Method and system for interactive billboard interpretation assistant based on large model

Abstract

The invention relates to the technical field of artificial intelligence, and particularly discloses a method and system for an interactive billboard interpretation assistant based on a large model, and the method comprises the steps: collecting data from different data sources at regular time, carrying out the preprocessing of the collected data, and inputting the collected data as original data into a bottom database; establishing analysis modules of different statistical analysis modes, processing the original data in the underlying database based on the analysis modules to obtain an analysis result, and storing the analysis result in the underlying database; original data and an analysis result thereof are obtained from the underlying database based on user requirements, a pre-trained large model is called to formulate an interpretation process and a splitting interpretation task, and an interpretation report is generated; and receiving specific data input by a user, training the fine tuning large model according to the specific data, and performing fine tuning on the large model in the interpretation process. The method has strong semantic comprehension and generalization ability, and the execution result can be changed by adjusting the input cue word, so that multiple tasks can be completed without heavy training.

Description

Technical Field

[0001] The present invention relates to the field of artificial intelligence technology, and in particular to a method and system for an interactive signboard interpretation assistant based on a large model. Background Art

[0002] The interactive dashboard interpretation assistant based on the big model is an intelligent tool that combines the powerful natural language processing (NLP) capabilities of the big model with the visual dashboard function to provide users with real-time data interpretation, analysis suggestions and interactive experience.

[0003] The existing technology mainly includes five modules, namely, a data acquisition module, which is responsible for collecting and entering data sources for different business scenarios; a database module, which is used to store and manage data, including business data, report template data, knowledge base data, etc.; a data analysis module, which is used to analyze business data, which may include basic statistical descriptions, change analysis, distribution analysis, comparative analysis, and predictive analysis, etc.; an interpretation generation module, which is used to generate data interpretation reports based on the analysis results using rule templates and pre-trained models; and a visualization module, which is used to display the statistical analysis results and interpretation reports of the data.

[0004] The shortcoming of the existing technology is that the interpretation process lacks intermediate steps: an interpretation process may need to be divided into multiple stages for step-by-step reasoning, while the existing methods all use the source data analysis results and templates as the input of the generative model. The lack of intermediate steps will amplify the error of the generated effect; in addition, the generative model lacks generalization: the generative model is further trained based on the pre-trained model, and the training data is related to the business scenario, so it does not have the generalization ability of cross-scenario interpretation, which limits the scalability of the interpretation system. Summary of the invention

[0005] The purpose of the present invention is to provide a method and system for an interactive signboard interpretation assistant based on a large model to solve the problems raised in the above-mentioned background technology.

[0006] To achieve the above object, the present invention provides the following technical solutions:

[0007] A method for an interactive kanban interpretation assistant based on a large model, the method comprising:

[0008] Create an underlying database based on the preset data structure, build an ETL task flow, collect data from different data sources at regular intervals, pre-process the collected data, and input it into the underlying database as raw data;

[0009] Establish analysis modules with different statistical analysis modes, process the raw data in the underlying database based on the analysis modules, obtain analysis results, and store them in the underlying database;

[0010] configuring a data table, and adjusting a data collection process and a data analysis process based on the data table;

[0011] Based on user needs, we obtain the original data and its analysis results from the underlying database, call the pre-trained large model to formulate the interpretation process and split the interpretation tasks, and generate the interpretation report;

[0012] Receive specific data input by the user, train and fine-tune the large model according to the specific data, and fine-tune the large model in the interpretation process.

[0013] As a further solution of the present invention: the steps of creating an underlying database according to a preset data structure, constructing an ETL task flow, collecting data from different data sources at a regular interval, and preprocessing the collected data before inputting the data into the underlying database as raw data include:

[0014] Receive a data structure input by a user, and create an underlying database based on the data structure;

[0015] Identify different data sources and determine the data source type;

[0016] Establish connections with data sources based on pre-defined connectors and drivers;

[0017] Based on the established connection, data is collected from different data sources at a regular interval, and after pre-processing, the collected data is input into the underlying database as raw data.

