Economic index visualization arrangement method based on large model multi-agent cooperation

Abstract

The application discloses an economic index visualization arrangement method based on large model multi-agent cooperation, relates to the technical field of multi-agent cooperation, and comprises the following steps: processing received input information through a first large language model, generating standardized requirements, and decomposing the standardized requirements into multiple subtasks; assigning the multiple subtasks to other agents in a preset agent role pool through a product manager agent, wherein the agent role pool comprises a data analysis and development agent and a visualization designer agent; determining a target data source through the data analysis and development agent and a first subtask assigned to the data analysis and development agent, and generating a data analysis result according to data in the target data source; and generating a visualization output corresponding to the data analysis result through a visualization designer agent and a second subtask assigned to the visualization designer agent. The application realizes a fully autonomous process from requirement understanding to final delivery, thereby improving the visualization efficiency of economic indexes.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of multi-agent collaboration, and particularly relates to an economic index visualization arrangement method based on large model multi-agent collaboration. BACKGROUND

[0002] With the development of artificial intelligence technology, AI (Artificial Intelligence) assisted development tools preliminarily realize automatic processing and visual presentation of structured data through natural language processing and machine learning technologies.

[0003] At present, for economic index analysis tasks, AI assisted tools still need manual input of detailed parameters, definition of rules and provision of specific algorithm models to analyze and predict economic indexes, and cannot independently perform data collection, analysis and visual result interpretation, that is, cannot realize a completely autonomous process from demand understanding to final delivery. In the entire economic index visualization process, manual intervention is required multiple times, resulting in low efficiency of visualization.

[0004] The above content is only used to assist in understanding the technical solutions of the present application and does not represent the acknowledgement of the above content as prior art. SUMMARY

[0005] The main purpose of the present application is to provide an economic index visualization arrangement method based on large model multi-agent collaboration, aiming to solve the technical problem of how to improve the efficiency of economic index visualization.

[0006] To achieve the above purpose, the present application provides an economic index visualization arrangement method based on large model multi-agent collaboration, which comprises:

[0007] The received input information is processed by a first large language model to generate standardized requirements, and the standardized requirements are decomposed into multiple subtasks, wherein the input information is the visualization requirement information of the economic index input by the user;

[0008] The multiple subtasks are assigned to other agents in a preset agent role pool through a product manager agent, wherein the agent role pool includes data analysis and development agents and visualization designer agents;

[0009] The target data source is determined through the data analysis and development agent and the first subtask assigned thereto, and the data analysis result is generated according to the data in the target data source;

[0010] The visualization output corresponding to the data analysis result is generated through the visualization designer agent and the second subtask assigned thereto.

[0011] In an embodiment, the step of determining, by the data analysis and development intelligent agent and the first subtask assigned to the data analysis and development intelligent agent, a target data source comprises:

[0012] determining, by the data analysis and development intelligent agent, a target economic entity according to the first subtask;

[0013] determining, by the data analysis and development intelligent agent, a target matching degree between the target economic entity and each data source in a preset data source library;

[0014] determining, by the data analysis and development intelligent agent and each target matching degree, a target data source.

[0015] In an embodiment, the step of determining, by the data analysis and development intelligent agent, a target matching degree between the target economic entity and each data source in a preset data source library comprises:

[0016] determining, by the data analysis and development intelligent agent, a standardized name of the target economic entity according to a preset economic knowledge graph;

[0017] determining, by the data analysis and development intelligent agent, a keyword matching degree between the standardized name and each index item of the data source;

[0018] determining, by the data analysis and development intelligent agent, a semantic matching degree between the input information and each index item of the data source;

[0019] determining, by the data analysis and development intelligent agent, the target matching degree according to the keyword matching degree and the semantic matching degree.

[0020] In an embodiment, the step of generating a data analysis result according to data in the target data source comprises:

[0021] pulling, by the data analysis and development intelligent agent, target data from the target data source;

[0022] performing, by the data analysis and development intelligent agent, data cleaning on the target data to obtain cleaned target data;

[0023] performing, by the data analysis and development intelligent agent, retrieval on the cleaned target data according to the first subtask to generate a data analysis result.

[0024] In an embodiment, the step of performing, by the data analysis and development intelligent agent, retrieval on the cleaned target data according to the first subtask to generate a data analysis result comprises:

[0025] The data analysis and development agent determines the target economic entity based on the first sub-task;

[0026] Based on a pre-defined economic knowledge graph, the extended information of the target economic entity is determined;

[0027] Based on the extended information, keyword retrieval is performed on the cleaned target data to obtain the first retrieval result;

[0028] The input information and the extended information are converted into text vectors using a fine-tuned large language model, wherein the fine-tuned large language model is obtained by fine-tuning training data in the economic field;

[0029] The second retrieval result is determined based on the vector similarity between the text vector and each indicator item in the cleaned target data;

[0030] Based on the first search result and the second search result, the target search result is determined, and based on the target search result, data analysis results are generated.

[0031] In one embodiment, the agent role pool includes: a data architect agent, and after the step of pulling target data from the target data source, it further includes:

[0032] In the case of multiple target data sources, the data architect agent determines whether the indicator dimensions of the target data pulled from each target data source are consistent.

[0033] In cases where the indicator dimensions of the target data are inconsistent, the target dimension is determined through the data architect agent and the first subtask, and the indicator dimensions of the target data are uniformly converted into the target dimension.

[0034] The data architect agent integrates the target data.

[0035] In one embodiment, the agent role pool includes: a quality assurance agent, and after the step of generating data analysis results, it further includes:

[0036] The quality assurance agent determines whether the prediction deviation between the predicted value in the data analysis result and the obtained actual value exceeds a preset threshold.

[0037] If the prediction deviation exceeds the preset threshold, the agent parameters of the data analysis and development agent are adjusted using a preset reinforcement learning strategy.

[0038] In one embodiment, the method further includes:

[0039] The input information and the output information of all agents in the agent role pool are stored in a unified manner to ensure that all agents in the agent role pool share the same context.

[0040] In one embodiment, the step of processing the received input information through a first large language model to generate standardized requirements includes:

[0041] The input information is processed by the large language model to determine the user intent and the semantic slots corresponding to the user intent;

[0042] The semantic slots are filled based on the input information to obtain the normalization requirements.

[0043] In one embodiment, prior to the step of processing the received input information through the first large language model to generate normalization requirements, the method further includes:

[0044] The soft target of the training data is obtained by performing intent recognition on the preset training data through the second language model.

[0045] The third language model is trained using the training data and the soft target to obtain the first language model.

[0046] Furthermore, to achieve the above objectives, this application also proposes an economic indicator visualization and orchestration system based on large-scale model multi-agent collaboration, the economic indicator visualization and orchestration system based on large-scale model multi-agent collaboration comprising:

[0047] The requirement parsing module is used to process the received input information through the first major language model, generate standardized requirements, and decompose the standardized requirements into multiple sub-tasks;

[0048] The task allocation module is used to allocate the multiple sub-tasks to other agents in a preset agent role pool through the product manager agent, wherein the agent role pool includes a data analysis and development agent and a visualization designer agent.

[0049] The data analysis module is used to determine the target data source through the data analysis and development agent and its assigned first sub-task, and generate data analysis results based on the data in the target data source;

[0050] The data visualization module is used to generate visualization outputs corresponding to the data analysis results through the visualization designer agent and its assigned second subtask.

[0051] In addition, to achieve the above objectives, this application also proposes an electronic device, the device comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, the computer program being configured to implement the steps of the economic indicator visualization orchestration method based on large model multi-agent collaboration as described above.

