Risk control data mart index analysis method and system based on AI

Through the AI-based risk control data mart indicator analysis method, the problem of multiple index versions and redundant development logic in the risk control data mart is solved, the efficiency of indicator development and the accuracy of risk management are achieved, and the operation of enterprises is improved the robustness.

CN118963712APending Publication Date: 2024-11-15WUHAN ZBANK CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202411038677.4
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-07-31
Publication Date
2024-11-15

AI Technical Summary

Technical Problem

The existing risk control data mart has many index versions, chaotic development logic, and inconsistent indicator data, which leads to the inability to judge the importance and attention, and the development work is duplicated, making it difficult to effectively manage risk control.

Method used

Using AI-based risk control data mart indicator analysis method, the index processing logic is obtained through the database and open source tools of the risk control data mart, the business indicator analysis script is generated and script optimization is performed, and the prompt words and role requirements are written using the AI ​​interface to perform Indicator development and optimization.

Benefits of technology

Improve the accuracy and efficiency of metric development, reduce duplicate work, help developers and analysts better understand and manage risks, and ensure the robustness and sustainability of enterprise operations.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118963712A_ABST
    Figure CN118963712A_ABST
Patent Text Reader

Abstract

The invention provides an AI-based risk control data mart index analysis method and system. The method comprises the steps of obtaining a risk control index development data dictionary by applying a risk control data mart metadata management platform; developing back-end connection and collecting data dictionary information in the MPP database, and collecting table and field distribution conditions through an MPP data dictionary table; using a java development back end to connect an MPP database, and obtaining a database execution plan; developing an access AI interface, writing prompt words to create roles and requirements, and generating index analysis and index development scripts; and risk control index development analysis visual development is carried out, and front-end development is carried out on index analysis development and optimization functions. Aiming at the problem that development and analysis standards of developers and analysts in index development in application layer reports of data marts of data warehouses are not uniform, the method is used for developing improvement and optimization schemes in the aspect of repetitive work, assists developers or analysts in index development, analysis and optimization, and improves the accuracy and efficiency of decision making.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the field of data warehouse and data development technology, and in particular to an AI-based risk control data mart indicator analysis method and system. Background Art

[0002] AI, or Artificial Intelligence, refers to the technology and science of machines simulating human intelligence.

[0003] It aims to enable computer systems to perform tasks that normally require human intelligence, such as learning, reasoning, problem solving, understanding language, recognizing images and sounds, making decisions, etc.

[0004] AI technology covers multiple fields and methods. Machine learning is an important branch of AI, including supervised learning, unsupervised learning, and reinforcement learning. Through these learning methods, computers can automatically discover patterns and rules from large amounts of data and make predictions and decisions based on them.

[0005] In image recognition, AI systems can learn to recognize different objects and scenes by studying a large number of pictures. Natural language processing is also one of the key areas of AI. It is committed to enabling computers to understand and generate human language. Applications such as intelligent customer service and machine translation are based on natural language processing technology. AI also plays an important role in many fields such as autonomous driving, medical diagnosis, financial forecasting, and intelligent recommendation systems. For example, self-driving cars rely on AI to perceive the surrounding environment and make driving decisions to ensure driving safety. It is a technology with great potential and influence that is constantly changing the way we live and work.

[0006] MPP (Massively Parallel Processing) is a computing technology that uses a large number of processors (or computing nodes) to perform multiple tasks simultaneously to improve processing speed and efficiency. MPP technology is widely used in big data processing, scientific computing, artificial intelligence and other fields. For example, many big data processing platforms (such as Hadoop and Spark) and database systems (such as Greenplum and Teradata) use MPP architecture.

[0007] Indicator metadata is data about business indicators. It describes the definition, calculation method, data source, data quality, related business rules and other information of the indicator. By managing and using indicator metadata, organizations can better understand their business performance, make more accurate decisions, and improve business efficiency. At the same time, indicator metadata is also an important part of data governance, which helps to improve the quality and consistency of data.

