotation results by checking the completeness of the required fields, the format requirements of the optional fields, and the conformity of the value range and the option range of the annotation results one by one. Through the status update function, the system effectively manages and tracks the qualified status of the annotation results to ensure the accuracy and consistency of the annotation data. The automatically generated inspection report provides an intuitive error record, which is convenient for users to modify and improve unqualified data.

[0117] In one embodiment, the above S50 includes:

[0118] S501, extracting business field information associated with the annotation result from the data set, and combining the business field information with the annotation result into a complete annotation record;

[0119] S502, generating a metadata entry including annotation field content, business field information and annotation task attribute information based on the annotation record;

[0120] S503: Store the metadata entry in a database.

[0121] In this embodiment, the system extracts business field information associated with the annotation results from the data set to provide context and business information support for the annotation results. Business field information can include source information of the data, correlation data, etc., for subsequent data analysis. The system extracts business field information (such as customer type, transaction time, product classification, etc.) from the original data set through the field association rules configured in the annotation task, and associates these fields with the annotation result content. The system automatically matches the business fields in the data set according to the field name or identifier, and temporarily stores them after extraction to generate a complete annotation record.

[0122] The system combines the extracted business field information with the annotation results to generate a complete annotation record containing the annotation field content, business field information, and annotation task attribute information. This record is the metadata entry and contains all relevant information. After the annotation results pass the inspection, the system combines the annotation field content with the extracted business field information into a complete annotation record. The system retrieves the attribute information of the annotation task (such as task ID, task name, creation time) from the annotation task configuration and appends this information to the annotation record to form a complete metadata entry structure. This entry contains both the annotation data content and business background information.

[0123] The system stores the generated metadata entries in the database to facilitate subsequent data retrieval, analysis and management operations. The persistent storage of metadata entries ensures the security, integrity and traceability of the annotation result data. The system writes the metadata entries to the database, and the database table structure is designed according to the fields required by the metadata entries for fast storage and retrieval. The system can assign a unique identifier (such as an automatically generated ID) to each metadata record for data tracking and management. During the storage process, the system backs up the entries or marks the storage status to ensure the reliability of the data.

[0124] This embodiment achieves an effective combination of annotation results and business background information by extracting business field information from the data set and generating metadata entries. Storing metadata entries in the database ensures the integrity and persistence of the annotated data, facilitating subsequent data retrieval and analysis. By integrating business fields, annotation fields, and task attributes, the system has stronger contextual support during data retrieval, providing a reliable data foundation for multi-dimensional data analysis and business applications.

[0125] In one embodiment, the above S60 includes:

[0126] S601, determining an analysis dimension based on the business field information and the annotation task attribute information in the metadata;

[0127] S602: Group the annotation results according to the analysis dimension, and compare and analyze the grouped annotation results based on a reference annotation data set to generate data analysis indicators.

[0128] In this embodiment, the system uses the business field information (such as customer type, transaction date) and annotation task attribute information (such as task name, risk level) in the metadata to determine the analysis dimensions, so that the annotation results can be effectively grouped and compared during subsequent data analysis. The analysis dimensions determine the way the data is aggregated and filtered to ensure that the analysis results are targeted and practical. The system extracts business field information and task attribute information from the metadata entries, and selects the analysis dimensions through preset rules or user-defined configurations. For example, the system can select "customer type" and "risk level" as analysis dimensions to observe the risk distribution of different customer types. The determination of the analysis dimensions can be adjusted based on business needs, and the system saves the selected dimensions for subsequent analysis.

[0129] The system groups the annotation results according to the determined analysis dimensions to form annotated data sets under different dimensions. This process enables data to be aggregated according to specific conditions, providing a structured basis for comparative analysis. The system groups the annotation results in the metadata according to the selected analysis dimensions. The system can use SQL queries or aggregate functions to group annotated data of the same dimension into the same group. For example, if "customer type" is used as a dimension, the system groups the annotation results by customer type to form an annotated data set corresponding to each customer type. After the grouping is completed, the system stores the data in a temporary analysis table or cache for subsequent processing.