[0052] In addition, to achieve the above objectives, this application also proposes a storage medium, which is a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, it implements the steps of the above-described method for visualizing and orchestrating economic indicators based on large-model multi-agent collaboration.

[0053] In addition, to achieve the above objectives, this application also provides a computer program product, which includes a computer program that, when executed by a processor, implements the steps of the above-described method for visualizing and arranging economic indicators based on large-scale model multi-agent collaboration.

[0054] The one or more technical solutions proposed in this application have at least the following technical effects: First, the system receives user input regarding visualization requirements for economic indicators and processes this input information through a first major language model to generate accurate and clear standardized requirements, thereby reducing the misunderstanding of user requirements by the intelligent agent. Then, through the first major language model, the standardized requirements are decomposed into multiple sub-tasks to reduce the time wasted on manual task breakdown. Next, through a product manager intelligent agent, these sub-tasks are assigned to other intelligent agents in a pre-defined pool of intelligent agent roles, providing a foundation for collaboration among multiple intelligent agents. Furthermore, through the data analysis and development intelligent agent in the pool of intelligent agent roles and its assigned first sub-task, the target data source is determined, and data analysis results are generated based on the data in the target data source, avoiding the tedious operation of manually inputting data sources for searching and analyzing data, thus improving the efficiency and intelligence level of economic indicator analysis. Finally, through the visualization designer intelligent agent in the pool of intelligent agent roles and its assigned second sub-task, the data analysis results are transformed into intuitive visualization output, ensuring that the visualization output meets user needs while reducing the manual selection of visualization charts, thereby improving the efficiency of visualization output. This application achieves a fully autonomous process from understanding the needs of economic indicators to final delivery through collaboration between a large model and multiple agents. The large model and each agent can automatically and seamlessly connect without human intervention, thereby improving the efficiency of economic indicator visualization in the overall operation process. Attached Figure Description

[0055] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application.

[0056] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, for those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0057] Figure 1 This is a flowchart illustrating an embodiment of the method for visualizing and orchestrating economic indicators based on large-scale multi-agent collaboration in this application.

[0058] Figure 2 A flowchart illustrating the visualization of economic indicators provided in Embodiment 1 of this application;

[0059] Figure 3 This is a flowchart illustrating the economic data processing provided in Embodiment 2 of this application;

[0060] Figure 4 This is a flowchart illustrating the multi-agent collaboration and communication process provided in Embodiment 2 of this application.

[0061] Figure 5 This is the overall flowchart of multi-agent collaboration provided in Embodiment 3 of this application;

[0062] Figure 6 This is a schematic diagram of the module structure of the economic indicator visualization and orchestration system based on large-model multi-agent collaboration, as described in an embodiment of this application.

[0063] Figure 7 This is a schematic diagram of the device structure of the hardware operating environment involved in the economic indicator visualization and orchestration method based on large model multi-agent collaboration in the embodiments of this application.

[0064] The purpose, features, and advantages of this application will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0065] It should be understood that the specific embodiments described herein are merely illustrative of the technical solutions of this application and are not intended to limit this application.

[0066] It should be noted that in the description of this application and the appended claims, the terms "first," "second," etc., are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0067] It should be noted that all actions involving the acquisition of signals, information, or data in this application are carried out in compliance with the relevant data protection laws and policies of the country where the application is located, and with the authorization granted by the owner of the relevant device.

[0068] To better understand the technical solution of this application, a detailed description will be provided below in conjunction with the accompanying drawings and specific implementation methods.

[0069] Currently, the entire process of using AI-assisted tools to visualize economic indicators requires multiple human interventions, resulting in low visualization efficiency.

[0070] This application achieves a fully autonomous process from understanding the needs of economic indicators to final delivery through collaboration between a large model and multiple agents. The large model and each agent can automatically and seamlessly connect without human intervention, thereby improving the efficiency of economic indicator visualization in the overall operation process.

[0071] It should be noted that the executing entity in this embodiment can be an electronic device with data processing, network communication and program execution functions, such as a tablet computer, personal computer, mobile phone, etc.

[0072] Based on this, embodiments of this application provide a method for visualizing and orchestrating economic indicators based on large-scale model multi-agent collaboration, referring to... Figure 1 , Figure 1 This is a flowchart illustrating the first embodiment of the economic indicator visualization and orchestration method based on large-scale multi-agent collaboration in this application.

[0073] In this embodiment, the method for visualizing and orchestrating economic indicators based on large-scale model multi-agent collaboration includes steps S10 to S40:

[0074] Step S10: Process the received input information through the first large language model to generate normalization requirements, and decompose the normalization requirements into multiple sub-tasks;

[0075] The input information consists of visualization requests for economic indicators provided by users via text boxes or voice input interfaces. This information is presented in unstructured natural language and typically includes the name of the economic indicator, the time range, and the data presentation method (such as chart type). Economic indicators are quantitative characteristics used to measure the level or trend of economic activity, including GDP (Gross Domestic Product), CPI (Consumer Price Index), and interest rates.

[0076] Large Language Model (LLM) refers to a pre-trained deep learning model based on the Transformer architecture, which has the ability to understand and generate natural language; while the first large language model is a large language model used to process user input information and generate standardized requirements.

[0077] Standardization requirements refer to the structured instruction set output by the first major language model, which includes quantifiable parameters such as indicator type and spatiotemporal range, and can be presented in the form of JSON (JavaScript Object Notation).

[0078] Optionally, after receiving the input information from the user, it can be preprocessed by cleaning, word segmentation, and part-of-speech tagging, and the potential entities in the input information can be identified and standardized. For example, the colloquial "gross domestic product" can be mapped to a unified database identifier GDP. Then, the standardized input information can be processed through the first language model.

[0079] Optionally, after generating multiple subtasks, they and the standardization requirements can be further output to the interactive interface for user confirmation or adjustment.

[0080] Step S20: Through the product manager agent, assign multiple subtasks to other agents in the preset agent role pool;

[0081] An intelligent agent is an artificial intelligence entity that can perceive its environment and take actions to achieve specific goals. An intelligent agent role pool refers to a pre-registered cluster of intelligent agents, each with independent task processing capabilities and bound to specific capability tags, used to match user needs during sub-task allocation.

[0082] The intelligent agent role pool includes product manager intelligent agents, data analysis and development intelligent agents, and visualization designer intelligent agents. Among them, the product manager intelligent agent is an intelligent agent with task scheduling and resource allocation algorithms, which can realize sub-task allocation through ability tag matching; the data analysis and development intelligent agent is an intelligent agent that integrates multiple data analysis algorithms and data processing tools, and is responsible for performing data-related processing tasks such as data collection, cleaning, transformation, feature engineering, model selection, algorithm implementation, and data analysis; the visualization designer intelligent agent is an intelligent agent with built-in visualization design templates and graphic drawing tools to realize data visualization, and is responsible for designing visualization solutions based on data and sub-task requirements, including chart type, layout, color scheme, etc.

[0083] For example, after the first language model converts the user's input of economic indicator visualization demand information into multiple sub-tasks, these sub-tasks are sent to the product manager agent. The product manager agent can then match each sub-task with the ability tags of each agent in the agent role pool, and for any sub-task, assign it to the agent with the highest matching score.

[0084] Optionally, multiple subtasks can be assigned to other agents in a preset agent role pool through a product manager agent and a preset scheduler, wherein the scheduler is used to manage the context information of the product manager agent and the output information of other agents.