[0018] As a further solution of the present invention: the steps of establishing analysis modules of different statistical analysis modes, processing raw data in the underlying database based on the analysis modules, obtaining analysis results, and storing them in the underlying database include:

[0019] Establish analysis modules for different statistical analysis modes;

[0020] Process the raw data in the underlying database based on the analysis module to obtain the analysis results;

[0021] The storage location of the original data is queried in the underlying database, the storage location of the analysis results is determined based on the storage location of the original data, and the analysis results are stored in the underlying database.

[0022] As a further solution of the present invention: the method further comprises:

[0023] Read the analysis results of the analysis module and visualize the analysis results; wherein the visualization process corresponds to the statistical analysis mode of the analysis module. When the statistical analysis mode adopts the cumulative statistical mode, the visualization process adopts a bar chart; when the statistical analysis mode adopts the time series statistical mode, the visualization process adopts a line chart.

[0024] As a further solution of the present invention: the method further comprises:

[0025] Built-in multi-agent submodule in large models;

[0026] When receiving a specific dimension input by the user, the data is filtered based on the multi-agent submodule;

[0027] The opening and closing ports of the multi-agent submodule are user-oriented.

[0028] The technical solution of the present invention also provides a system of interactive signboard interpretation assistant based on a large model, the system comprising:

[0029] The data collection module is used to create an underlying database according to the preset data structure, build an ETL task flow, collect data from different data sources at a regular interval, and input the collected data into the underlying database as raw data after preprocessing;

[0030] Data analysis module, used to establish analysis modules of different statistical analysis modes, process the raw data in the underlying database based on the analysis module, obtain analysis results, and store them in the underlying database;

[0031] An operation adjustment module is used to configure a data table and adjust a data collection process and a data analysis process based on the data table;

[0032] The task splitting and interpretation module is used to obtain the original data and its analysis results in the underlying database based on user needs, call the pre-trained large model to formulate the interpretation process and split the interpretation tasks, and generate the interpretation report;

[0033] The model fine-tuning module is used to receive specific data input by the user, train and fine-tune the large model according to the specific data, and fine-tune the large model during the interpretation process.

[0034] As a further solution of the present invention: the data acquisition module includes:

[0035] A database creation unit, used to receive a data structure input by a user, and create an underlying database based on the data structure;

[0036] A data source identification unit, used to identify different data sources and determine the data source type;

[0037] A connection establishing unit, used for establishing a connection with a data source based on a preset connector and a driver level;

[0038] The raw data generation unit is used to collect data from different data sources based on the established connection timing, and after pre-processing the collected data, input it into the underlying database as raw data.

[0039] As a further solution of the present invention: the data analysis module includes:

[0040] An analysis module creation unit, used to establish analysis modules of different statistical analysis modes;

[0041] An analysis execution unit, used to process the raw data in the underlying database based on the analysis module to obtain analysis results;

[0042] The storage connection unit is used to query the storage location of the original data in the underlying database, determine the storage location of the analysis result according to the storage location of the original data, and store the analysis result in the underlying database.

[0043] As a further solution of the present invention: the working content of the system also includes:

[0044] Read the analysis results of the analysis module and visualize the analysis results; wherein the visualization process corresponds to the statistical analysis mode of the analysis module. When the statistical analysis mode adopts the cumulative statistical mode, the visualization process adopts a bar chart; when the statistical analysis mode adopts the time series statistical mode, the visualization process adopts a line chart.

[0045] As a further solution of the present invention: the working content of the system also includes:

[0046] Built-in multi-agent submodule in large models;

[0047] When receiving a specific dimension input by the user, the data is filtered based on the multi-agent submodule;

[0048] The opening and closing ports of the multi-agent submodule are user-oriented.

