Artificial intelligence modeling method and device based on report system, equipment and medium

Abstract

The invention relates to an artificial intelligence modeling method and device based on a report system, equipment and a medium. The method comprises the steps that a data set, a control and a page in the report system are converted into a target formatted language description document; obtaining a data dictionary, performing data preprocessing based on the data dictionary and the target formatted language description document, and generating metadata information catalog of all current data; obtaining a query request input by a target user, and identifying a query keyword in the query request; obtaining a target control and target data corresponding to the query keyword based on the metadata information catalog; constructing a user portrait of the target user, and determining recommendation information of the target user according to the user portrait; and generating target report information according to the target control, the target data and the recommendation information, and displaying the target report information. According to the invention, artificial intelligence modeling in the report system is realized, and the accuracy of generation of the control and recommendation data information is improved.

Description

Technical Field

[0001] The present application relates to the field of artificial intelligence technology, and in particular to an artificial intelligence modeling method, device, equipment and medium based on a reporting system. Background Art

[0002] In the reporting system (including embedded reports in the business system, independent BI reports, etc.), the reports prepared by the implementation personnel are usually difficult to meet the needs of all inquirers. Especially when performing overview queries or statistical queries, managers often put forward various unique reporting requirements. When the AI ​​big model appeared, someone proposed to introduce the AI ​​big model into the reporting system, and the user directly described the requirements to the AI. The AI ​​is responsible for converting the requirements into query instructions, and after querying the data, the results are sorted, analyzed, and presented to the user. With the improvement and popularization of the capabilities of the AI ​​big model, there are also many solutions built using this type of model, including open source solutions provided by large companies. However, due to various practical limitations, the application effect of the products made based on this solution is unsatisfactory, and there are few customer cases of successful actual deployment.

[0003] The existing AI big models need to understand the specific data meaning of the customer before they can provide query functions. In terms of databases and data warehouses, customers can usually get the corresponding data dictionary provided by the application provider or the data warehouse builder. The data dictionary can provide the definition and function of each table and each field, but lacks guidance for specific business scenarios. In addition, the AI ​​big model lacks a global understanding of complete application functions or enterprise data, which is specifically manifested in: a. Lack of logical data relationships between businesses, such as lack of knowledge of the relationship between order data and inventory data, so it is impossible to provide relevant related information query and analysis; b. Lack of user scenario awareness, for example, when a financial staff member starts to search, the AI ​​big model does not know which information the financial staff is concerned about, so it is impossible to provide data in a targeted manner. Therefore, it is difficult for the AI ​​big model in the existing reporting system to improve accurate data. Summary of the invention

[0004] The purpose of the embodiments of the present application is to propose an artificial intelligence modeling method, device, equipment and medium based on a reporting system to realize artificial intelligence modeling within the reporting system and improve the accuracy of control and recommended data information generation.

[0005] In order to solve the above technical problems, the embodiment of the present application provides an artificial intelligence modeling method based on a reporting system, including: Convert data sets, controls and pages in the report system into target formatting language description documents; Acquire a data dictionary, perform data preprocessing based on the data dictionary and the target formatting language description document, and generate a metadata information catalog of all current data; Obtaining a query request input by a target user, and identifying a query keyword in the query request; Acquire a target control corresponding to the query keyword based on the metadata information catalog; Constructing a user profile of the target user, and determining recommended information for the target user based on the user profile; Target report information is generated according to the target control and the recommendation information, and the target report information is displayed.

[0006] In order to solve the above technical problems, the embodiment of the present application provides an artificial intelligence modeling device based on a reporting system, including: Document conversion module, used to convert data sets, controls and pages in the report system into target formatting language description documents; A preprocessing module, used to obtain a data dictionary, perform data preprocessing based on the data dictionary and the target formatting language description document, and generate a metadata information catalog of all current data; A request acquisition module, used to acquire a query request input by a target user and identify query keywords in the query request; A control acquisition module, used for acquiring a target control corresponding to the query keyword based on the metadata information catalog; A user portrait building module, used to build a user portrait of the target user and determine the recommended information of the target user according to the user portrait; An information display module is used to generate target report information according to the target control and the recommendation information, and display the target report information.

[0007] In order to solve the above technical problems, a technical solution adopted by the present invention is: to provide a computer device, including one or more processors; a memory for storing one or more programs, so that one or more processors can implement any one of the above-mentioned artificial intelligence modeling methods based on the reporting system.

[0008] In order to solve the above technical problems, a technical solution adopted by the present invention is: a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, it implements any one of the above-mentioned artificial intelligence modeling methods based on the reporting system.