[0085] For example, when the product manager agent is assigning subtasks, long-term operation accumulation or sudden system failure may cause some of its data to be lost. At this time, the scheduler can be used to obtain the above information to recover the lost task assignment data and ensure the continuity and accuracy of task assignment.

[0086] Optionally, the product manager agent determines the dependencies between multiple subtasks and assigns the subtasks to other agents in the agent role pool; then, controls the other agents to execute the assigned subtasks according to the dependencies.

[0087] Step S30: Through the data analysis and development agent and its assigned first subtask, determine the target data source, and generate data analysis results based on the data in the target data source;

[0088] The first subtask is used to characterize the data processing-related subtasks assigned to the data analysis and development agent, which typically include explicit economic indicators, analysis requirements (such as growth rate), and spatiotemporal constraints.

[0089] The target data source refers to the data access point determined based on the analysis of the first subtask. It can be a database, an API (Application Programming Interface) provided by an official statistical structure, an unstructured document library, a real-time data stream, etc. This implementation method does not impose specific restrictions on it.

[0090] Data analysis results refer to the information and conclusions obtained by a data analysis and development agent after retrieving, analyzing, and / or predicting the target data source based on the first sub-task. The data analysis and development agent can select an appropriate analytical model (for tasks such as analyzing relationships between economic indicators) or a predictive model (for tasks such as economic trend forecasting and risk assessment) based on the first sub-task to analyze or predict the retrieved data and generate data analysis results.

[0091] For example, the data analysis and development agent can analyze information such as economic indicators and time ranges contained in the first sub-task, search and match in a known list of data sources or databases to determine the target data source containing the required data; then, it can construct a data request instruction according to the first sub-task and send the instruction to the target data source to obtain a dataset that meets the requirements of the first sub-task; then, based on the data type and task requirements, it can determine the data analysis method and model, and use the determined data analysis method and model to process and analyze the retrieved dataset to generate data analysis results.

[0092] In one feasible implementation, step S30, which involves determining the target data source through the data analysis and development agent and its assigned first subtask, includes:

[0093] Step S31: Through data analysis and development, the intelligent agent determines the target economic entity based on the first sub-task;

[0094] Economic entities refer to the objects of data correlation analysis. They can be economic indicators such as GDP and CPI, microeconomic entities such as industries and enterprises, or economic events, economic policies, geographical regions, time dimensions, etc. This implementation method does not impose specific restrictions on them.

[0095] The target economic entity can be obtained by semantic analysis of the first sub-task through data analysis and the development of intelligent agents, which typically includes the indicator name, region, and time.

[0096] Step S32: Through data analysis and development of intelligent agents, determine the target matching degree between the target economic entity and each data source in the preset data source database;

[0097] A data source repository is a knowledge base that stores information about multiple data sources in a structured manner. Each entry records metadata such as the name of each data source, API interface, list of economic indicators covered by the data source (using standardized names), time granularity (year / quarter / month / day), geographical granularity (country / province / city), data update frequency, authority level of the data provider, and data quality level.

[0098] Target matching degree refers to the strength of the association between the target economic entity and various data sources, which can be determined based on indicators such as semantic similarity, data dimension coverage and / or time granularity alignment.

[0099] In one feasible implementation, step S32 includes:

[0100] Step S321: Through data analysis and development of intelligent agents, determine the standardized name of the target economic entity based on the preset economic knowledge graph;

[0101] An economic knowledge graph is a graph-structured data structure consisting of nodes (economic entities) and edges (relationships between entities). Node attributes include name, type, region, and business scope, while edge attributes include relationships such as affiliation, cooperation, and competition. A standardized name refers to a unique and standardized name that identifies a specific economic entity within the economic knowledge graph.

[0102] Optionally, prior to step S321, economic data can be obtained using open APIs provided by various economic data sources (including statistical yearbooks, financial reports, policy documents, research papers, news reports, industry analysis reports, etc.). Then, a large-scale model is used to perform entity identification, relation extraction, and event extraction on this economic data to obtain economic entities (including economic events) and economic relationships within the economic field, and an economic knowledge graph is constructed based on this. The large-scale model is obtained by fine-tuning economic text data labeled with entities, relationships, and events.

[0103] Optionally, new data can be dynamically obtained from the aforementioned economic data sources, the knowledge extraction process described above can be repeated, and the economic knowledge graph can be dynamically updated and expanded based on the new knowledge extracted.

[0104] Optionally, during the dynamic construction of the economic knowledge graph, a portion of the knowledge graph content can be manually sampled to check for issues such as entity recognition errors or inaccurate relationship extraction, thereby ensuring the accuracy of the information in the economic knowledge graph.

[0105] Optionally, during the dynamic construction of the economic knowledge graph, economic data extracted can be analyzed and processed according to pre-set conflict detection and resolution rules. For example, when data extracted from different economic data sources conflict (such as different economic data sources having different records for the same economic indicator), the more credible knowledge can be determined based on factors such as the authority of the economic data source (such as prioritizing data from official statistical departments) and the freshness of the data (such as selecting more recent data), and the economic knowledge graph can be updated accordingly.

[0106] Understandably, by defining standardized names for target economic entities, the inconsistency in the expression of economic indicator names across different sources and scenarios can be effectively addressed. This helps reduce errors and duplication of work caused by name differences, thereby improving the efficiency of the entire economic indicator analysis and visualization process.

[0107] Step S322: Through data analysis and intelligent agent development, determine the keyword matching degree between the standardized name and the indicator items of each data source;

[0108] Keyword matching degree refers to the degree of lexical overlap between standardized names and data source metrics. It can be determined using algorithms such as TF-IDF (Term Frequency-Inverse Document Frequency) and BM25 (BestMatching 25).

[0109] Step S323: Through data analysis and development of intelligent agents, determine the semantic matching degree between the input information and the indicator items of each data source;

[0110] Semantic matching degree refers to the degree of matching obtained by comparing the deeper meaning of the concepts expressed by the input information and the data source indicators using natural language processing technology and semantic analysis algorithms. It is used to measure the consistency between the original indicator requirements and the data source indicators.

[0111] For example, for any data source, the BERT (Bidirectional Encoder Representations from Transformers) model can be used to convert the input information and the indicators of the data source into vector form respectively, and the cosine similarity between the two vectors can be calculated as the semantic matching degree between the input information and the indicators of the data source.

[0112] Step S324: Through data analysis and development of intelligent agents, the target matching degree is determined based on keyword matching degree and semantic matching degree.

[0113] For example, for any data source, the keyword matching degree and semantic matching degree between the target economic entity and the data source can be weighted and summed according to the preset weight allocation rules to obtain the target matching degree between the target economic entity and the data source.

[0114] Understandably, the hybrid matching algorithm that combines keyword matching and semantic matching overcomes the limitations of relying solely on name matching or semantic matching. It effectively avoids the misselection of data sources caused by similar names but inconsistent meanings, or semantically similar names but significantly different names, thereby improving the accuracy and reliability of data acquisition and ultimately enhancing the accuracy of economic indicator analysis.

[0115] Step S33: Determine the target data source by analyzing and developing intelligent agents and matching degrees with each target through data analysis.

[0116] For example, a data analysis and development agent can select the data source with the highest target matching degree or that meets a certain matching degree threshold from the data source library as the target data source.

[0117] Optionally, when multiple data sources can provide the same economic indicators, the priority of each data source that can provide the same economic indicators can be determined based on preset priority rules such as the authority score of the data source, the timeliness of data updates, the matching degree of data granularity, and the stability of API, and the data source with the highest priority can be determined as the target data source.