[0049] Compared with the prior art, the beneficial effects of the present invention are: the large model has strong semantic understanding and generalization capabilities, and can change the execution results by adjusting the input prompt words, so multiple tasks can be completed without heavy training; for the interpretation of dashboard data, splitting multiple tasks and arranging them in series can improve flexibility and accuracy, and the large model can play the role of each sub-module at the same time, which improves the effect while reducing the complexity of customized training sub-algorithm modules; in addition, the ability to fine-tune the large model and configure the interpretation process is retained, so that the system can more flexibly adapt to different scenarios and different data. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings in the following description are only some embodiments of the present invention.

[0051] Figure 1 A flowchart of the method for interpreting an assistant for interactive kanban boards based on a large model. DETAILED DESCRIPTION

[0052] In order to make the technical problems, technical solutions and beneficial effects to be solved by the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0053] Figure 1 The flowchart of the method of the interactive billboard interpretation assistant based on the large model is as follows. In an embodiment of the present invention, a method of the interactive billboard interpretation assistant based on the large model includes:

[0054] Create an underlying database based on the preset data structure, build an ETL task flow, collect data from different data sources at regular intervals, pre-process the collected data, and input it into the underlying database as raw data;

[0055] Establish analysis modules with different statistical analysis modes, process the raw data in the underlying database based on the analysis modules, obtain analysis results, and store them in the underlying database;

[0056] configuring a data table, and adjusting a data collection process and a data analysis process based on the data table;

[0057] Based on user needs, we obtain the original data and its analysis results from the underlying database, call the pre-trained large model to formulate the interpretation process and split the interpretation tasks, and generate the interpretation report;

[0058] Receive specific data input by the user, train and fine-tune the large model according to the specific data, and fine-tune the large model in the interpretation process.

[0059] This system provides a method and system implementation for completing analytical data interpretation through big model related technologies (including SFT and Agent, etc.). It mainly utilizes the semantic understanding and generalization capabilities of the big model itself, and through a small amount of data fine-tuning and Agent orchestration, a unified interpretation system for different businesses and different data types is realized.

[0060] Specifically, this system is built around four core functions, which are described as follows:

[0061] i. Data collection and storage, including the underlying database and the ETL task flow for writing data.

[0062] ii. Various statistical analysis modes and visual displays, including cumulative statistics and pie chart and bar chart comparison, time series statistics and line chart, etc.

[0063] iii. Interpretation generation module: input analysis data and original data, the big model formulates the interpretation process, splits the interpretation tasks, summarizes the results and generates an interpretation report.

[0064] iv. Large model fine-tuning and orchestration: Supports customized fine-tuning of large models to align with specific requirements; and supports the design and orchestration of the analysis and interpretation process.

[0065] The process of using this system is as follows:

[0066] First, for data tasks, according to the dimensions of the input data, configure the data table and configure the scheduled task execution of data collection and analysis, that is, determine the frequency of the data collection process and the frequency of the data analysis process; after the user clicks the button, the background big model service executes a series of interpretation algorithms, generates the final interpretation report, and displays it on the front-end dashboard. At the same time, fine-tuning the big model requires preparing specific training data, writing it to the database data table, and executing fine-tuning tasks in the background; further, the big model service has a built-in multi-agent sub-module and provides configuration parameters, allowing users to choose whether to filter specific dimensions in the input data, or start and close specific Agent sub-modules to orchestrate the interpretation algorithm process.

[0067] Regarding the above-mentioned drawings, there are seven parts in the drawings, which are described as follows:

[0068] (1) Data collection: Configure ETL scheduled tasks around the database and write business data according to interface standards

[0069] (2) Database: underlying storage of business data

[0070] (3) Data analysis: Perform statistical analysis on business data as needed to generate analytical data such as cumulative features and time series features

[0071] (4) Visual dashboard: used to visualize analytical data and interpret reports

[0072] (5) Interpretation of the big model: The original data and analysis data are used as input, and the corresponding interpretation report is generated through the big model. The analysis process is broken down into multiple stages within the module.

[0073] Interpretation planning and distribution: Multiple analysis data may correspond to different interpretation angles, and are suitable for splitting into independent and complete subtasks. Here, the big model is first used to split the task and then distribute it to different sub-modules.