[0009] The embodiment of the present invention provides an artificial intelligence modeling method, device, equipment and medium based on a reporting system. The method includes: converting data sets, controls and pages in the reporting system into target formatting language description documents; obtaining a data dictionary, performing data preprocessing based on the data dictionary and the target formatting language description document, and generating a metadata information catalog of all current data; obtaining a query request input by a target user, and identifying the query keyword in the query request; obtaining a target control corresponding to the query keyword based on the metadata information catalog; constructing a user portrait of the target user, and determining the recommended information of the target user according to the user portrait; generating target report information according to the target control and the recommended information, and displaying the target report information. In the reporting system of the embodiment of the present invention, data sets, controls and pages are converted into target formatting language description documents, and a metadata information catalog of all current data is generated. At the same time, when a user performs a query, the corresponding control can be accurately queried according to the metadata information catalog, and the recommended information can be accurately identified according to the user portrait, thereby realizing artificial intelligence modeling in the reporting system, which is conducive to improving the accuracy of generating controls and recommended data information. BRIEF DESCRIPTION OF THE DRAWINGS

[0010] In order to more clearly illustrate the scheme in the present application, a brief introduction is given below to the drawings required for use in the description of the embodiments of the present application. Obviously, the drawings described below are some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0011] Figure 1 It is a flowchart of the implementation process of the artificial intelligence modeling method based on the reporting system provided in the embodiment of the present application; Figure 2 It is a flowchart for implementing the first sub-process in the artificial intelligence modeling method based on the reporting system provided in the embodiment of the present application; Figure 3 It is a flowchart for implementing the second sub-process in the artificial intelligence modeling method based on the reporting system provided in the embodiment of the present application; Figure 4 It is a flowchart for implementing the third sub-process in the artificial intelligence modeling method based on the reporting system provided in the embodiment of the present application; Figure 5 It is a flowchart for implementing the fourth sub-process in the artificial intelligence modeling method based on the reporting system provided in the embodiment of the present application; Figure 6 It is a flowchart for implementing the fifth sub-process in the artificial intelligence modeling method based on the reporting system provided in the embodiment of the present application; Figure 7It is a flowchart for implementing the sixth sub-process in the artificial intelligence modeling method based on the reporting system provided in the embodiment of the present application; Figure 8 It is a schematic diagram of an artificial intelligence modeling device based on a reporting system provided in an embodiment of the present application; Fig. 9 It is a schematic diagram of a computer device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0012] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by technicians in the technical field of the present application; the terms used in the specification of the application herein are only for the purpose of describing specific embodiments and are not intended to limit the present application; the terms "including" and "having" and any variations thereof in the specification and claims of the present application and the above-mentioned drawings are intended to cover non-exclusive inclusions. The terms "first", "second", etc. in the specification and claims of the present application or the above-mentioned drawings are used to distinguish different objects, not to describe a specific order.

[0013] Reference to "embodiments" herein means that a particular feature, structure, or characteristic described in conjunction with the embodiments may be included in at least one embodiment of the present application. The appearance of the phrase in various locations in the specification does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment that is mutually exclusive with other embodiments. It is explicitly and implicitly understood by those skilled in the art that the embodiments described herein may be combined with other embodiments.

[0014] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the accompanying drawings.

[0015] The present invention is described in detail below with reference to the accompanying drawings and embodiments.

[0016] It should be noted that the artificial intelligence modeling method based on the reporting system provided in the embodiment of the present application is generally executed by a server, and accordingly, the artificial intelligence modeling device based on the reporting system is generally configured in the server.

[0017] Specifically, most of the current AI (artificial intelligence) solutions are independently built and can be connected to any software system. The AI ​​solution of the embodiment of the present application is only an extended function of the reporting system and is integrated with the reporting system. If other applications (applications that require an embedded reporting system) need to be used, the entire reporting system should be used; separately deployed reporting systems (such as smart cockpits that can be applied to enterprise digital transformation) can be used directly. This application is an independently deployed AI system, and the knowledge base it constructs requires the user to provide a data dictionary for the database to be connected. The embodiment of the present application can realize artificial intelligence modeling within the reporting system, which is conducive to improving the accuracy of the generation of controls and recommended data information.

[0018] See also Figure 1 , Figure 1 A specific implementation of an artificial intelligence modeling method based on a reporting system is shown.

[0019] It should be noted that if there are substantially the same results, the method of the present invention is not limited to Figure 1 The method shown is limited to the process sequence, and the method comprises the following steps: S1: Convert the data sets, controls and pages in the report system into target formatting language description documents.

[0020] Specifically, the reporting system will first save the query as a data set, which includes a name, a database link, and a query instruction. Each displayed data element in the reporting system is a control, which includes a name, a style (for example, a pie chart that needs to display specific values ​​and proportions; for example, a form, which field is used for sorting, etc.) and an associated data set. The smallest unit of a single presentation is a page, which includes a name and multiple controls. In an embodiment of the present application, the data sets, controls, and pages in the reporting system are converted into a target formatting language description document. The data sets, controls, and pages in the reporting system are all equipped with their own permission controls. These permission control information will be associated with the data sets, controls, and pages.