[0008] The existing risk control data mart has multiple indicator versions, complicated development logic, inconsistent indicator data, inability to judge the importance and attention of indicators, and repeated indicator development work. Through effective risk control indicator management, we can better grasp the risk situation, take corresponding measures in time, and ensure the stability and sustainability of enterprise operations. Summary of the invention

[0009] In response to the technical problems existing in the prior art, the present invention provides an AI-based risk control data mart indicator analysis method and system, which obtains the indicator processing logic through the risk control data mart, the database and open source tools based on the mart, and outputs business indicator analysis scripts through AI customization, and performs indicator script optimization prompt functions.

[0010] According to a first aspect of the present invention, the present invention provides an AI-based risk control data mart indicator analysis method, comprising:

[0011] Step 1. Obtain risk control indicators and develop a data dictionary by using the risk control data mart metadata management platform;

[0012] Step 2. Develop the backend to connect and collect data dictionary information in the MPP database, and collect table and field distribution information through the MPP data dictionary table;

[0013] Step 3. Use Java to develop the backend to connect to the MPP database and obtain the database execution plan for SQL script optimization;

[0014] Step 4. Develop access AI interfaces, write prompt words to create roles and requirements, and use them to generate indicator analysis and indicator development scripts;

[0015] Step 5. Conduct visual development of risk control indicator analysis, and perform front-end development of indicator analysis and optimization functions.

[0016] Based on the above technical solution, the present invention can also make the following improvements.

[0017] Optionally, the risk control indicator development data dictionary includes: table structure, table business logic, field type, field business logic, and table association relationship.

[0018] Optionally, the step of acquiring risk control indicators and developing a data dictionary by applying the risk control data mart metadata management platform includes the following steps:

[0019] 1.1. Access the metadata management platform and conduct in-depth analysis of the data structure, data volume, and data distribution in the database;

[0020] 1.2. Find risk control data marts, locate and select data marts or data warehouses containing risk control data;

[0021] 1.3. Find the table structure and business logic to find the data table containing risk control indicators;

[0022] 1.4. Understand the field type and business logic. For each field, obtain the field type and field business logic;

[0023] 1.5. Determine the relationship between tables and describe the relationship between different tables;

[0024] 1.6. Document the acquired data dictionary information and ensure regular updates;

[0025] 1.7. Ensure the data dictionary is visible and accessible to the entire team.

[0026] Optionally, the table and field distribution information such as partitions, data volume, and maximum and minimum value data are collected through the MPP data dictionary table.

[0027] Optionally, collecting table and field distribution information through the MPP data dictionary table includes:

[0028] 2.1. Log in and connect to the MPP database, make sure you have permission to access the MPP database and have the credentials to connect to the database;

[0029] 2.2. Query system tables or metadata tables

[0030] 2.3. Use SQL query statements to retrieve information about tables from system tables or metadata tables;

[0031] 2.4. Get the statistical information of the fields from the system table;

[0032] 2.5. Query based on foreign key constraints or index information in the system table;

[0033] 2.6. Export the data dictionary information obtained from the MPP database as a document or spreadsheet.

[0034] Optionally, the use of Java to develop a backend to connect to an MPP database and obtain a database execution plan includes the following:

[0035] 3.1. Connect to the database, execute SQL statements through client tools, and obtain the execution plan;

[0036] 3.2. Analyze the execution plan;

[0037] 3.3. Analyze and optimize queries, identify bottlenecks and optimization opportunities in queries.

[0038] Optionally, the parsing execution plan includes:

[0039] Extract key features from the execution plan, use the key features as optimization prompts for the AI ​​model, and use the execution plan optimization suggestion table as an alternative to assist in SQL script optimization; the key features include execution time, resource consumption, and data access method.

[0040] Optionally, the writing prompt to create a role and requirements includes:

[0041] To create prompts for SQL developers and data analysts, you need to add and modify the prompts based on the characteristics of different roles.