[0130] The system compares the grouped annotation results with the reference annotation data set and calculates the key indicators of each group of data to evaluate the accuracy, coverage, etc. of the annotation results. This process helps users understand the overall performance of the annotation data and identify the completion and effectiveness of the annotation tasks. The system retrieves the corresponding standard results from the reference annotation data set and compares the grouped annotation results with the standard data item by item. During the comparison analysis, the system calculates indicators such as accuracy and recall for each grouped data. For example, the system compares the correct annotation items in each group of annotation results with the reference data, generates accuracy and recall results, and records the comparison results in a temporary analysis table.

[0131] The system generates data analysis indicators such as accuracy, recall, risk distribution, etc. based on the results of the comparison analysis. Data analysis indicators provide users with intuitive quality assessment of annotation results and provide data support for subsequent business decisions. The system calculates the accuracy (number of correct annotations / total number of annotations) and recall (number of correct annotations / number of annotations that should be marked) under each analysis dimension based on the comparison analysis results, and writes the generated indicators into the analysis result table. The system can display the analysis results in the form of charts or reports. After the analysis indicators are summarized, the system can generate a report file or directly display the results in the interface.

[0132] This embodiment extracts business fields and annotation task attribute information from metadata, dynamically generates analysis dimensions, performs group analysis and comparison on the annotation results, and generates data analysis indicators. The system realizes automated data analysis of annotation results, and users can intuitively understand key indicators such as accuracy and recall rate of annotation results. Through the flexible setting of analysis dimensions, the system supports multiple analysis perspectives to meet different business needs, improves the efficiency and reliability of annotation result analysis, and provides strong data support for subsequent business decisions.

[0133] In one embodiment, a data labeling and analysis device is provided, and the data labeling and analysis device corresponds one-to-one to the data labeling and analysis method in the above embodiment. Figure 3 , Figure 3 This is a functional module diagram of a preferred embodiment of the data annotation and analysis device of the present invention. Data management module 10, task management module 20, annotation processing module 30, inspection module 40, metadata generation module 50 and data analysis module 60. The functional modules are described in detail as follows:

[0134] The data management module 10 is used to obtain a data set including fields to be annotated and business fields;

[0135] A task management module 20, used to create a labeling task based on the data set, and determine a labeling template according to the labeling task;

[0136] A labeling processing module 30, used to label the to-be-labeled field according to the labeling task and the labeling template, and generate a labeling result;

[0137] A verification module 40, used to verify the annotation result according to the annotation template;

[0138] A metadata generation module 50, for generating metadata corresponding to the annotation result when the annotation result passes the inspection, the metadata including business field information and annotation task attribute information;

[0139] The data analysis module 60 is used to perform data analysis on the annotation results based on the metadata to generate analysis results.

[0140] In one embodiment, the data management module 10 is specifically used for:

[0141] Receive data set upload requests and verify the upload permissions of the requesting account;

[0142] When the verification passes, the data set upload operation is performed;

[0143] When the upload operation of the data set is completed, the data set is parsed, the fields to be marked and the business fields are identified from the data set, and a unique identifier of the data set is generated.

[0144] In one embodiment, the task management module 20 is specifically used to:

[0145] Generate basic information of the annotation task based on the unique identifier of the data set, and add the to-be-annotated field and the business field to the annotation task;

[0146] Setting a task name, task description, task priority and task processing parameters in the labeling task, and assigning the labeling task to a preset resource group;

[0147] According to the requirements of the annotation task, a annotation template matching the type and structure of the field to be annotated is selected from an annotation template library;

[0148] Configure the required and optional fields of the annotation template.

[0149] In one embodiment, the annotation processing module 30 is specifically used to:

[0150] Based on the annotation template, the fields to be annotated and the corresponding required fields and optional fields are displayed in the annotation interface;

[0151] Displaying the data content of the field to be annotated one by one in the annotation interface, and generating an input box or selection item matching the annotation template;

[0152] Annotate the field to be annotated based on the input box or selection item to generate an annotation result;

[0153] The marking results are stored in a database, and the status of the marked fields to be marked is set to marked.