[0118] In this embodiment, the target data source is determined by combining economic knowledge graphs and hybrid matching algorithms to improve the accuracy of data source matching and ensure that economic indicator data that meets the user's needs can be obtained.

[0119] Step S40: Generate visualization output corresponding to the data analysis results through the visualization designer agent and its assigned second subtask.

[0120] The second subtask is used to characterize the data visualization-related subtasks assigned to the visualization designer agent, which typically include requirements for visualization form, style, elements, etc.

[0121] Visual output refers to the data analysis results generated by the visualization designer's intelligent agent and presented in an intuitive visual form, including interactive charts, analysis reports, large screen components, etc.

[0122] For example, the visualization designer agent can determine the horizontal and vertical axes and the corresponding visualization format based on the second subtask, such as a line chart, pie chart, or stacked bar chart. Assuming the visualization format is determined to be a line chart, the visualization designer agent can then use data analysis results on GDP changes over the past five years and its built-in drawing tools and algorithms to create a line chart, using the year as the horizontal axis and the GDP value as the vertical axis to reflect the changes in GDP values ​​over each year. Simultaneously, the visualization designer agent can automatically optimize the chart's layout, colors, fonts, labels, and other visual elements based on the potential audience and user preferences, ensuring the professionalism and readability of the visualization output.

[0123] For example, a visualization designer agent and its assigned second subtask can generate visualization charts corresponding to the data analysis results, and automatically generate chart titles, legends, and key insights based on the chart content, avoiding rigid data piling up; then, a product manager agent can generate an introduction and background information based on the user's input information; at the same time, a data analysis and development agent can generate key findings on economic indicators based on the data analysis methods and processes it uses, and generate conclusions and recommendations based on the data analysis results; then, the product manager agent can generate an economic analysis report by combining the above introduction, background information, data analysis methods, key findings, chart interpretation, conclusions, and recommendations.

[0124] Optionally, before generating the visualization output through the visualization designer agent, a preset large model can be used to select target charts from a preset chart library based on the data type of each data in the data analysis results and the user's visualization preferences, and generate a preset number of candidate layout schemes that arrange the target charts; then, through the visualization designer agent and the second subtask, the target layout scheme is determined from the candidate layout schemes, and the data analysis results are filled into the charts of the target layout scheme to obtain the visualization output corresponding to the data analysis results.

[0125] Optionally, when the input information includes multiple economic indicators, the visualization designer agent can generate corresponding visualization charts for each economic indicator; then, the visualization designer agent can arrange and layout the generated multiple visualization charts to generate a comprehensive interactive economic indicator dashboard, making it convenient for users to obtain information in one stop.

[0126] For example, please refer to Figure 2 , Figure 2 This document provides a flowchart illustrating the visualization of economic indicators. First, data input is provided to the visualization designer agent, including data analysis results and user visualization preferences. Then, the visualization designer agent designs the visualization based on the data input, leveraging a large model to provide various chart selections and layout options. Next, the visualization designer agent determines the final chart composition and target layout from the multiple layout options provided by the large model. Chart compositions can include line charts, GIS (Geographic Information System) maps, multi-dimensional bar charts, composite charts, etc., and this implementation does not impose specific limitations on these. Next, data is populated based on the data analysis results to obtain the visualization chart, and a chart interpretation is generated based on the chart content and data analysis results. Finally, according to the presentation format in the second subtask, the visualization chart and chart interpretation are combined to determine the final visualization output. The presentation format can include integrated dashboards, data dashboards, integrated large screens, analysis reports, etc.

[0127] In this embodiment, through the collaboration between the large model and multiple agents, a fully autonomous process from understanding the needs of economic indicators to final delivery is realized. The large model and each agent can automatically and seamlessly connect without human intervention, thereby improving the efficiency of economic indicator visualization in the overall operation process.

[0128] Based on the first embodiment of this application, in the second embodiment of this application, the content that is the same as or similar to that in the first embodiment described above can be referred to the above description and will not be repeated hereafter. Based on this, step S30, which generates data analysis results based on data from the target data source, includes:

[0129] Step S34: Through data analysis and development of intelligent agents, target data is retrieved from the target data source;

[0130] Target data refers to the data set pulled from the target data source, which may include different data formats such as tables and JSON.

[0131] For example, the data analysis and development agent can select the appropriate protocol based on the type of the target data source, construct an API request and send it to the target data source; then it can receive the data stream returned by the target data source and store it in a temporary cache for further processing.

[0132] In one feasible implementation, the agent role pool includes: a data architect agent, and after step S34, it further includes:

[0133] Step S341: In the case of multiple target data sources, the data architect agent determines whether the indicator dimensions of the target data pulled from each target data source are consistent.

[0134] A data architect agent is an intelligent agent with the ability to analyze data structures and validate dimensions. It typically integrates algorithms such as data lineage analysis and data dimension alignment, and is responsible for designing data pipelines, determining data storage schemes, and planning data flow.

[0135] Metric dimensions refer to the set of metadata attributes for each field in the target data, including data type, unit, etc. For example, the "sales amount" field may be counted as "day / month" in different data sources.

[0136] Step S342: In the case where the indicator dimensions of each target data are inconsistent, the target dimension is determined through the data architect agent and the first subtask, and the indicator dimensions of each target data are uniformly converted into the target dimension.

[0137] The target dimension refers to the standard dimension for achieving unified analysis and comparison of data.

[0138] For example, when the data architect agent detects inconsistencies in the indicator dimensions of the target data, it parses the first subtask and determines the target dimension that the user expects to display from the indicator dimensions corresponding to each target data according to the user's needs. Then, based on the dimension differences, the data architect agent uses professional conversion algorithms in the economic field (such as deflator, exchange rate conversion, etc.) to unify the indicator dimensions of each target data into the target dimension.

[0139] Step S343: Integrate the target data through the data architect intelligent agent.

[0140] For example, the data architect agent can integrate the target data based on the standard names of the fields in the target data. For instance, it can integrate the data from the past three years and the data from three years ago for a single field based on the standard names, so that it can be viewed and accessed in a unified manner.

[0141] In this implementation, the data architect agent automatically performs dimensional transformation and data integration, eliminating dimensional differences and data inconsistencies between different data sources, constructing a unified and comprehensive dataset, providing a reliable data foundation for economic indicator analysis, and improving the accuracy of data analysis.

[0142] Step S35: Through data analysis and development of intelligent agents, the target data is cleaned to obtain the cleaned target data;

[0143] Data cleaning refers to the process of detecting and correcting anomalies in target data based on economic rules or economic knowledge graphs, including data correction, outlier handling, missing value imputation, and data quality verification.

[0144] For example, data analysis and development agents can use large language models to perform semantic understanding and contextual analysis on target data, identify and correct typos and inconsistent formats in the target data; then use statistical methods and machine learning algorithms (such as the Isolation Forest algorithm) to automatically detect and process outliers; then intelligently fill in missing data based on time series forecasts, regression models or contextual information; and then combine economic knowledge graphs to perform logical consistency verification (e.g., GDP cannot be less than CPI), data timeliness verification (ensuring that the data is up-to-date and available), and data integrity checks.

[0145] Understandably, data cleaning can effectively remove noise, errors, and duplicate information from data, improving the accuracy, completeness, and consistency of target data, thereby enhancing the accuracy of economic indicator analysis.

[0146] Step S36: The data analysis and development agent retrieves the cleaned target data according to the first sub-task and generates data analysis results.