[0074] Sub-module interpretation: Sub-modules are only for specific data and can be divided by type or by business. For example, time series data usually focuses on trends and anomalies, while bar charts focus more on dimensional comparisons.

[0075] Interpretation summary: Collect the conclusions of all sub-modules, and then submit them to the big model to summarize and generate an interpretation report

[0076] (6) Instruction fine-tuning module: For scenarios with customized requirements, a small amount of data instructions can be used to fine-tune the alignment requirements.

[0077] (7) Multi-Agent: The interpretation process involves multiple steps and supports customized sub-module arrangement to complete the task, such as focusing the interpretation on a specific product.

[0078] Furthermore, the steps of creating an underlying database according to a preset data structure, constructing an ETL task flow, collecting data from different data sources at a regular interval, and preprocessing the collected data before inputting it into the underlying database as raw data include:

[0079] Receive a data structure input by a user, and create an underlying database based on the data structure;

[0080] Identify different data sources and determine the data source type;

[0081] Establish connections with data sources based on pre-defined connectors and drivers;

[0082] Based on the established connection, data is collected from different data sources at a regular interval, and after pre-processing, the collected data is input into the underlying database as raw data.

[0083] The data structure in the underlying database is preset and set by the user. After the data structure is set, the underlying database can be created based on the data structure. Then, the present invention needs to collect data from different data sources, identify the data sources of different data sources, determine the data source type, and establish a connection with the data source according to the preset connector and drive level. There is a data structure conversion and matching process in this process. Data is collected from different data sources at a fixed time based on the established connection. After pre-processing the collected data, it is input into the underlying database as raw data.

[0084] As a preferred embodiment of the technical solution of the present invention, the steps of establishing analysis modules of different statistical analysis modes, processing raw data in the underlying database based on the analysis modules, obtaining analysis results, and storing them in the underlying database include:

[0085] Establish analysis modules for different statistical analysis modes;

[0086] Process the raw data in the underlying database based on the analysis module to obtain the analysis results;

[0087] The storage location of the original data is queried in the underlying database, the storage location of the analysis results is determined based on the storage location of the original data, and the analysis results are stored in the underlying database.

[0088] Different analysis modules are established for different statistical analysis modes. The analysis modules process the original data in the underlying database to obtain analysis results. The analysis results correspond one-to-one to the original data. When storing, the analysis results need to be stored according to the location of the original data, that is, the storage locations of the corresponding analysis results and the original data are also corresponding.

[0089] As a preferred embodiment of the technical solution of the present invention, the method further includes:

[0090] Read the analysis results of the analysis module and visualize the analysis results; wherein the visualization process corresponds to the statistical analysis mode of the analysis module. When the statistical analysis mode adopts the cumulative statistical mode, the visualization process adopts a bar chart; when the statistical analysis mode adopts the time series statistical mode, the visualization process adopts a line chart.

[0091] The visualization display process is used to display the analysis results. It is connected to the visualization dashboard, can display the analysis statistical data, and provide interactive buttons for users to choose to generate an interpretation report. This application provides a variety of statistical analysis modes and supporting visualization display solutions, including cumulative statistics and pie chart bar chart comparisons, as well as time series statistics and line chart comparisons.

[0092] As a preferred embodiment of the technical solution of the present invention, the method further includes:

[0093] Built-in multi-agent submodule in large models;

[0094] When receiving a specific dimension input by the user, the data is filtered based on the multi-agent submodule;

[0095] The opening and closing ports of the multi-agent submodule are user-oriented.

[0096] By configuring a multi-agent system, users can choose whether to filter data dimensions or enable / disable specific interpretation algorithm submodules to customize the interpretation process; multi-agent can be implemented using reinforcement learning or rule-based systems.