[0021] See also Figure 2 , Figure 2 A specific implementation of step S1 is shown, which is described in detail as follows: S11: Generate a first formatted language description document from the non-embedded query data set in the report system based on the query instruction.

[0022] Among them, the first formatting language description document includes data basic information, data result information, query information, permission control information, business rule information and metadata version control information. Basic information includes data set name and unique identifier, associated database link information, creation time and last modification time, creator and maintainer information. Data structure information includes the definitions of involved tables and fields, the association relationship between fields (foreign keys, references, etc.), the business meaning and constraints of fields, and the calculation logic of calculated fields. Query-related information includes original query instructions, query conditions and filtering rules, sorting and grouping rules, aggregate functions and calculation formulas, and performance optimization tips (such as index usage recommendations). Permission control includes: field-level access rights, data row-level filtering rules, permission rules for special roles, and sensitive data processing strategies. Business rules include: business logic relationships between fields, data quality rules, business scenario tags, and commonly used analysis dimensions. Metadata version control includes: version number and change records, dependency tracking, deprecated field tags, and compatibility instructions.

[0023] In the embodiment of the present application, for a data set, since its main content is a query instruction, firstly, an AI model is used to analyze the query instruction and generate a document described in a formatted language (such as json). Since the query language is not a natural language, a specially tuned AI model or a general AI model supplemented with specific prompt words is used to decompose the query intent of the query language (including but not limited to SQL) to generate a formatted language. This avoids the differences, complexity, and irrelevance of the query language. Differences: Relational databases use SQL, but data warehouses, object-oriented databases, streaming databases, big data processing services, third-party APIs, etc. usually have their own query language standards. It needs to be organized into a unified format for subsequent processing. Complexity: Due to differences in databases, the same business data relationship may have different data affiliations in different data sources. The present invention only needs to focus on the results presented by its data set (flat data table), and it is necessary to avoid the hierarchy of the data source interfering with the subsequent model processing. Irrelevant: In addition to fields, the query usually also contains conditions, joint queries, embedded subqueries, function calls and other functions. Since the present invention uses these data sets as raw data instead of directly operating the database, AI does not need to understand these additional functions but only needs to focus on the fields.

[0024] Since the system of the present invention does not generate query commands and execute them in the data source to obtain the original data, but reuses and integrates the existing data sources, the system does not need to understand the calculations, built-in fixed conditions, fixed groupings and summary formulas when obtaining the data set, but needs to try to understand the business relationship between the data sets; S12: recursively analyzing the embedded query data set in the report system to generate multi-level relationship description information, and storing the relationship description information in a vector database.

[0025] Specifically, for complex data sets with embedded queries, the system will recursively analyze and generate multi-level relationship descriptions for subsequent reuse in different query levels (such as explicit fact queries and implicit fact queries). This information will be stored in a vector database to support semantic retrieval and similarity matching.

[0026] In a specific embodiment, for a data set with embedded queries (such as embedded queries with SQL conditions), it is necessary to perform an independent analysis on the relationship between each embedded subquery. If there are still subqueries in the embedded query, repeat the recursive execution. The process of this analysis is similar to the outer layer analysis, analyzing the tables and fields used in the query. The commands of the subqueries may be highly similar to the queries of some independent databases. This information can determine the relationship between different data sources and be used in combination in subsequent queries; S13: converting the controls and pages in the report system into a second formatting language description document and a third formatting language description document respectively; S14: generating the target formatted language description document based on the first formatted language description document, the second formatted language description document and the third formatted language description document;

[0027] Specifically, the controls in the report system are converted into a second formatting language description document through program code. The second formatting language description document will describe the name of the control, the name of the related data set, the display style, filtering conditions, sorting and grouping, access rights, etc. The pages in the report system are converted into a third formatting language description document through program code. The third formatting language description document will describe the name of the page, the names of related controls, the hierarchical relationship (such as the grouping of controls under the page), the placement, access rights, etc. S2: Acquire a data dictionary, perform data preprocessing based on the data dictionary and the target formatting language description document, and generate a metadata information catalog of all current data.

[0028] Specifically, the data dictionary provided by the original data provider (eg, system developer) is obtained, data preprocessing is performed based on the data dictionary and the target formatting language description document, and the data is integrated into a unified metadata information catalog for all current data.

[0029] See also Figure 3 , Figure 3 A specific implementation of step S2 is shown, which is described in detail as follows: S21: Obtain the data dictionary provided by the original data provider; S22: performing data cleaning and standardization processing on the data dictionary and the target formatting language description document to generate target data information.