[0147] For example, the data analysis and development agent parses the first subtask to determine the search conditions, such as time range and field filtering rules; then, it converts the search conditions into the corresponding database query language or search instructions to search the cleaned target data and obtain the search results; then, according to the user requirements in the first subtask, it calls the relevant analysis algorithm or model to analyze and process the search results and generate data analysis results.

[0148] In one feasible implementation, step S36 includes:

[0149] Step S361: Through data analysis and development, the intelligent agent determines the target economic entity based on the first sub-task;

[0150] The specific implementation of step S361 can be referred to the specific implementation of step S31 in the first embodiment above, and will not be repeated here.

[0151] Step S362: Determine the extended information of the target economic entity based on the preset economic knowledge graph;

[0152] It should be noted that step S362 can use the economic knowledge graph constructed in the first embodiment.

[0153] Extended information refers to other economic concepts, entities, or information related to the target economic entity, obtained through reasoning from an economic knowledge graph.

[0154] For example, a data analysis and development intelligent agent can automatically expand a target economic entity to its synonyms, related entities, or higher-level concepts based on entities and relationships in an economic knowledge graph, thereby obtaining extended information corresponding to the target economic entity and improving retrieval recall. For instance, the entity "digital economy" can be expanded to "information industry," "big data," "artificial intelligence," etc.

[0155] Step S363: Based on the extended information, perform keyword retrieval on the cleaned target data to obtain the first retrieval result;

[0156] The first search result is used to characterize the structured data obtained through keyword retrieval.

[0157] For example, the data development and analysis agent can use a preset text retrieval algorithm (such as the BM25 algorithm) to match different fields in the extended information with the indicator list of the cleaned target data using keywords to determine the matched target indicator; then, it can extract the data corresponding to the target indicator from the cleaned target data, and combine the target indicator and the corresponding data to obtain the first retrieval result.

[0158] Understandably, by using keywords from the extended information to retrieve the cleaned target data, one can further focus on data related to the target economic entity, thereby improving the effectiveness of data analysis.

[0159] Step S364: Using the fine-tuned large language model, the input information and extended information are converted into text vectors. The fine-tuned large language model is obtained by fine-tuning training data in the economic field.

[0160] A fine-tuned large language model refers to a specialized language model obtained by further training on training data in the economic field, based on the original large language model. Text vectors, on the other hand, are vector representations of semantic features extracted from the fine-tuned large language model to reflect input and extended information.

[0161] Step S365: Determine the second search result based on the vector similarity between the text vector and each indicator item in the cleaned target data;

[0162] Vector similarity is a quantitative indicator used to measure the degree of similarity between two vectors in a high-dimensional space. It can be calculated using methods such as cosine similarity and Euclidean distance.

[0163] The second search result is used to characterize the structured data determined through vector similarity calculation.

[0164] For example, the data development and analysis agent can use a fine-tuned large language model to vectorize the input and extended information to obtain text vectors; similarly, using the same fine-tuned large language model, it can vectorize each indicator item in the cleaned target data to obtain multiple indicator vectors; then, using a cosine similarity algorithm, it can calculate the vector similarity between the text vectors and each indicator vector; then, the indicator items corresponding to the indicator vectors whose vector similarity exceeds a preset threshold can be identified as target indicators; then, the data corresponding to the target indicators can be extracted from the cleaned target data, and the target indicator and the corresponding data can be combined to obtain the second retrieval result.

[0165] Understandably, by calculating the vector similarity between the text vector and the target data indicator item to determine the second search result, accurate data retrieval based on semantic understanding is achieved, overcoming the problem that traditional keyword retrieval may miss semantically related but literally mismatched data.

[0166] Step S366: Determine the target search result based on the first search result and the second search result, and generate data analysis results based on the target search result.

[0167] For example, the data development and analysis agent can combine the first and second search results according to the indicator name, remove duplicate indicator items, and obtain the target search result; then, it can call the data analysis algorithm to analyze the target search result and generate the data analysis result.

[0168] Optionally, after obtaining the first and second search results, the indicators can be ranked based on factors such as the matching degree between each indicator and the input information, the authority of the corresponding data source, the publication time, and the connectivity in the economic knowledge graph (i.e., the strength of the association with other important economic concepts in the economic knowledge graph). The ranked indicators can then be combined to obtain the target search results. Furthermore, in subsequent visualization processes, indicators that meet user needs and have strong relevance can be prioritized for display to improve the user's visualization experience.

[0169] For example, please refer to Figure 3 , Figure 3 This document provides a flowchart illustrating the economic data processing process. First, the data input is determined, which can be read from a data source database. Each data source can be from the National Bureau of Statistics, provincial data interfaces, industry reports, real-time financial data streams, etc. Next, a data analysis and development agent is used to discover and connect data sources. After identifying the target data source, data is retrieved from it. In cases where multiple target data sources exist, the target data from different sources undergoes dimensional transformation and integration to obtain a complete set of target data (which can be presented in the form of a dataset / data table, etc.). Then, the integrated target data undergoes data cleaning and verification, including data correction, outlier handling, missing value imputation, and data quality verification, resulting in cleaned target data. Finally, data retrieval is performed using a knowledge graph, which can be an integration of multiple knowledge graphs, such as an economic knowledge graph, a basic statistical knowledge graph, or an industry-specific private domain knowledge graph. Data retrieval can include keyword retrieval and vector retrieval. The target retrieval results, obtained by combining the results of both methods, represent economic data related to the economic indicators in the input information.

[0170] In this embodiment, by combining keyword retrieval and vector retrieval, the breadth of data is ensured, as well as its relevance and accuracy, thereby improving the comprehensiveness and precision of the retrieval results and thus enhancing the accuracy of economic indicator analysis.

[0171] In one feasible implementation, the agent role pool includes: a quality assurance agent, and after step S30, it further includes:

[0172] Step A10: The quality assurance agent determines whether the prediction deviation between the predicted value in the data analysis results and the obtained actual value exceeds a preset threshold.

[0173] A quality assurance agent is an intelligent agent with capabilities such as state perception, deviation calculation, root cause analysis, and model optimization. It is responsible for comprehensive testing and verification of data accuracy, model effectiveness, visualization presentation, and system functionality.

[0174] Optionally, throughout the entire lifecycle of visualizing economic indicators, a quality assurance agent can collaboratively monitor the entire visualization orchestration process in real time. This includes: monitoring the data source connection status to ensure stable data interfaces and uninterrupted data flow; monitoring the efficiency and resource consumption of data cleaning, transformation, and integration processes to monitor data processing performance; tracking the running speed and accuracy of predictive models or analytical algorithms; monitoring visualization generation efficiency to ensure that the generation speed of charts and reports meets requirements; and monitoring the usage of resources such as CPU (Central Processing Unit), memory, and storage.

[0175] For example, a quality assurance agent can assess data quality by monitoring the integrity, accuracy, consistency, and timeliness of target data.

[0176] For example, the quality assurance agent can evaluate the predictive model used by the data development and analysis agent during data analysis using metrics such as R-squared, MAE (Mean Absolute Error), RMSE (Root Mean Squared Error), and MAPE (Mean Absolute Percentage Error). Similarly, it can evaluate the analytical model used by the data development and analysis agent during data analysis using metrics such as model fit, significance level, and residuals. Furthermore, the backpropagation algorithm and evaluation results can be used to adjust the predictive or analytical model, or replace the model used by the data development and analysis agent.