[0042] According to a second aspect of the present invention, there is provided an AI-based risk control data mart indicator analysis system, comprising:

[0043] The data acquisition module is used to acquire risk control indicators and develop a data dictionary by using the risk control data mart metadata management platform;

[0044] The first data processing module is used to develop backend connections and collect data dictionary information in the MPP database, and collect table and field distribution information through the MPP data dictionary table;

[0045] The second data processing module is used to use Java to develop the backend to connect to the MPP database and obtain the database execution plan for SQL script optimization;

[0046] The third data processing module is used to develop access AI interfaces, write prompt words to create roles and requirements, and generate indicator analysis and indicator development scripts;

[0047] The indicator data development module is used for visual development of risk control indicator analysis, and front-end development of indicator analysis and optimization functions.

[0048] The present invention also provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, an AI-based risk control data mart indicator analysis method as described in any one of the above is implemented.

[0049] Technical effects and advantages of the present invention:

[0050] The present invention provides an AI-based risk control data mart indicator analysis method and system, which obtains the indicator processing logic through the risk control data mart, the database based on the mart and the open source tool, and realizes the output of business indicator analysis scripts through AI customization, and performs indicator script optimization prompt functions. The present invention is used to improve and optimize solutions for the development and analysis standards of developers and analysts in the indicator development in the application layer report of the data warehouse data mart, and develop repetitive work. Assist developers or analysts in indicator development, analysis and optimization, and improve the accuracy and efficiency of decision-making.

[0051] Other features and advantages of the present invention will be described in the following description, and partly become apparent from the description, or understood by practicing the present invention. The purpose and other advantages of the present invention can be realized and obtained by the structures pointed out in the description, claims and drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] Figure 1 It is a flow chart of an AI-based risk control data mart indicator analysis method provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0053] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be clearly and completely described below in conjunction with the drawings of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0054] It is understandable that based on the defects in the background technology, the embodiment of the present invention proposes an AI-based risk control data mart indicator analysis method, specifically as follows: Figure 1 As shown, the method comprises the following steps:

[0055] Step 1. Obtain the existing risk control indicators and develop a data dictionary through the existing application risk control data mart metadata management platform.

[0056] Step 1. Obtain risk control indicators and develop a data dictionary by using the risk control data mart metadata management platform;

[0057] The risk control indicator development data dictionary specifically includes: table structure, table business logic, field type, field business logic and table association relationship;

[0058] To obtain the data dictionary for risk control indicator development through the existing application risk control data mart metadata management platform, you can follow the steps below:

[0059] 1.1. Access the metadata management platform and understand indicator data;

[0060] Conduct in-depth analysis of the data structure, data volume, data distribution, etc. in the database. For example, understand the relationship between tables, the data type and value range of the fields, and use access permissions to log in to the metadata database and corresponding tables to access the metadata information stored therein.

[0061] 1.2. Find a risk control data mart;

[0062] Locate and select data marts or data warehouses that contain risk control data. This may be accomplished through search capabilities, directories, or specific data mart identifiers.

[0063] 1.3. Lookup table structure and business logic;

[0064] Find the data tables containing risk control indicators. For each table, you can view the following information:

[0065] Table structure: lists the fields in the table and their types;

[0066] Table business logic: describes the purpose and meaning of the data stored in the table, as well as its role and function in risk control.

[0067] 1.4. Understand field types and business logic;

[0068] For each field, obtain the following information:

[0069] Field type: The data dictionary should detail the data type of each field (such as string, integer, date, etc.).

[0070] Field business logic: describes the meaning of the field, data source, processing method, and specific use in risk control models or strategies.

[0071] 1.5. Determine the relationship between tables;

[0072] In the data dictionary, the relationships between different tables need to be clearly described, especially in risk control data, which may involve joint analysis of data, risk assessment or report generation. The description of the relationship should include the associated fields, relationship type (such as one-to-one, one-to-many) and the associated business logic.

[0073] 1.6. Documentation and updating;

[0074] Document the acquired data dictionary information and ensure that it is updated regularly. This helps team members understand the structure and meaning of the data while ensuring the consistency and understandability of the data.