[0097] As a preferred embodiment of the technical solution of the present invention, the present invention also provides a system of interactive signboard interpretation assistant based on a large model, the system comprising:

[0098] The data collection module is used to create an underlying database according to the preset data structure, build an ETL task flow, collect data from different data sources at a regular interval, and input the collected data into the underlying database as raw data after preprocessing;

[0099] Data analysis module, used to establish analysis modules of different statistical analysis modes, process the raw data in the underlying database based on the analysis module, obtain analysis results, and store them in the underlying database;

[0100] An operation adjustment module is used to configure a data table and adjust a data collection process and a data analysis process based on the data table;

[0101] The task splitting and interpretation module is used to obtain the original data and its analysis results in the underlying database based on user needs, call the pre-trained large model to formulate the interpretation process and split the interpretation tasks, and generate the interpretation report;

[0102] The model fine-tuning module is used to receive specific data input by the user, train and fine-tune the large model according to the specific data, and fine-tune the large model during the interpretation process.

[0103] Furthermore, the data acquisition module includes:

[0104] A database creation unit, used to receive a data structure input by a user, and create an underlying database based on the data structure;

[0105] A data source identification unit, used to identify different data sources and determine the data source type;

[0106] A connection establishing unit, used for establishing a connection with a data source based on a preset connector and a driver level;

[0107] The raw data generation unit is used to collect data from different data sources based on the established connection timing, and after pre-processing the collected data, input it into the underlying database as raw data.

[0108] Specifically, the data analysis module includes:

[0109] An analysis module creation unit, used to establish analysis modules of different statistical analysis modes;

[0110] An analysis execution unit, used to process the raw data in the underlying database based on the analysis module to obtain analysis results;

[0111] The storage connection unit is used to query the storage location of the original data in the underlying database, determine the storage location of the analysis result according to the storage location of the original data, and store the analysis result in the underlying database.

[0112] On the basis of the above contents, the working contents of the system also include:

[0113] Read the analysis results of the analysis module and visualize the analysis results; wherein the visualization process corresponds to the statistical analysis mode of the analysis module. When the statistical analysis mode adopts the cumulative statistical mode, the visualization process adopts a bar chart; when the statistical analysis mode adopts the time series statistical mode, the visualization process adopts a line chart.

[0114] The system also includes:

[0115] Built-in multi-agent submodule in large models;

[0116] When receiving a specific dimension input by the user, the data is filtered based on the multi-agent submodule;

[0117] The opening and closing ports of the multi-agent submodule are user-oriented.

[0118] For the above system, its work content mainly includes the following four parts:

[0119] 1. Data collection and storage:

[0120] Underlying database: Choose a suitable database (such as MySQL, PostgreSQL) to store raw data and analytical data.

[0121] ETL task flow: Set up ETL task flow to collect data from different data sources at regular intervals, pre-process the data, and write the data to the database. Tools such as Apache NiFi and Airflow can be used to automate the data pipeline.

[0122] 2. Multiple statistical analysis modes and visual displays:

[0123] Statistical analysis: supports analysis modes such as cumulative statistics and time series statistics. Libraries such as Pandas and NumPy can be used for data analysis.

[0124] Visual display: integrated chart libraries (such as Matplotlib, Plotly, ECharts), supporting common visualization forms such as bar charts, pie charts, and line charts.

[0125] 3. Interpretation of the generation module:

[0126] Input analysis data and original data: Input analysis results and original data as required.

[0127] Large model: Call pre-trained large models (such as GPT, BERT) to formulate interpretation processes, split interpretation tasks, and generate interpretation reports. Models can be called using Python APIs or through microservices.

[0128] 4. Fine-tuning and choreography of large models:

[0129] Fine-tuning large models: Supports loading customized data sets and fine-tuning large models to meet specific task requirements. You can use frameworks such as Hugging Face and TensorFlow for fine-tuning operations.

[0130] Interpretation Process Orchestration: By configuring a multi-agent system, users can choose whether to filter data dimensions or enable / disable specific interpretation algorithm submodules to customize the interpretation process. Multi-agent can be implemented using Reinforcement Learning or rule-based systems.

[0131] The functions that can be achieved by the method of the interactive signboard interpretation assistant based on the large model are all completed by a computer device, and the computer device includes one or more processors and one or more memories, and at least one program code is stored in the one or more memories. The program code is loaded and executed by the one or more processors to achieve the functions of the method of the interactive signboard interpretation assistant based on the large model.