[0177] Step A20: When the prediction deviation exceeds a preset threshold, the agent parameters of the data analysis and development agent are adjusted using a preset reinforcement learning strategy.

[0178] Prediction bias refers to the absolute or relative error between the model's predicted value and the actual value.

[0179] Reinforcement learning is a machine learning method that uses an agent to interact with its environment and update its behavioral policy (agent parameters) based on the feedback reward signals. It includes algorithms such as PPO (Proximal Policy Optimization) and SAC (Soft Actor-Critic).

[0180] Among them, agent parameters define the internal variables for data analysis and development of agent behavior and decision-making logic, which may include the called model, model parameters, feature weights, time decay factors, etc.

[0181] For example, when a data analysis and development agent calls a prediction model to generate data analysis results, the quality assurance agent periodically compares the predicted value in the data analysis results with the currently acquired time to determine the prediction deviation; and when the prediction deviation exceeds a preset threshold, a penalty signal is generated based on the prediction deviation; then, the agent parameters of the data analysis and development agent can be adjusted through the PPO algorithm and the penalty signal, such as adjusting the model parameters or the called prediction model.

[0182] In this embodiment, the agent parameters are automatically adjusted through reinforcement learning strategies, enabling the data analysis and development agent to automatically optimize its behavior when faced with excessive prediction deviations. This gradually improves the accuracy and reliability of predictions, reduces human intervention, and enhances the reliability of economic indicator analysis while ensuring its efficiency.

[0183] In one feasible implementation, the method further includes:

[0184] Step A30: Store the input information and the output information of all agents in the agent role pool in a unified manner to ensure that all agents in the agent role pool share the context.

[0185] For example, any agent in the agent role pool can be composed of an independent large language model instance. When executing the corresponding sub-task, it can perform independent analysis within its large language model. When the sub-task is completed, the corresponding output information is saved to a preset shared database so that other agents can retrieve the data.

[0186] Optionally, for any agent in the agent role pool, when a conflict is detected (such as lack of file creation permission), a timer is started, and the agent communicates with other agents in the agent role pool to resolve the conflict, such as asking the product manager agent if it is possible to avoid creating a new file; the timer stops when the conflict is resolved, but if the time for detecting the conflict exceeds a preset time threshold, an alarm message can be output to seek help from the user.

[0187] For example, please refer to Figure 4 , Figure 4A flowchart illustrating multi-agent collaboration and communication is provided. First, the input information is processed through a primary language model to determine standardization requirements and multiple sub-tasks. This input information, standardization requirements, and sub-tasks are all stored in a unified shared memory. Next, the product manager agent assigns the sub-tasks to different agents, and the corresponding sub-task assignment information is also stored in the shared memory. Then, each agent executes its assigned sub-task. For example, the data analysis and development agent can determine the target data source based on its assigned sub-task, retrieve data, and perform data retrieval and analysis. After retrieving data from different data sources, it can leverage the data architect agent to perform data dimension transformation and data integration. The output information and communication context of the data analysis and development agent and the data architect agent are stored in the shared memory for other agents to access. Finally, the visualization designer agent can perform visualization design and chart analysis based on its assigned sub-tasks and the data analysis results retrieved from the shared memory, and output the final visualization results. Throughout the entire lifecycle of the visualization output, the quality assurance agent can monitor and comprehensively evaluate the execution process and data of other agents, and adjust the agent parameters of each agent in reverse based on the evaluation results.

[0188] In this embodiment, by uniformly storing input information and agent output information, information flow between agents is realized, avoiding the problems of duplication of work and information inconsistency, thereby ensuring the continuity of the entire analysis process and improving the efficiency of economic indicator visualization and the reliability of analysis results.

[0189] Based on the first and / or second embodiments of this application, in the third embodiment of this application, the content that is the same as or similar to the first and second embodiments described above can be referred to the above description and will not be repeated hereafter. Based on this, the step S10 of processing the received input information through the first large language model to generate standardized requirements includes:

[0190] Step S11: Process the input information through the first large language model to determine the user intent and the semantic slots corresponding to the user intent;

[0191] User intent refers to the core user request extracted by the large language model, while semantic slots refer to the key parameters to be filled under the user intent, which are determined autonomously by the large language model based on the training data.

[0192] In one feasible implementation, prior to step S10, the method further includes:

[0193] Step S01: Use the second language model to perform intent recognition on the preset training data to obtain the soft target of the training data;

[0194] Soft target refers to the probability distribution output by the teacher model during knowledge distillation, including the probability distribution of intent, entity type, and relation type corresponding to each word or phrase in the training data, rather than a single category label.

[0195] The second large language model represents the teacher model used to output soft objectives. It can be the original large language model, such as DeepSeek or Qwen, or it can be a large language model that has been fine-tuned with training data from the economic field. This implementation does not impose any specific restrictions on it.

[0196] Step S02: Train the third language model using training data and soft targets to obtain the first language model.

[0197] The third major language model represents the learning model in the knowledge distillation process, and it is usually set as the original major language model.

[0198] For example, for a set of pre-defined training data containing various economic analysis-related text samples, the second language model performs intent recognition on these samples, outputting the probability distribution of each sample belonging to different intent categories, forming a soft target. Then, these training data and soft targets are input into the training framework to train the third language model, enabling it to learn how to generate an intent probability distribution similar to the soft target based on the input text, that is, to learn the "experience" of the teacher model (the second language model) in generalizing and processing fuzzy information. After multiple rounds of iterative optimization, the performance of the third language model reaches the expected level or the loss function converges, resulting in the first language model, which is used to process actual user input.

[0199] For example, the first large language model trained by this distillation can be used to process the input information. For the input information "I want to see the trend of China's GDP growth rate and CPI index over the past five years, preferably in a chart for comparison", the corresponding user intent identified is "comparison of economic indicator trends". The semantic slots under this user intent include economic indicators, regions, time ranges, visualization forms, etc.

[0200] In this embodiment, the first language model is trained through knowledge distillation, which can not only identify economic entities from input information, but also identify the relationships between different entities, thereby more accurately identifying the user intent corresponding to the input information.

[0201] Step S12: Extract data content from the input information to fill semantic slots and obtain the standardization requirements.

[0202] Data content refers to the entity or attribute values ​​in the input information that are related to the semantic slots.

[0203] For example, after extracting the data content, it can be filled in directly, or it can be filled in after inference through the first large language model. For instance, for the input information "I want to see the trend of China's GDP growth rate and CPI index over the past five years", the data content corresponding to the region slot is determined to be "China", which can be filled in directly; while the data content corresponding to the time range slot is "the past five years", which can be filled in after determining the specific year information (such as 2020-2025) through the large language model. Furthermore, by combining the user intent and the filled semantic slots, the standardized requirements corresponding to the input information are obtained.

[0204] In this embodiment, a first large language model is trained using knowledge distillation technology, and this first large language model is used for intent recognition to generate standardized requirements. This improves the accuracy of user intent recognition, ensures that the overall economic indicator visualization process is based on accurate user requirements, reduces user intervention, and improves the efficiency of economic indicator visualization and the accuracy of economic indicator analysis results.