[0154] In one embodiment, the inspection module 40 is specifically configured to:

[0155] Check whether the required fields in the annotation result have been fully filled in, and check whether the optional fields in the annotation result meet the format requirements in the annotation template;

[0156] Performing a compliance check on the numerical range, data type and option range of the annotation result according to the compliance requirements in the annotation template;

[0157] The status of the annotation results that passed the inspection is updated to inspected, and the status of the annotation results that failed the inspection is updated to pending modification.

[0158] In one embodiment, the metadata generation module 50 is specifically used to:

[0159] Extracting business field information associated with the annotation result from the data set, and combining the business field information with the annotation result into a complete annotation record;

[0160] Generate metadata entries including annotation field content, business field information, and annotation task attribute information based on the annotation record;

[0161] The metadata entry is stored in a database.

[0162] In one embodiment, the data analysis module 60 is specifically used to:

[0163] Determine the analysis dimension based on the business field information and the annotation task attribute information in the metadata;

[0164] The annotation results are grouped according to the analysis dimension, and the grouped annotation results are compared and analyzed based on a reference annotation data set to generate data analysis indicators.

[0165] In one embodiment, a computer device is provided. The computer device may be a server, and its internal structure diagram may be as follows: Figure 4 As shown. The computer device includes a processor, a memory, a network interface and a database connected via a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile and / or volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The network interface of the computer device is used to communicate with an external user terminal through a network connection. When the computer program is executed by the processor, it realizes the functions or steps of a data labeling and analysis method service side.

[0166] In one embodiment, a computer device is provided. The computer device may be a user terminal, and its internal structure diagram may be as follows: Figure 5As shown. The computer device includes a processor, a memory, a network interface, a display screen and an input device connected via a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The network interface of the computer device is used to communicate with an external server via a network connection. When the computer program is executed by the processor, it realizes the functions or steps on the user side of a data annotation and analysis method

[0167] In one embodiment, a computer device is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, the following steps are implemented:

[0168] Get the data set containing the fields to be annotated and the business fields;

[0169] Based on the data set, create a labeling task, and determine a labeling template according to the labeling task;

[0170] Annotate the field to be annotated according to the annotation task and the annotation template to generate an annotation result;

[0171] Checking the annotation result according to the annotation template;

[0172] When the annotation result passes the inspection, metadata corresponding to the annotation result is generated, the metadata including business field information and annotation task attribute information;

[0173] The annotation result is analyzed based on the metadata to generate an analysis result.

[0174] In one embodiment, a computer readable storage medium is provided, on which a computer program is stored, and when the computer program is executed by a processor, the following steps are implemented:

[0175] Get the data set containing the fields to be annotated and the business fields;

[0176] Based on the data set, create a labeling task, and determine a labeling template according to the labeling task;

[0177] Annotate the field to be annotated according to the annotation task and the annotation template to generate an annotation result;

[0178] Checking the annotation result according to the annotation template;

[0179] When the annotation result passes the inspection, metadata corresponding to the annotation result is generated, the metadata including business field information and annotation task attribute information;

[0180] The annotation result is analyzed based on the metadata to generate an analysis result.

[0181] It should be noted that the above functions or steps that can be implemented by the computer-readable storage medium or computer device can refer to the relevant descriptions on the server side and the user side in the aforementioned method embodiment. To avoid repetition, they will not be described one by one here.

[0182] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. As an illustration and not limitation, RAM is available in many forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).

[0183] Those skilled in the art can clearly understand that for the convenience and simplicity of description, only the division of the above-mentioned functional units and modules is used as an example. In actual applications, the above-mentioned functions can be distributed and completed by different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above.

[0184] It should be noted that if software tools or components other than those of the Company appear in the embodiments of the present application, they are only used for illustration and do not represent actual use. The above-described embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit them; although the present invention has been described in detail with reference to the above-mentioned embodiments, a person of ordinary skill in the art should understand that the technical solutions described in the above-mentioned embodiments can still be modified, or some of the technical features thereof can be replaced by equivalents; and these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be included in the protection scope of the present invention.