[0030] See also Figure 4 , Figure 4 A specific implementation of step S22 is shown, which is described in detail as follows: S221: unifying the field naming and data types of the data dictionary and the target formatting language description document to obtain basic data information; S222: removing abnormal values ​​and missing values ​​in the basic data information, and processing the business terms and calculation formulas in the basic data information to obtain pre-processed data information; S223: Constructing a field mapping relationship based on the preprocessed data information to generate the target data information.

[0031] Specifically, in order to standardize data and reduce data redundancy and data errors, it is necessary to unify the field naming and data types of the data dictionary and the target formatting language description document to obtain basic data information, then remove abnormal values ​​and missing values ​​in the basic data information, and process the business terms and calculation formulas in the basic data information to obtain pre-processed data information, and then build a field mapping relationship based on the pre-processed data information to generate target data information; S23: Use the RAGFlow engine to build a knowledge graph based on the target data information.

[0032] The embodiment of the present application uses the RAGFlow engine to build a knowledge graph based on target data information, wherein RAGFlow is an open source RAG (Retrieval-Augmented Generation) engine built based on deep document understanding. In the embodiment of the present application, intelligent entity recognition, deep relationship mining, multi-hop reasoning architecture, spatiotemporal causal modeling and LLM collaborative construction can be realized. Among them, intelligent entity recognition combines the spaCy NER model with the BLINK entity linking technology to achieve accurate identification and disambiguation of business entities. Deep relationship mining uses remote supervision plus pre-training models to extract explicit / implicit business relationships (such as order-inventory associations). The multi-hop reasoning architecture is based on the graph neural network to achieve reasoning capabilities across 3 levels of nodes (such as: order → product → supplier). Spatiotemporal causal modeling is to encode time series and causal relationships in business rules (such as the impact of sales fluctuations on inventory). LLM collaborative construction is to establish a feedback optimization closed loop by assisting relationship discovery and pattern recognition through a large model. Since the embodiment of the present application performs relationship mining, LLM collaborative relationship verification can be used, that is, the relationship mined by the algorithm is handed over to LLM for business rationality verification (such as the causal relationship of "customer satisfaction → logistics timeliness"). At the same time, LLM can also judge the rationality after the multi-hop reasoning is calculated, and extend new multi-hop reasoning assumptions, such as: path assumptions (such as the "inventory backlog→capital turnover→financial expenses" link).

[0033] The knowledge graph in the embodiment of the present application includes a node structure and a relationship network. Among them, the node structure entity nodes, relationship edges and community divisions. Entity nodes include name, description, type (such as organization / personnel / location). Relationship edges are explicit / implicit associations between entities. Community divisions are entity communities that are automatically clustered by graph algorithms. The entities therein are determined based on actual business scenarios. In one example, the business entity layer includes core business objects (orders / inventory / products), data elements (fields / tables / queries), and system components (datasets / controls / pages). The relationship network includes data logical relationships (order→inventory impact relationship), business rule relationships (calculation formula dependency chain), and permission inheritance relationships (page→control→field permission transfer); S24: Perform vectorization processing based on the topological structure of the knowledge graph to generate index information.

[0034] See also Figure 5 , Figure 5 A specific implementation of step S24 is shown, which is described in detail as follows: S241: Encoding the topological structure of the knowledge graph into a vector space to vectorize the result information of the knowledge graph and generate the multi-dimensional vector index; S242: Use a graph attention network to construct vector representation information of the graph context based on the multi-dimensional vector index; S243: Construct the progressive hierarchical index based on the vector representation information.

[0035] Specifically, the topological structure of the knowledge graph is encoded into the vector space to vectorize the result information of the knowledge graph, generate a multi-dimensional vector index, and then use the graph attention network to construct the vector representation information of the graph context based on the multi-dimensional vector index, and then construct a progressive hierarchical index based on the vector representation information. The hierarchical index is a progressive index of the concept layer → entity layer → attribute layer. In the embodiment of the present application, the topological structure of the knowledge graph can also be used to optimize the vector retrieval performance and implement the incremental update mechanism; S25: Constructing a query module library and cache access information based on the target data information according to a preset cache strategy.

[0036] Specifically, commonly used query results are cached according to the preset cache strategy, and typical analysis scenarios are pre-generated, query template libraries are built, and hot data access is optimized. Pre-generated typical analysis scenarios can be sales weekly reports, inventory warnings, etc. By pre-generating analysis scenarios, users can directly call them when querying, reducing the time for real-time processing. Manual analysis of historical query data or industry general templates determines the analysis scenario.

[0037] Establishing a query template library is to abstract common query patterns into reusable templates, such as filtering by time range, summarizing by department, etc. The establishment of a template library requires extracting patterns from existing reports and user queries, and LLM natural language processing or pattern recognition technology is needed to classify similar query requests. Optimize hot data access. Hot data refers to data that is frequently accessed. Optimization strategies include caching strategies, index optimization, or preloading mechanisms. Determine which data is hot data based on the monitored access frequency, and use caching technology, such as Redis, or adjust database indexes to speed up querying data; S26: Construct the metadata information catalog of all current data according to the target data information, the knowledge graph, the index information, the query template library and the cache access information.