[0075] 1.7. Use and Sharing;

[0076] Ensuring that the data dictionary is visible and accessible to the entire team can help developers, analysts, and other stakeholders better understand and leverage risk management data.

[0077] Through the above steps, the data dictionary for risk control indicator development can be effectively extracted and managed from the metadata management platform to support application risk control data management and analysis needs.

[0078] Provides a data foundation for AI-assisted SQL script generation through indicator metadata tables.

[0079] Step 2. Develop the backend to connect and collect data dictionary information in the MPP database, and collect table and field distribution information through the MPP data dictionary table;

[0080] The distribution of the tables and fields, such as partitions, data volume, maximum and minimum values, etc.;

[0081] The connection and collection of data dictionary information in the MPP database to understand the distribution of tables and fields can be completed by the following steps:

[0082] 2.1. Log in and connect to the MPP database;

[0083] First, make sure you have access to the MPP database and have the credentials to connect to the database (such as username and password).

[0084] 2.2. Query system tables or metadata tables;

[0085] MPP databases usually have dedicated system tables or metadata tables to store information about objects (such as tables and columns) in the database. The specific table and field names may vary, depending on the specific MPP database system used (such as Greenplum, Vertica, Redshift, etc.).

[0086] 2.3.Query table information;

[0087] Use SQL query statements to retrieve information about tables from system tables or metadata tables. The following are some examples of commonly used queries and information items:

[0088] Basic information of the query table:

[0089] SELECT table_name,table_schema,table_type,create_time

[0090] FROM information_schema.tables

[0091] WHERE table_schema='your_schema';

[0092] Query the number of rows and partitions of the table:

[0093] SELECT schemaname,tablename,partitiontablename,partitionname,partitiontype,partition_column,row_count

[0094] FROM pg_partitions

[0095] Where schemaname='your_schema'and tablename='your_table';

[0096] Query table field information:

[0097] SELECT table_name,column_name,data_type,is_nul label,character_maximum_length

[0098] FROM information_schema.columns

[0099] WHERE table_schema='your_schema'AND table_name='your_table';

[0100] 2.4.Query the statistical information of the field;

[0101] For some MPP databases, you can directly obtain the statistical information of the fields from the system table, such as the maximum value, minimum value, average value, etc. Here are some examples:

[0102] Query the maximum and minimum values ​​of a field

[0103] SELECT column_name,MAX(column_name)AS max_value,MIN(column_name)ASmin_value

[0104] FROM your_schema.your_table;

[0105] 2.5. Combine multi-table association information (if necessary);

[0106] If you need to know the relationship between tables or the association information of specific fields, you can query them in combination with the foreign key constraints or index information in the system table.

[0107] 2.6. Export and document data dictionary;

[0108] Export the data dictionary information obtained from the MPP database as a document or spreadsheet for sharing and reference by team members. This can help the team understand the data structure, distribution, and statistical information, so as to manage and analyze data more effectively.

[0109] Different MPP database systems may have different system table structures and field naming, which need to be adjusted according to the specific system to ensure that performance issues are taken into account when executing queries, especially when large amounts of data or complex queries are involved. Through the above steps, you can effectively connect to the MPP database and collect and manage the distribution and statistical information of tables and fields from it to support back-end development and data management.

[0110] Step 3. Use Java to develop the backend to connect to the MPP database and obtain the database execution plan for SQL script optimization;

[0111] It should be noted that obtaining the execution plan includes: obtaining the execution plan of the current query or operation from the database system; this includes the query execution steps, index usage, and detailed information on how to read data.

[0112] In this embodiment, connecting to the backend processing MPP database and executing statements to obtain a database execution plan involves the following steps:

[0113] 3.1. Connect to the database, execute the statement and obtain the execution plan;

[0114] The method to execute SQL statements through client tools and obtain the execution plan is as follows:

[0115] Using EXPLAIN or EXPLAIN ANALYZE

[0116] In most MPP databases, you can use the EXPLAIN keyword to analyze the execution plan of a SQL query. For example:

[0117] EXPLAIN SELECT*FROM your_schema.your_table WHERE condition_column=

[0118] 'some_value';

[0119] EXPLAIN returns the execution plan of the query, which describes how the database engine executes the query, including the indexes used, scanning methods, connection methods, etc.