Claims

1. A data annotation and analysis method, characterized in that: The following steps are involved: Get the data set containing the fields to be annotated and the business fields; Based on the data set, create a labeling task, and determine a labeling template according to the labeling task; Annotate the field to be annotated according to the annotation task and the annotation template to generate an annotation result; Checking the annotation result according to the annotation template; When the annotation result passes the inspection, metadata corresponding to the annotation result is generated, the metadata including business field information and annotation task attribute information; The annotation result is analyzed based on the metadata to generate an analysis result.

2. The data annotation and analysis method according to claim 1, characterized in that: Get a data set containing the fields to be annotated and business fields, including: Receive data set upload requests and verify the upload permissions of the requesting account; When the verification passes, the data set upload operation is performed; When the upload operation of the data set is completed, the data set is parsed, the fields to be marked and the business fields are identified from the data set, and a unique identifier of the data set is generated.

3. The data annotation and analysis method according to claim 1, characterized in that: Based on the data set, a labeling task is created, and a labeling template is determined according to the labeling task, including: Generate basic information of the annotation task based on the unique identifier of the data set, and add the to-be-annotated field and the business field to the annotation task; Setting a task name, task description, task priority and task processing parameters in the labeling task, and assigning the labeling task to a preset resource group; According to the requirements of the annotation task, a annotation template matching the type and structure of the field to be annotated is selected from an annotation template library; Configure the required and optional fields of the annotation template.

4. The data annotation and analysis method as claimed in claim 1, characterized in that: The field to be annotated is annotated according to the annotation task and the annotation template to generate an annotation result, including: Based on the annotation template, the fields to be annotated and the corresponding required fields and optional fields are displayed in the annotation interface; Displaying the data content of the field to be annotated one by one in the annotation interface, and generating an input box or selection item matching the annotation template; Annotate the field to be annotated based on the input box or selection item to generate an annotation result; The marking results are stored in a database, and the status of the marked fields to be marked is set to marked.

5. The data annotation and analysis method according to claim 1, characterized in that: Checking the annotation result according to the annotation template includes: Check whether the required fields in the annotation result have been fully filled in, and check whether the optional fields in the annotation result meet the format requirements in the annotation template; Performing a compliance check on the numerical range, data type and option range of the annotation result according to the compliance requirements in the annotation template; The status of the annotation results that passed the inspection is updated to inspected, and the status of the annotation results that failed the inspection is updated to pending modification.

6. The data annotation and analysis method according to claim 1, characterized in that: When the annotation result passes the inspection, metadata corresponding to the annotation result is generated, and the metadata includes business field information and annotation task attribute information, including: Extracting business field information associated with the annotation result from the data set, and combining the business field information with the annotation result into a complete annotation record; Generate metadata entries including annotation field content, business field information, and annotation task attribute information based on the annotation record; The metadata entry is stored in a database.

7. The data annotation and analysis method according to claim 1, characterized in that: Performing data analysis on the annotation result based on the metadata to generate analysis results includes: Determine the analysis dimension based on the business field information and the annotation task attribute information in the metadata; The annotation results are grouped according to the analysis dimension, and the grouped annotation results are compared and analyzed based on a reference annotation data set to generate data analysis indicators.

8. A data labeling and analysis device, characterized in that: The data annotation and analysis device comprises: The data management module is used to obtain the data set containing the fields to be annotated and the business fields; A task management module, used to create a labeling task based on the data set, and determine a labeling template according to the labeling task; A labeling processing module, used to label the to-be-labeled field according to the labeling task and the labeling template, and generate a labeling result; A verification module, used for verifying the annotation result according to the annotation template; A metadata generation module, used to generate metadata corresponding to the annotation result when the annotation result passes the inspection, the metadata including business field information and annotation task attribute information; A data analysis module is used to perform data analysis on the annotation results based on the metadata to generate analysis results.

9. A computer device, characterized in that: The computer device includes a memory, a processor, and a data labeling and analysis program stored in the memory and executable on the processor. When the data labeling and analysis program is executed by the processor, the steps of the data labeling and analysis method according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium, characterized in that: The storage medium stores a data labeling and analysis program, which, when executed by a processor, implements the steps of the data labeling and analysis method according to any one of claims 1 to 7.

Images