[0038] Specifically, the preprocessed data (including target data information, knowledge graph, index information, query template library and cache access information, etc.) are integrated into a unified metadata information catalog.

[0039] Furthermore, in another specific embodiment, for small-scale models running locally, the system will perform model tuning based on preprocessed data, including tasks such as intent recognition, entity extraction, and relationship reasoning. For large-scale deployment scenarios, the preprocessed data will be used as the RAG knowledge base to support query requirements at different levels. Among them, query requirements include: explicit fact queries, implicit fact queries, explainable reasoning, and implicit reasoning queries. Explicit fact queries are to retrieve relevant information directly from vector indexes. Implicit fact queries are multi-hop reasoning through knowledge graphs. Explainable reasoning is logical deduction combined with business rules. Implicit reasoning queries are to use historical data for pattern recognition; S3: Obtain a query request input by a target user, and identify query keywords in the query request.

[0040] Specifically, the above steps have preprocessed the data in the report system and generated metadata information catalog and knowledge base information. Therefore, after the target user re-enters the query request in the report system, the query keywords in the query request need to be identified. The query keywords in the query request can be identified through the natural language model; S4: Obtain a target control corresponding to the query keyword based on the metadata information catalog.

[0041] Specifically, in the metadata information catalog, the corresponding controls or control sets can be matched according to the query keywords, and then the target controls can be determined from these controls or control sets.

[0042] See also Figure 6 , Figure 6 A specific implementation of step S4 is shown, which is described in detail as follows: S41: Matching a control or a control set corresponding to the query keyword from the metadata information catalog; S42: performing data query according to the style of the control or the control set and the associated data set to determine the target control, and sorting the target controls according to the grouping; S43: If the control or the control set is not matched, generate supplementary recording prompt information, and feed back to the target user based on the supplementary recording prompt information.

[0043] Among them, the control or control set corresponding to the query keyword is matched from the metadata information catalog. Then, a data query is performed according to the style and associated data set of the control or the control set to determine the target control, and the target controls are sorted according to the group. The control grouping can be grouped according to the name in the page corresponding to the control. The control sorting can be sorted according to the generation time or according to the similarity. If the control or the control set is not matched, a supplementary recording prompt information is generated, and based on the supplementary recording prompt information, it is fed back to the target user to remind the target user to supplement the control; S5: Construct a user profile of the target user, and determine recommendation information for the target user based on the user profile.

[0044] See also Figure 7 , Figure 7 A specific implementation of step S5 is shown, which is described in detail as follows: S51: Acquire historical data of the target user, and identify the user type, preference tag, display preference information, business field, and authority information of the target user based on the historical data; S52: constructing a user profile of the target user according to the user type, the preference tag, the display preference information, the business field, and the authority information; S53: Determine the recommended information of the target user according to the user portrait.

[0045] Specifically, the user portrait of each user can be pre-built. When the user makes a query, the user portrait corresponding to the queried user can be determined, and the user's recommended information can be determined based on the user portrait, so as to generate recommended information of interest to the user and improve the accuracy of data recommendation. In an embodiment of the present application, the historical data of the target user is obtained, and the user type, preference label, display preference information, business field and permission information of the target user are identified based on the historical data. The user portrait of the target user is constructed based on the user type, preference label, display preference information, business field and permission information, and finally the recommended information of the target user is determined based on the user portrait.

[0046] Among them, user type is to automatically identify user type (such as financial personnel, sales manager) by analyzing the query keywords commonly used by users. Preference tags are to record the types of data that users often view (such as sales, inventory) to form preference tags. Display preference information is to remember the chart styles that users have used recently (such as the pie chart / bar chart selected in the last 10 times). Business areas are to count the business areas that users query most frequently (such as frequent queries for order-related data). Permission information is to record the scope of data that users can access to ensure that they do not exceed their authority.

[0047] In a specific embodiment, a "standard toolkit" is automatically recommended based on the target user's position (e.g., the sales director sees the customer analysis dashboard by default). The route can be predicted based on the target user's navigation software, and the weekly sales report that the user must check every Monday for the past three months can be automatically prepared in advance; S6: Generate target report information according to the target control and the recommendation information, and display the target report information.

[0048] Specifically, in the report system, target report information is generated based on the target control and the recommendation information, and the target report information is displayed to the target user.

[0049] Furthermore, the embodiments of the present application can record the actual use of recommended content by users, automatically update user feature information at regular intervals, and regularly compare the effects of different recommendation strategies. At the same time, it is also possible to clean up display templates that have not been used for a long time, discover new business rules, and update the rules. In the reporting system, each time a query request is generated and the location of the original data is displayed, a button is added to allow the user to save this query as a data set and control for subsequent direct use in report design. For the same or highly similar content that is queried by multiple users and multiple times, AI will give a reminder and recommend saving it.