[0120] Use EXPLAIN ANALYZE to obtain detailed execution plans and actual execution times

[0121] The MPP database supports EXPLAIN ANALYZE, which executes queries and returns detailed execution plans and actual execution time information. This is very helpful for analyzing query performance bottlenecks and optimizing them.

[0122] EXPLAIN ANALYZE SELECT*FROM your_schema.your_table WHERE condition_column='some_value';

[0123] 3.2. Analyze the execution plan;

[0124] The execution plan may return a lot of information, including query plan tree, cost estimation, data distribution, connection type, etc. To deeply understand how to parse the execution plan, you usually need to be familiar with the execution plan output format and keyword meaning of the specific MPP database.

[0125] 3.3. Analyze and optimize queries;

[0126] By analyzing the execution plan, you can identify bottlenecks and optimization opportunities in the query. Common optimizations include using appropriate indexes, optimizing join methods, and refactoring queries.

[0127] Privilege management: Ensure that the account used for the connection has sufficient privileges to execute EXPLAIN and EXPLAINANALYZE.

[0128] Query Complexity: For complex queries, the execution plan can be complex. Patience and understanding are required to parse and understand the output.

[0129] Performance impact: Executing EXPLAIN ANALYZE actually executes the query, which may have a certain impact on the production system. Therefore, use it with caution in a production environment.

[0130] Through the above steps, you can connect to the MPP database and obtain and parse the execution plan of the SQL query to perform performance analysis and optimization.

[0131] In the above technical solutions, parsing the MPP execution plan and providing optimization suggestions usually involve the following key steps and considerations:

[0132] 1) Obtain the execution plan; import the MPP jdbc driver jar file, and load the PostgreSQL driver in the Java backend development logic:

[0133] Class.forName("org.postgresql.Driver")

[0134] Then establish the connection;

[0135] Connection connection=DriverManager.getConnection(ur l,user,password);

[0136] Execute the following SQL statement to obtain the execution plan:

[0137] EXPLAIN SELECT*FROM your_schema.your_table WHERE condition_column=

[0138] 'some_value';

[0139] Based on the experience of execution plans before and after optimization, such as table scanning method (index, partition or full table scan), table joining method (nested loop join, hash join, etc.);

[0140] Extract important node types and sequences: Execution plans are usually organized according to node types (such as Seq Scan, Index Scan, Hash Join, etc.) and their execution order.

[0141] 2) Analyze the execution plan and extract key features from it, such as execution time, resource consumption, data access method, etc.

[0142] These features will be used as optimization prompts for the AI ​​model, and the execution plan optimization suggestion table will be used as an alternative to assist in SQL script optimization.

[0143] The execution plan is usually returned in a tree structure, and the specific output may vary depending on different MPP databases.

[0144] The key points include: Node type and order: Execution plans are usually organized by node type (such as Seq Scan, IndexScan, Hash Join, etc.) and their execution order.

[0145] Cost estimation: Each execution node may estimate the cost, which reflects the resource consumption of query execution.

[0146] Data distribution and join method: The execution plan also displays the data distribution (such as distribution key, partition key) and the join method used (such as Nested Loop Join, Hash Join).

[0147] 2. Optimization suggestion strategy

[0148] Based on the analysis of the execution plan, such as table scans, sorts, joins, aggregations, indexes, partitions, etc., as well as the amount of data, number of fields, and cost of each step, the following optimization suggestions can be made:

[0149] Index optimization: Check whether the execution plan uses indexes, especially on frequently queried columns. If indexes are not used, consider creating new indexes or optimizing existing indexes.

[0150] Join optimization: Analyze the join methods in the execution plan, especially the join operations on large tables. You can consider adjusting the join order and using appropriate join algorithms (such as Hash Join and Merge Join) to optimize performance.