[0205] For example, to help understand the implementation process of the economic indicator visualization and orchestration method based on large-model multi-agent collaboration obtained in this embodiment combined with the above embodiment one, please refer to... Figure 5 , Figure 5This paper presents a general flowchart for multi-agent collaboration. First, an interactive interface is provided where users can input their economic analysis needs and visualization goals through natural language text boxes or voice input. Then, an intelligent intent and requirement parsing module parses the user input and outputs a preliminary standardized requirement and task decomposition plan to the interactive interface for user confirmation or fine-tuning. Next, the standardized requirement and multiple sub-tasks confirmed by the user are sent to the multi-agent collaborative management and orchestration module. The product manager agent allocates tasks, the data analysis and development agent mines target data sources, performs data retrieval and analysis, and the data architect agent... The system transforms and integrates data from different data sources, uses a visualization designer agent to select and lay out visualization charts, and outputs the final visualization results (including visualization dashboards, interactive charts, and economic analysis reports). The multi-agent heterogeneous economic data processing and knowledge fusion module provides access interfaces to heterogeneous data sources and economic knowledge graphs for each agent, enabling them to obtain comprehensive and accurate economic indicator data for analysis. Simultaneously, the real-time progress of each agent is displayed on the interactive interface, such as "The data architect agent is designing data pipelines" and "The data analysis and development agent is processing data and building models," allowing users to understand the current project progress. Furthermore, the multi-agent collaborative management and orchestration module provides a quality assurance agent to continuously monitor the performance of the visualization output process and the stability of the underlying data sources, and optimizes each agent based on user suggestions for improving published results. In addition, within the multi-agent collaborative management and orchestration module, when an agent detects a long-standing unresolved conflict or an inability to reach a consensus, it can send a request to a human supervisor (user) and resolve the conflict or negotiate based on user suggestions. Finally, once the complete visualization output is displayed on the interactive interface, users can export the visualization results as documents, images, or other formats with a single click by triggering the publishing control, or deploy it to an internal data platform for others to view and interact with.

[0206] It should be noted that the above examples are only for understanding this application and do not constitute a limitation on the method for visualizing and arranging economic indicators based on large-scale multi-agent collaboration. Any simple modifications based on this technical concept are within the scope of protection of this application.

[0207] It should be noted that the economic indicator visualization and arrangement method based on large-scale model multi-intelligence collaboration proposed in this application is not limited to the field of economic indicator visualization. It can also be widely applied to various scenarios that require efficient and autonomous management of complex information processing and decision-making processes. For example, in supply chain and logistics optimization systems, intelligent agent teams can collaboratively analyze data from different links (such as production, transportation, inventory, and sales), autonomously identify bottlenecks, and generate optimization solutions and visualization reports. In smart city construction, intelligent agent teams can autonomously analyze and visualize urban operation data (such as traffic flow, environmental indicators, public safety incidents, and population density), assisting city managers in understanding urban operation, predicting development trends, and assisting in the formulation of urban planning and emergency response strategies. It can also be applied to scenarios such as scientific research project management and data analysis, financial risk management and investment analysis, and intelligent marketing and customer relationship management to improve management efficiency, accuracy, and intelligence levels in various scenarios. This application does not impose specific limitations on specific application scenarios.

[0208] This application also provides an economic indicator visualization and orchestration system based on large-scale model multi-agent collaboration. Please refer to... Figure 6 The economic indicator visualization and orchestration system based on large-scale model multi-agent collaboration includes:

[0209] The requirement parsing module 10 is used to process the received input information through the first major language model, generate standardized requirements, and decompose the standardized requirements into multiple sub-tasks;

[0210] The task assignment module 20 is used to assign multiple sub-tasks to other agents in a preset agent role pool through the product manager agent. The agent role pool includes a data analysis and development agent and a visualization designer agent.

[0211] The data analysis module 30 is used to determine the target data source through the data analysis and development agent and its assigned first sub-task, and generate data analysis results based on the data in the target data source;

[0212] The data visualization module 40 is used to generate visualization outputs corresponding to the data analysis results through the visualization designer agent and its assigned second subtask.

[0213] The economic indicator visualization and orchestration system based on large-model multi-agent collaboration provided in this application adopts the economic indicator visualization and orchestration method based on large-model multi-agent collaboration in the above embodiments, and can solve the technical problem of how to improve the efficiency of economic indicator visualization. Compared with the prior art, the beneficial effects of the economic indicator visualization and orchestration system based on large-model multi-agent collaboration provided in this application are the same as the beneficial effects of the economic indicator visualization and orchestration method based on large-model multi-agent collaboration provided in the above embodiments, and other technical features of the economic indicator visualization and orchestration system based on large-model multi-agent collaboration are the same as the features disclosed in the methods of the above embodiments, and will not be repeated here.

[0214] This application provides an electronic device, which includes: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, which are executed by the at least one processor to enable the at least one processor to execute the economic indicator visualization orchestration method based on large model multi-agent collaboration in Embodiment 1 described above.

[0215] The following is for reference. Figure 7 The diagram illustrates a structural schematic of an electronic device suitable for implementing embodiments of this application. The electronic devices in these embodiments may include, but are not limited to, mobile terminals such as mobile phones, laptops, digital broadcast receivers, PDAs (Personal Digital Assistants), PADs (Portable Application Descriptions), PMPs (Portable Media Players), in-vehicle terminals (e.g., in-vehicle navigation terminals), and fixed terminals such as digital TVs and desktop computers. Figure 7 The electronic device shown is merely an example and should not impose any limitation on the functionality and scope of use of the embodiments of this application.

[0216] like Figure 7As shown, the electronic device may include a processing unit 1001 (e.g., a central processing unit, a graphics processing unit, etc.), which can perform various appropriate actions and processes according to a program stored in a read-only memory 1002 or a program loaded from a storage device 1003 into a random access memory 1004. The random access memory 1004 also stores various programs and data required for the operation of the electronic device. The processing unit 1001, the read-only memory 1002, and the random access memory 1004 are interconnected via a bus 1005. An input / output interface 1006 is also connected to the bus. Typically, the following systems can be connected to the input / output interface 1006: input devices 1007 including, for example, touchscreens, touchpads, keyboards, mice, image sensors, microphones, accelerometers, gyroscopes, etc.; output devices 1008 including, for example, liquid crystal displays (LCDs), speakers, vibrators, etc.; storage devices 1003 including, for example, magnetic tapes, hard disks, etc.; and communication devices 1009. The communication device 1009 allows the electronic device to communicate wirelessly or wiredly with other devices to exchange data. Although the diagrams show electronic devices with various systems, it should be understood that it is not required to implement or have all of the systems shown. More or fewer systems may be implemented alternatively.

[0217] Specifically, according to the embodiments disclosed in this application, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments disclosed in this application include a computer program product comprising a computer program carried on a computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via a communication device, or installed from storage device 1003, or installed from read-only memory 1002. When the computer program is executed by processing device 1001, it performs the functions defined in the methods of the embodiments disclosed in this application.

[0218] The electronic device provided in this application adopts the economic indicator visualization and arrangement method based on large model multi-agent collaboration in the above embodiments, which can solve the technical problem of how to improve the efficiency of economic indicator visualization. Compared with the prior art, the beneficial effects of the electronic device provided in this application are the same as those of the economic indicator visualization and arrangement method based on large model multi-agent collaboration provided in the above embodiments, and other technical features of the electronic device are the same as those disclosed in the previous embodiment method, and will not be repeated here.

[0219] It should be understood that the various parts disclosed in this application can be implemented using hardware, software, firmware, or a combination thereof. In the description of the above embodiments, specific features, structures, materials, or characteristics can be combined in any suitable manner in one or more embodiments or examples.

[0220] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0221] This application provides a computer-readable storage medium having computer-readable program instructions (i.e., a computer program) stored thereon, which are used to execute the economic indicator visualization and orchestration method based on large model multi-agent collaboration in the above embodiments.