[0050] In an embodiment of the present application, the data sets, controls and pages in the reporting system are converted into target formatting language description documents; a data dictionary is obtained, data preprocessing is performed based on the data dictionary and the target formatting language description document, and a metadata information catalog of all current data is generated; a query request input by a target user is obtained, and query keywords in the query request are identified; a target control corresponding to the query keyword is obtained based on the metadata information catalog; a user portrait of the target user is constructed, and recommended information for the target user is determined based on the user portrait; target report information is generated based on the target control and the recommended information, and the target report information is displayed. In an embodiment of the present invention, the data sets, controls and pages in the reporting system are converted into target formatting language description documents, and a metadata information catalog of all current data is generated. At the same time, when a user performs a query, the corresponding controls and data can be accurately queried based on the metadata information catalog, and the recommended information can be accurately identified based on the user portrait, thereby realizing artificial intelligence modeling in the reporting system, which is conducive to improving the accuracy of control and recommended data information generation.

[0051] Please refer to Figure 8 , as a response to the above Figure 1 The present application provides an embodiment of an artificial intelligence modeling device based on a reporting system, and the embodiment of the device is Figure 1 Corresponding to the embodiment of the method shown, the device can be specifically applied to various electronic devices.

[0052] like Figure 8 As shown, the artificial intelligence modeling device based on the reporting system of this embodiment includes: a document conversion module 71, a preprocessing module 72, a request acquisition module 73, a control acquisition module 74, a user portrait construction module 75 and an information display module 76, wherein: The document conversion module 71 is used to convert the data sets, controls and pages in the report system into target formatting language description documents; A preprocessing module 72 is used to obtain a data dictionary, perform data preprocessing based on the data dictionary and the target formatting language description document, and generate a metadata information catalog of all current data; A request acquisition module 73, used to acquire a query request input by a target user and identify a query keyword in the query request; A control acquisition module 74, configured to acquire a target control corresponding to the query keyword based on the metadata information catalog; A user portrait building module 75, used to build a user portrait of the target user and determine the recommended information of the target user according to the user portrait; The information display module 76 is used to generate target report information according to the target control and the recommendation information, and display the target report information.

[0053] Furthermore, the document conversion module 71 includes: A first conversion unit is used to generate a first formatted language description document based on a query instruction from a non-embedded query data set in the report system, wherein the first formatted language description document includes data basic information, data result information, query information, authority control information, business rule information and metadata version control information; A recursive analysis unit, used for recursively analyzing the embedded query data set in the report system, generating multi-level relationship description information, and storing the relationship description information in a vector database; A second conversion unit, used to convert the controls and pages in the report system into a second formatting language description document and a third formatting language description document respectively; A description document generating unit is used to generate the target formatting language description document based on the first formatting language description document, the second formatting language description document and the third formatting language description document.

[0054] Furthermore, the preprocessing module 72 includes: A data dictionary acquisition unit, used to acquire the data dictionary provided by the original data provider; A target data information generating unit, used for performing data cleaning and standardization processing on the data dictionary and the target formatting language description document to generate target data information; A knowledge graph construction unit, used to construct a knowledge graph based on the target data information using a RAGFlow engine; An index information generating unit, used for performing vectorization processing based on the topological structure of the knowledge graph to generate index information; A query module library construction unit, used to construct a query module library and cache access information based on the target data information according to a preset cache strategy; A catalog generation unit is used to build the metadata information catalog of all current data based on the target data information, the knowledge graph, the index information, the query template library and the cache access information.

[0055] Furthermore, the target data information generating unit includes: A basic data information generating unit, used for unifying the field naming and data types of the data dictionary and the target formatting language description document to obtain basic data information; A preprocessed data information generating unit, used to remove abnormal values ​​and missing values ​​in the basic data information, and process the business terms and calculation formulas in the basic data information to obtain preprocessed data information; A mapping relationship building unit is used to build a field mapping relationship based on the preprocessed data information to generate the target data information.

[0056] Furthermore, the index information includes a multi-dimensional vector index and a hierarchical index, and the index information generating unit includes: A vectorization processing unit, used to encode the topological structure of the knowledge graph into a vector space, so as to perform vectorization processing on the result information of the knowledge graph and generate the multi-dimensional vector index; A vector representation information construction unit, configured to construct vector representation information of a graph context based on the multi-dimensional vector index using a graph attention network; A layered index building unit is used to build the layered index in a progressive manner based on the vector representation information.