[0151] Filter conditions: Analyze the filter conditions in the query to confirm whether the indexes and partition keys are used effectively. Optimizing the filter conditions can significantly reduce the amount of data scanned.

[0152] Parallelization settings: For MPP databases, you can improve query performance by adjusting the degree of parallelism and resource allocation. The execution plan may show parallel scan or parallel join operations, so make sure the configuration is reasonable.

[0153] Performance testing and monitoring

[0154] After implementing the optimization suggestions, it is recommended to conduct performance testing and monitoring to ensure the optimization effect. Use EXPLAIN to obtain the actual execution time to help evaluate the optimization effect.

[0155] Example optimization suggestion:

[0156] If the execution plan shows a full table scan (Seq Scan), consider whether you can add indexes or use more effective filtering conditions to reduce the amount of scanned data.

[0157] If a high-cost join operation (such as Nested Loop Join) is used, consider whether performance can be improved by adjusting the query order or using Hash Join instead.

[0158] In summary, parsing the MPP execution plan and giving optimization suggestions requires combining the specific execution plan and database architecture characteristics, deeply analyzing the query execution path and cost estimation, and thus proposing effective performance optimization strategies.

[0159] Create a database execution plan optimization suggestion table;

[0160] The characteristics and runtime consumption of historical optimization cases are recorded in the data table, and the optimization details are recorded as manual optimization suggestions. Manual optimization suggestions are used to assist AI in SQL script optimization.

[0161] Provide suggestions for optimizing database operations based on the prediction results and model analysis. - This includes creating or modifying indexes, adjusting query statement structures, partitioning tables, etc. For example, in the database of a financial company, complex transaction queries are often executed. By analyzing the database execution plan through AI and historical optimization specific features, it can be found that some queries take too long to execute due to insufficient use of indexes. The AI ​​model predicts the performance of new queries based on historical data and recommends creating appropriate indexes to improve query efficiency. For another example, when an e-commerce platform's database is processing a large amount of order data, AI analysis finds that some data read operations use full table scans, resulting in performance degradation, and recommends partitioning related tables to improve performance. In short, AI analysis of database execution plans helps improve database performance and efficiency, optimize resource utilization, and provide better support for the business operations of enterprises.

[0162] Step 4. Develop the access AI interface, write the prompt word Prompt to create roles and requirements, and use it to generate indicator analysis and indicator development scripts;

[0163] Here are some tips for creating SQL developers and data analysts:

[0164] SQL Developer Prompt:

[0165] An experienced and skilled SQL developer, proficient in PostGreSQL-based MPP database management system. Able to efficiently write complex SQL queries, stored procedures and triggers, with good database design and optimization capabilities. When solving problems, pay attention to code readability, maintainability and performance optimization. Able to quickly propose innovative and effective solutions to new requirements and technical challenges.

[0166] Prompt for data analysts:

[0167] A keen and professional data analyst, proficient in data analysis tools and techniques, such as Excel, Python (including libraries such as Pandas and NumPy), SQL, and data visualization tools (such as Tableau, PowerBI, etc.). Able to extract valuable information from large amounts of data and conduct in-depth data analysis and insights. Good at discovering problems and proposing hypotheses through data, and verifying and solving problems through rigorous analysis. Able to clearly and accurately explain the results and conclusions of data analysis to non-technical personnel, providing strong support for decision-making.

[0168] Later, you will need to add and modify the features of Prompt according to the characteristics of different roles.

[0169] Step 5. Conduct visual development of risk control indicator analysis, and perform front-end development of indicator analysis and optimization functions.

[0170] Developing the back-end and front-end of risk control indicator analysis visualization involves different technologies and tools. The following discusses the key points and suggestions of each.

[0171] Backend Development:

[0172] Back-end development is mainly responsible for processing data and business logic, providing API interfaces for front-end calls, and ensuring data security and performance.