[0132] The processor takes out instructions from the memory one by one, analyzes the instructions, and then completes the corresponding operations according to the instruction requirements, generating a series of control commands, so that the various parts of the computer can automatically, continuously and coordinately move to become an organic whole, realize the input of programs, the input of data, and the calculation and output of results. The arithmetic operations or logical operations generated in this process are all completed by the operator; the memory includes a read-only memory (ROM), which is used to store computer programs, and a protection device is provided outside the memory.

[0133] Exemplarily, the computer program may be divided into one or more modules, one or more modules are stored in a memory and executed by a processor to implement the present invention. One or more modules may be a series of computer program instruction segments capable of implementing specific functions, and the instruction segments are used to describe the execution process of the computer program in a terminal device.

[0134] Those skilled in the art will understand that the description of the above service equipment is merely an example and does not constitute a limitation on the terminal equipment. It may include more or fewer components than described above, or a combination of certain components, or different components, for example, it may include input and output devices, network access devices, buses, etc.

[0135] The processor may be a central processing unit (CPU), other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or any conventional processor, etc. The processor is the control center of the terminal device, and uses various interfaces and lines to connect various parts of the entire user terminal.

[0136] The memory can be used to store computer programs and / or modules. The processor can realize various functions of the terminal device by running or executing the computer programs and / or modules stored in the memory and calling the data stored in the memory. The memory can mainly include a program storage area and a data storage area, wherein the program storage area can store an operating system, an application required for at least one function (such as an information collection template display function, a product information release function, etc.); the data storage area can store data created according to the use of the berth status display system (such as product information collection templates corresponding to different product types, product information that different product providers need to release, etc.). In addition, the memory can include a high-speed random access memory, and can also include a non-volatile memory, such as a hard disk, a memory, a plug-in hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) card, a flash card (Flash Card), at least one disk storage device, a flash memory device, or other volatile solid-state storage devices.

[0137] If the module / unit integrated in the terminal device is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the present invention implements all or part of the modules / units in the above-mentioned embodiment system, and can also be completed by instructing the relevant hardware through a computer program. The above-mentioned computer program can be stored in a computer-readable storage medium, and the computer program can realize the functions of the above-mentioned various system embodiments when executed by the processor. Among them, the computer program includes computer program code, and the computer program code can be in source code form, object code form, executable file or some intermediate form. Computer-readable media may include: any entity or device capable of carrying computer program code, recording medium, U disk, mobile hard disk, disk, optical disk, computer memory, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), electrical carrier signal, telecommunication signal and software distribution medium, etc.

[0138] It should be noted that, in this article, the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, an element defined by the sentence "comprises a ..." does not exclude the existence of other identical elements in the process, method, article or device including the element.

[0139] The above are only preferred embodiments of the present invention, and are not intended to limit the patent scope of the present invention. Any equivalent structure or equivalent process transformation made using the contents of the present invention specification and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present invention.

Claims

1. A method for interactive signboard interpretation assistant based on a large model, characterized in that: The method comprises: Create an underlying database based on the preset data structure, build an ETL task flow, collect data from different data sources at regular intervals, pre-process the collected data, and input it into the underlying database as raw data; Establish analysis modules with different statistical analysis modes, process the raw data in the underlying database based on the analysis modules, obtain analysis results, and store them in the underlying database; configuring a data table, and adjusting a data collection process and a data analysis process based on the data table; Based on user needs, we obtain the original data and its analysis results from the underlying database, call the pre-trained large model to formulate the interpretation process and split the interpretation tasks, and generate the interpretation report; Receive specific data input by the user, train and fine-tune the large model according to the specific data, and fine-tune the large model in the interpretation process.

2. The method of interactive signboard interpretation assistant based on large model according to claim 1, characterized in that: The steps of creating an underlying database according to a preset data structure, constructing an ETL task flow, collecting data from different data sources at a regular interval, and pre-processing the collected data before inputting it into the underlying database as raw data include: Receive a data structure input by a user, and create an underlying database based on the data structure; Identify different data sources and determine the data source type; Establish connections with data sources based on pre-defined connectors and drivers; Based on the established connection, data is collected from different data sources at a regular interval, and after pre-processing, the collected data is input into the underlying database as raw data.