[0057] Furthermore, the control acquisition module 74 includes: A module space construction unit, used to match the control or control set corresponding to the query keyword from the metadata information catalog; A control sorting unit, used for performing data query according to the style of the control or the control set and the associated data set to determine the target control, and sorting the target controls according to the grouping; A supplementary recording prompt information generating unit is used to generate supplementary recording prompt information if the control or the control set is not matched, and feed back to the target user based on the supplementary recording prompt information.

[0058] Furthermore, the user portrait building module 75 includes: A historical data acquisition unit, used to acquire the historical data of the target user, and identify the user type, preference label, display preference information, business field and authority information of the target user based on the historical data; A user portrait generating unit, configured to construct a user portrait of the target user according to the user type, the preference tag, the display preference information, the business domain and the authority information; A recommendation information generating unit is used to determine the recommendation information of the target user according to the user portrait.

[0059] To solve the above technical problems, the present application also provides a computer device. Fig. 9 , Fig. 9 This is a basic structural block diagram of the computer device in this embodiment.

[0060] The computer device 8 includes a memory 81, a processor 82, and a network interface 83 which are interconnected through a system bus. It should be noted that: Fig. 9 Only a computer device 8 having three components, memory 81, processor 82, and network interface 83, is shown, but it should be understood that it is not required to implement all the components shown, and more or fewer components may be implemented instead. Among them, those skilled in the art can understand that the computer device here is a device that can automatically perform numerical calculations and / or information processing according to pre-set or stored instructions, and its hardware includes but is not limited to microprocessors, application specific integrated circuits (ASIC), field-programmable gate arrays (FPGA), digital signal processors (DSP), embedded devices, etc.

[0061] Computer devices can be computing devices such as desktop computers, notebooks, PDAs, and cloud servers. Computer devices can interact with users through keyboards, mice, remote controls, touch pads, or voice control devices.

[0062] The memory 81 includes at least one type of readable storage medium, and the readable storage medium includes flash memory, hard disk, multimedia card, card-type memory (for example, SD or DX memory, etc.), random access memory (RAM), static random access memory (SRAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), programmable read-only memory (PROM), magnetic memory, disk, optical disk, etc. In some embodiments, the memory 81 can be an internal storage unit of the computer device 8, such as a hard disk or memory of the computer device 8. In other embodiments, the memory 81 can also be an external storage device of the computer device 8, such as a plug-in hard disk, a smart memory card (SmartMedia Card, SMC), a secure digital (Secure Digital, SD) card, a flash card (Flash Card), etc. equipped on the computer device 8. Of course, the memory 81 can also include both the internal storage unit of the computer device 8 and its external storage device. In this embodiment, the memory 81 is generally used to store the operating system and various application software installed on the computer device 8, such as the program code of the artificial intelligence modeling method based on the report system, etc. In addition, the memory 81 can also be used to temporarily store various types of data that have been output or are to be output.

[0063] In some embodiments, the processor 82 may be a central processing unit (CPU), a controller, a microcontroller, a microprocessor, or other data processing chips. The processor 82 is generally used to control the overall operation of the computer device 8. In this embodiment, the processor 82 is used to run the program code stored in the memory 81 or process data, such as running the program code of the artificial intelligence modeling method based on the reporting system to implement various embodiments of the artificial intelligence modeling method based on the reporting system.

[0064] The network interface 83 may include a wireless network interface or a wired network interface, and the network interface 83 is generally used to establish a communication connection between the computer device 8 and other electronic devices.

[0065] The present application also provides another implementation, namely, providing a computer-readable storage medium, which stores a computer program, and the computer program can be executed by at least one processor to enable the at least one processor to perform the steps of an artificial intelligence modeling method based on a reporting system as described above.

[0066] Through the description of the above implementation methods, those skilled in the art can clearly understand that the above-mentioned embodiment methods can be implemented by means of software plus a necessary general hardware platform, and of course by hardware, but in many cases the former is a better implementation method. Based on such an understanding, the technical solution of the present application, or the part that contributes to the prior art, can be embodied in the form of a software product, which is stored in a storage medium (such as ROM / RAM, a disk, or an optical disk), and includes a number of instructions for a terminal device (which can be a mobile phone, a computer, a server, an air conditioner, or a network device, etc.) to execute the methods of each embodiment of the present application.

[0067] Obviously, the embodiments described above are only some embodiments of the present application, rather than all embodiments. The preferred embodiments of the present application are given in the accompanying drawings, but they do not limit the scope of the present application. The present application can be implemented in many different forms. On the contrary, the purpose of providing these embodiments is to make the understanding of the disclosure of the present application more thorough and comprehensive. Although the present application is described in detail with reference to the aforementioned embodiments, for those skilled in the art, it is still possible to modify the technical solutions recorded in the aforementioned specific implementation methods, or to perform equivalent replacement of some of the technical features therein. Any equivalent structure made using the contents of the present application specification and the accompanying drawings, directly or indirectly used in other related technical fields, is similarly within the scope of protection of the present application.