[0173] Framework and language: Choose a framework and language suitable for big data processing and high concurrency, such as Java's Spring framework.

[0174] Database: Mysql.

[0175] API design: Design RESTful APIs to provide to the front end for data acquisition and operation. API design should comply with standards, including good URL naming conventions, appropriate HTTP methods (GET, POST, PUT, DELETE, etc.) and error handling mechanisms.

[0176] Data processing: Implement data processing, analysis, and calculation to generate risk control indicator data. Consider using task queues (such as Celery) to handle background tasks and asynchronous processing.

[0177] Security: Ensure the security of the API interface, including authentication and authorization mechanisms, data transmission encryption (HTTPS), input validation, and prevention of SQL injection.

[0178] Performance optimization: Optimize database query and data processing algorithms, as well as the design of cache strategies, to improve the system's response speed and concurrent processing capabilities.

[0179] Front-end development

[0180] Front-end development is responsible for the design and interaction of the user interface, as well as the connection with the back-end API and data display.

[0181] Technology selection and recommendations:

[0182] Framework: Choose a modern JavaScript framework such as React.js, Vue.js, or Angular to improve development efficiency and user experience.

[0183] Data visualization library: Select a data visualization library suitable for data display and interaction, such as D3.js, Chart.js, ECharts, etc., to draw charts and display risk control indicator data.

[0184] UI design: Design a user-friendly interface, considering elements such as data tables, charts, filters, searches, etc., so that users can view and analyze data intuitively.

[0185] API call: interact with the backend API to obtain risk control indicator data and display it. Use asynchronous requests and state management (such as Redux, Vuex) to manage data flow.

[0186] Responsive design: Ensure that the interface can be well displayed and interacted on different devices (such as desktops, tablets, and mobile phones) to achieve responsive design.

[0187] Performance optimization: optimize the loading speed and interactive performance of the front-end pages, reduce the number and size of resource requests, and adopt code splitting and lazy loading strategies.

[0188] Testing and deployment: After development is completed, comprehensive unit testing, integration testing, and performance testing are carried out to ensure system stability and performance.

[0189] In addition, an embodiment of the present invention provides an AI-based risk control data mart indicator analysis system, including:

[0190] The data acquisition module is used to acquire risk control indicators and develop a data dictionary by using the risk control data mart metadata management platform;

[0191] The first data processing module is used to develop backend connections and collect data dictionary information in the MPP database, and collect table and field distribution information through the MPP data dictionary table;

[0192] The second data processing module is used to use Java to develop the backend to connect to the MPP database and obtain the database execution plan for SQL script optimization;

[0193] The third data processing module is used to develop access AI interfaces, write prompt words to create roles and requirements, and generate indicator analysis and indicator development scripts;

[0194] The indicator data development module is used for visual development of risk control indicator analysis, and front-end development of indicator analysis and optimization functions.

[0195] It can be understood that the AI-based risk control data mart indicator analysis system provided by the present invention corresponds to the AI-based risk control data mart indicator analysis method provided in the aforementioned embodiments. The relevant technical features of an AI-based risk control data mart indicator analysis system can refer to the relevant technical features of an AI-based risk control data mart indicator analysis method, which will not be repeated here.

[0196] The present invention provides a physical structure of an electronic device, which may include: a processor, a communication interface, a memory, and a communication bus, wherein the processor, the communication interface, and the memory communicate with each other through the communication bus. The processor may call the logic instructions in the memory to execute the above-mentioned AI-based risk control data mart indicator analysis method.

[0197] In addition, the logic instructions in the above-mentioned memory can be implemented in the form of software functional units and can be stored in a computer-readable storage medium when sold or used as an independent product. Based on such an understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product, which is stored in a storage medium and includes several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc., which can store program code.

[0198] Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0199] Although the preferred embodiments of the present invention have been described, those skilled in the art may make other changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications that fall within the scope of the present invention.

[0200] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, 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.