3. The method of interactive signboard interpretation assistant based on large model according to claim 1, characterized in that: The steps of establishing analysis modules of different statistical analysis modes, processing raw data in the underlying database based on the analysis modules, obtaining analysis results, and storing them in the underlying database include: Establish analysis modules for different statistical analysis modes; Process the raw data in the underlying database based on the analysis module to obtain the analysis results; The storage location of the original data is queried in the underlying database, the storage location of the analysis results is determined based on the storage location of the original data, and the analysis results are stored in the underlying database.

4. The method of interactive signboard interpretation assistant based on large model according to claim 1, characterized in that: The method further comprises: Read the analysis results of the analysis module and visualize the analysis results; wherein the visualization process corresponds to the statistical analysis mode of the analysis module. When the statistical analysis mode adopts the cumulative statistical mode, the visualization process adopts a bar chart; when the statistical analysis mode adopts the time series statistical mode, the visualization process adopts a line chart.

5. The method of interactive signboard interpretation assistant based on large model according to claim 1, characterized in that: The method further comprises: Built-in multi-agent submodule in large models; When receiving a specific dimension input by the user, the data is filtered based on the multi-agent submodule; The opening and closing ports of the multi-agent submodule are user-oriented.

6. A system of interactive signboard interpretation assistant based on a large model, characterized in that: The system comprises: The data collection module is used to create an underlying database according to the preset data structure, build an ETL task flow, collect data from different data sources at a regular interval, and input the collected data into the underlying database as raw data after preprocessing; Data analysis module, used to establish analysis modules of different statistical analysis modes, process the raw data in the underlying database based on the analysis module, obtain analysis results, and store them in the underlying database; An operation adjustment module is used to configure a data table and adjust a data collection process and a data analysis process based on the data table; The task splitting and interpretation module is used to obtain the original data and its analysis results in the underlying database based on user needs, call the pre-trained large model to formulate the interpretation process and split the interpretation tasks, and generate the interpretation report; The model fine-tuning module is used to receive specific data input by the user, train and fine-tune the large model according to the specific data, and fine-tune the large model during the interpretation process.

7. The system of interactive signboard interpretation assistant based on large model according to claim 6, characterized in that: The data acquisition module comprises: A database creation unit, used to receive a data structure input by a user and create an underlying database based on the data structure; A data source identification unit, used to identify different data sources and determine the data source type; A connection establishing unit, used for establishing a connection with a data source based on a preset connector and a driver level; The raw data generation unit is used to collect data from different data sources based on the established connection timing, and after pre-processing the collected data, input it into the underlying database as raw data.

8. The system of interactive signboard interpretation assistant based on large model according to claim 6, characterized in that: The data analysis module includes: An analysis module creation unit, used to establish analysis modules of different statistical analysis modes; An analysis execution unit, used to process the raw data in the underlying database based on the analysis module to obtain analysis results; The storage connection unit is used to query the storage location of the original data in the underlying database, determine the storage location of the analysis result according to the storage location of the original data, and store the analysis result in the underlying database.

9. The system of interactive signboard interpretation assistant based on large model according to claim 6, characterized in that: The system also includes: Read the analysis results of the analysis module and visualize the analysis results; wherein the visualization process corresponds to the statistical analysis mode of the analysis module. When the statistical analysis mode adopts the cumulative statistical mode, the visualization process adopts a bar chart; when the statistical analysis mode adopts the time series statistical mode, the visualization process adopts a line chart.

10. The system of interactive signboard interpretation assistant based on large model according to claim 6, characterized in that: The system also includes: Built-in multi-agent submodule in large models; When receiving a specific dimension input by the user, the data is filtered based on the multi-agent submodule; The opening and closing ports of the multi-agent submodule are user-oriented.

Images