Claims

1. An artificial intelligence modeling method based on a reporting system, characterized in that: include: Convert data sets, controls and pages in the report system into target formatting language description documents; Acquire a data dictionary, perform data preprocessing based on the data dictionary and the target formatting language description document, and generate a metadata information catalog of all current data; Obtaining a query request input by a target user, and identifying a query keyword in the query request; Acquire a target control corresponding to the query keyword based on the metadata information catalog; Constructing a user profile of the target user, and determining recommended information for the target user based on the user profile; Target report information is generated according to the target control and the recommendation information, and the target report information is displayed.

2. The artificial intelligence modeling method based on the reporting system according to claim 1 is characterized in that: The conversion of data sets, controls and pages in the report system into target formatting language description documents includes: Generate a first formatted language description document based on the query instruction from the non-embedded query data set in the report system, wherein the first formatted language description document includes data basic information, data result information, query information, authority control information, business rule information and metadata version control information; Recursively analyzing the embedded query data set in the report system to generate multi-level relationship description information, and storing the relationship description information in a vector database; Converting the controls and pages in the report system into a second formatting language description document and a third formatting language description document respectively; The target formatted language description document is generated based on the first formatted language description document, the second formatted language description document, and the third formatted language description document.

3. The artificial intelligence modeling method based on the reporting system according to claim 1 is characterized in that: The step of obtaining a data dictionary and performing data preprocessing based on the data dictionary and the target formatting language description document to generate a metadata information catalog of all current data includes: Obtaining the data dictionary provided by the original data provider; Performing data cleaning and standardization processing on the data dictionary and the target formatting language description document to generate target data information; Using the RAGFlow engine to build a knowledge graph based on the target data information; Performing vectorization processing based on the topological structure of the knowledge graph to generate index information; Constructing a query module library and cache access information based on the target data information according to a preset cache strategy; The metadata information catalog of all current data is constructed according to the target data information, the knowledge graph, the index information, the query template library and the cache access information.

4. The artificial intelligence modeling method based on the reporting system according to claim 3 is characterized in that: The data dictionary and the target formatting language description document are subjected to data cleaning and standardization processing to generate target data information, including: Unify the field naming and data types of the data dictionary and the target formatting language description document to obtain basic data information; Removing abnormal values ​​and missing values ​​from the basic data information, and processing the business terms and calculation formulas in the basic data information to obtain pre-processed data information; A field mapping relationship is constructed based on the preprocessed data information to generate the target data information.

5. The artificial intelligence modeling method based on the reporting system according to claim 3 is characterized in that: The index information includes a multi-dimensional vector index and a hierarchical index. The vectorization processing based on the topological structure of the knowledge graph is performed to generate the index information, including: Encoding the topological structure of the knowledge graph into a vector space to vectorize the result information of the knowledge graph and generate the multi-dimensional vector index; Using a graph attention network to construct vector representation information of the graph context based on the multi-dimensional vector index; The hierarchical index is constructed progressively based on the vector representation information.

6. The artificial intelligence modeling method based on the reporting system according to any one of claims 1 to 5, characterized in that: The step of obtaining a target control corresponding to the query keyword based on the metadata information catalog includes: Matching the control or control set corresponding to the query keyword from the metadata information catalog; Performing data query according to the style of the control or the control set and the associated data set to determine the target control, and sorting the target controls according to the grouping; If the control or the control set is not matched, supplementary recording prompt information is generated, and feedback is given to the target user based on the supplementary recording prompt information.

7. The artificial intelligence modeling method based on the reporting system according to any one of claims 1 to 5, characterized in that: The step of constructing a user profile of the target user and determining the recommended information of the target user according to the user profile includes: Acquire historical data of the target user, and identify the user type, preference tag, display preference information, business field, and authority information of the target user based on the historical data; Building a user profile of the target user according to the user type, the preference tag, the display preference information, the business field, and the permission information; Determine the recommended information for the target user based on the user portrait.

8. An artificial intelligence modeling device based on a reporting system, characterized in that: include: Document conversion module, used to convert data sets, controls and pages in the report system into target formatting language description documents; A preprocessing module, used to obtain a data dictionary, perform data preprocessing based on the data dictionary and the target formatting language description document, and generate a metadata information catalog of all current data; A request acquisition module, used to acquire a query request input by a target user and identify query keywords in the query request; A control acquisition module, used for acquiring a target control corresponding to the query keyword based on the metadata information catalog; A user portrait building module, used to build a user portrait of the target user and determine the recommended information of the target user according to the user portrait; An information display module is used to generate target report information according to the target control and the recommendation information, and display the target report information.

9. A computer device, characterized in that: It comprises a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, it implements the artificial intelligence modeling method based on the reporting system as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed by the processor, the artificial intelligence modeling method based on the reporting system as described in any one of claims 1 to 7 is implemented.

Images