[0222] The computer-readable storage medium provided in this application embodiment may be, for example, a USB flash drive, but is not limited to, electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems or devices, or any combination thereof. More specific examples of computer-readable storage media may include, but are not limited to: electrical connections having one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fibers, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof. In this embodiment, the computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system or device. The program code contained on the computer-readable storage medium may be transmitted using any suitable medium, including but not limited to: wires, optical cables, RF (Radio Frequency), etc., or any suitable combination thereof.

[0223] The aforementioned computer-readable storage medium may be included in an electronic device or may exist independently without being assembled into an electronic device.

[0224] The aforementioned computer-readable storage medium carries one or more programs, which, when executed by an electronic device, cause the electronic device to perform the functions defined in the methods of the embodiments disclosed in this application.

[0225] Computer program code for performing the operations of this application can be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages ​​such as Java, Smalltalk, and C++, and conventional procedural programming languages ​​such as the "C" language or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer can be connected to the user's computer via any type of network—including a Local Area Network (LAN) or a Wide Area Network (WAN)—or can be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0226] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this application. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0227] The modules described in the embodiments of this application can be implemented in software or hardware. The names of the modules do not necessarily limit the functionality of the unit itself.

[0228] The readable storage medium provided in this application is a computer-readable storage medium. This computer-readable storage medium stores computer-readable program instructions (i.e., a computer program) for executing the above-described method for visualizing and arranging economic indicators based on large-model multi-agent collaboration, thus solving the technical problem of how to improve the efficiency of economic indicator visualization. Compared with the prior art, the beneficial effects of the computer-readable storage medium provided in this application are the same as those of the economic indicator visualization and arranging method based on large-model multi-agent collaboration provided in the above-described embodiments, and will not be elaborated upon here.

[0229] This application also provides a computer program product, including a computer program that, when executed by a processor, implements the steps of the above-described method for visualizing and orchestrating economic indicators based on large-model multi-agent collaboration.

[0230] The computer program product provided in this application can solve the technical problem of how to improve the efficiency of economic indicator visualization. Compared with the prior art, the beneficial effects of the computer program product provided in this application are the same as those of the economic indicator visualization arrangement method based on large model multi-agent collaboration provided in the above embodiments, and will not be repeated here.

[0231] The above description is only a part of the embodiments of this application and does not limit the patent scope of this application. All equivalent structural transformations made under the technical concept of this application and using the contents of the specification and drawings of this application, or direct / indirect applications in other related technical fields, are included in the patent protection scope of this application.

Claims

1. A method for visualizing and orchestrating economic indicators based on large-scale multi-agent collaboration, characterized in that, The method for visualizing and orchestrating economic indicators based on large-model multi-agent collaboration includes: The received input information is processed by the first major language model to generate standardized requirements, and the standardized requirements are decomposed into multiple sub-tasks. The input information is the visualization requirements for economic indicators input by the user. The product manager agent assigns the multiple sub-tasks to other agents in a preset agent role pool. The agent role pool includes a data analysis and development agent and a visualization designer agent. The product manager agent is an agent with task scheduling and resource allocation algorithms. Sub-task allocation is achieved through capability tag matching. The data analysis and development agent is an agent that integrates multiple data analysis algorithms and data processing tools and is responsible for executing data-related processing tasks. Through data analysis and development, intelligent agents identify target economic entities based on their assigned first sub-tasks; Through the data analysis and development intelligent agent, the standardized name of the target economic entity is determined based on the preset economic knowledge graph; The data analysis and development intelligent agent determines the keyword matching degree between the standardized name and the indicator items of each data source in the preset data source library; Through the data analysis and development intelligent agent, the semantic matching degree between the input information and the indicator items of each of the data sources is determined; The data analysis and development intelligent agent determines the target matching degree between the target economic entity and each of the data sources based on the keyword matching degree and the semantic matching degree. The target data source is determined by analyzing and developing the intelligent agent and the matching degree of each target; Data analysis results are generated based on the data from the target data source; The visualization designer agent and its assigned second subtask generate the visualization output corresponding to the data analysis results. The visualization designer agent is an agent with built-in visualization design templates and graphic drawing tools to realize data visualization, and is responsible for designing visualization schemes based on data and subtask requirements.

2. The method for visualizing and orchestrating economic indicators based on large-scale model multi-agent collaboration as described in claim 1, characterized in that, The step of generating data analysis results based on the data in the target data source includes: The data analysis and development intelligent agent retrieves target data from the target data source. The data analysis and development intelligent agent cleans the target data to obtain cleaned target data. The data analysis and development agent retrieves the cleaned target data based on the first sub-task and generates data analysis results.

3. The method for visualizing and orchestrating economic indicators based on large-scale model multi-agent collaboration as described in claim 2, characterized in that, The step of retrieving the cleaned target data and generating data analysis results through the data analysis and development intelligent agent based on the first sub-task includes: The data analysis and development agent determines the target economic entity based on the first sub-task; Based on a pre-defined economic knowledge graph, the extended information of the target economic entity is determined; Based on the extended information, keyword retrieval is performed on the cleaned target data to obtain the first retrieval result; The input information and the extended information are converted into text vectors using a fine-tuned large language model, wherein the fine-tuned large language model is obtained by fine-tuning training data in the economic field; The second retrieval result is determined based on the vector similarity between the text vector and each indicator item in the cleaned target data; Based on the first search result and the second search result, the target search result is determined, and based on the target search result, data analysis results are generated.

4. The method for visualizing and orchestrating economic indicators based on large-scale model multi-agent collaboration as described in claim 2, characterized in that, The pool of intelligent agent roles includes: a data architect intelligent agent, and after the step of pulling target data from the target data source, it further includes: In the case of multiple target data sources, the data architect agent determines whether the indicator dimensions of the target data pulled from each target data source are consistent. In cases where the indicator dimensions of the target data are inconsistent, the target dimension is determined through the data architect agent and the first subtask, and the indicator dimensions of the target data are uniformly converted into the target dimension. The data architect agent integrates the target data.

5. The method for visualizing and orchestrating economic indicators based on large-scale model multi-agent collaboration as described in claim 1, characterized in that, The agent role pool includes: a quality assurance agent, and after the step of generating data analysis results, it also includes: The quality assurance agent determines whether the prediction deviation between the predicted value in the data analysis result and the obtained actual value exceeds a preset threshold. If the prediction deviation exceeds the preset threshold, the agent parameters of the data analysis and development agent are adjusted using a preset reinforcement learning strategy.

6. The method for visualizing and orchestrating economic indicators based on large-scale model multi-agent collaboration as described in claim 1, characterized in that, The method further includes: The input information and the output information of all agents in the agent role pool are stored in a unified manner to ensure that all agents in the agent role pool share the same context.

7. The method for visualizing and orchestrating economic indicators based on large-scale model multi-agent collaboration as described in claim 1, characterized in that, The step of processing the received input information through the first large language model to generate standardized requirements includes: The input information is processed by the first large language model to determine the user intent and the semantic slots corresponding to the user intent; The semantic slots are filled based on the input information to obtain the normalization requirements.

8. The method for visualizing and orchestrating economic indicators based on large-scale model multi-agent collaboration as described in claim 7, characterized in that, Before the step of processing the received input information through the first large language model to generate standardized requirements, the following steps are also included: The soft target of the training data is obtained by performing intent recognition on the preset training data through the second language model. The third language model is trained using the training data and the soft target to obtain the first language model.

Images