Claims

1. An AI-based risk control data mart indicator analysis method, characterized in that: The following steps are involved: Step 1. Obtain risk control indicators and develop a data dictionary by using the risk control data mart metadata management platform; Step 2. Develop the backend to connect and collect data dictionary information in the MPP database, and collect table and field distribution information through the MPP data dictionary table; Step 3. Use Java to develop the backend to connect to the MPP database and obtain the database execution plan for SQL script optimization; Step 4. Develop access AI interfaces, write prompt words to create roles and requirements, and use them to generate indicator analysis and indicator development scripts; Step 5. Conduct visual development of risk control indicator analysis, and perform front-end development of indicator analysis and optimization functions.

2. The AI-based risk control data mart indicator analysis method according to claim 1 is characterized in that: The risk control indicator development data dictionary includes: table structure, table business logic, field type, field business logic and table association relationship.

3. The AI-based risk control data mart indicator analysis method according to claim 2 is characterized in that: The risk control indicators are obtained by applying the risk control data mart metadata management platform to develop a data dictionary The following steps are involved: 1.

1. Access the metadata management platform to analyze the data structure, data volume, and data distribution in the database; 1.

2. Find risk control data marts, locate and select data marts or data warehouses containing risk control data; 1.

3. Find the table structure and business logic to find the data table containing risk control indicators; 1.

4. Understand the field type and business logic. For each field, obtain the field type and field business logic; 1.

5. Determine the relationship between tables and describe the relationship between different tables; 1.

6. Document the acquired data dictionary information and ensure regular updates; 1.

7. Ensure the data dictionary is visible and accessible to the entire team.

4. The AI-based risk control data mart indicator analysis method according to claim 1 is characterized in that: The table and field distribution includes partitions, data volume, and maximum and minimum value data.

5. The AI-based risk control data mart indicator analysis method according to claim 1 is characterized in that: The table and field distribution information collected through the MPP data dictionary table includes: 2.

1. Log in and connect to the MPP database, make sure you have permission to access the MPP database and have the credentials to connect to the database; 2.

2. Query system tables or metadata tables; 2.

3. Use SQL query statements to retrieve information about tables from system tables or metadata tables; 2.

4. Get the statistical information of the fields from the system table; 2.

5. Query based on foreign key constraints or index information in the system table; 2.

6. Export the data dictionary information obtained from the MPP database as a document or spreadsheet.

6. According to the AI-based risk control data mart indicator analysis method of claim 1, it is characterized in that the use of Java to develop a backend to connect to the MPP database and obtain a database execution plan includes the following: 3.

1. Connect to the database, execute SQL statements through client tools, and obtain the execution plan; 3.

2. Analyze the execution plan; 3.

3. Analyze and optimize queries, identify bottlenecks and optimization opportunities in queries.

7. The AI-based risk control data mart indicator analysis method according to claim 6, wherein the analysis execution plan comprises: Extract key features from the execution plan, use the key features as optimization prompts for the AI ​​model, and use the execution plan optimization suggestion table as an alternative to assist in SQL script optimization; the key features include execution time, resource consumption, and data access method.

8. The AI-based risk control data mart indicator analysis method according to claim 1, characterized in that: The writing prompt word creation roles and requirements include: Create prompt words for SQL developers and data analysts, and add and modify prompt words based on the characteristics of different roles.

9. An AI-based risk control data mart indicator analysis system, characterized in that: include: The data acquisition module is used to acquire risk control indicators and develop a data dictionary by using the risk control data mart metadata management platform; The first data processing module is used to develop backend connections and collect data dictionary information in the MPP database, and collect table and field distribution information through the MPP data dictionary table; The second data processing module is used to use Java to develop the backend to connect to the MPP database and obtain the database execution plan for SQL script optimization; The third data processing module is used to develop access AI interfaces, write prompt words to create roles and requirements, and generate indicator analysis and indicator development scripts; The indicator data development module is used for visual development of risk control indicator analysis, and front-end development of indicator analysis and optimization functions.

10. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the program, an AI-based risk control data mart indicator analysis method as described in any one of claims 1 to 8 is implemented.