Data analysis method and device, computer equipment, storage medium and computer program product

Abstract

The invention relates to a data analysis method and device, computer equipment, a storage medium and a computer program product. The method comprises the following steps: acquiring a data analysis request message, and determining corresponding data analysis demand information based on the data analysis request message; based on the data analysis demand information, obtaining a target index model including a data analysis index and a data analysis instruction template corresponding to the target index model; obtaining a target data table model matched with the target index model, wherein the target data table model comprises target analysis data meeting a data analysis index; and performing data analysis based on the target analysis data and the data analysis instruction template to obtain a data analysis result corresponding to the data analysis request message. By adopting the method, the accuracy of data analysis can be improved.

Description

Technical Field

[0001] The present application relates to the field of computer technology, and in particular to a data analysis method, apparatus, computer equipment, storage medium and computer program product. Background Art

[0002] With the rapid development of computer technology, data has become a core resource in modern society. In all walks of life, data structures have shown unprecedented complexity and diversity. From simple text and numbers to complex images, videos, network logs, sensor data, etc., the data types are numerous and the formats are different. Data analysis based on various data structures has become an important tool for decision-making, business optimization and market insights in all walks of life. However, complex data structures increase the difficulty of data analysis, resulting in low accuracy of data analysis. Summary of the invention

[0003] Based on this, it is necessary to provide a data analysis method, device, computer equipment, computer-readable storage medium and computer program product that can improve the accuracy of data analysis in response to the above technical problems.

[0004] In a first aspect, the present application provides a data analysis method. The method comprises:

[0005] Obtaining a data analysis request message, and determining corresponding data analysis requirement information based on the data analysis request message;

[0006] Based on the data analysis requirement information, obtaining a target indicator model including data analysis indicators and a data analysis instruction template corresponding to the target indicator model;

[0007] Obtaining a target data table model that matches the target indicator model, wherein the target data table model includes target analysis data that meets the data analysis indicator;

[0008] Perform data analysis based on the target analysis data and the data analysis instruction template to obtain the data analysis result corresponding to the data analysis request message.

[0009] In a second aspect, the present application also provides a data analysis device. The device comprises:

[0010] A demand determination module, used to obtain a data analysis request message and determine corresponding data analysis demand information based on the data analysis request message;

[0011] An indicator model acquisition module is used to acquire a target indicator model including data analysis indicators and a data analysis instruction template corresponding to the target indicator model based on data analysis requirement information;

[0012] A data table model acquisition module is used to obtain a target data table model that matches the target indicator model, wherein the target data table model includes target analysis data that meets the data analysis indicator;

[0013] The analysis result acquisition module is used to perform data analysis based on the target analysis data and the data analysis instruction template to obtain the data analysis result corresponding to the data analysis request message.

[0014] In a third aspect, the present application further provides a computer device, wherein the computer device comprises a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the steps of the above data analysis method are implemented.

[0015] In a fourth aspect, the present application further provides a computer-readable storage medium having a computer program stored thereon, and the computer program implements the steps of the above data analysis method when executed by a processor.

[0016] In a fifth aspect, the present application further provides a computer program product, which includes a computer program, and when the computer program is executed by a processor, the steps of the above data analysis method are implemented.

[0017] The above-mentioned data analysis method, apparatus, computer equipment, storage medium and computer program product, based on the data analysis requirement information determined according to the data analysis request message, obtains the target indicator model including the data analysis indicator and the data analysis instruction template corresponding to the target indicator model, obtains the target data table model matching the target indicator model, and performs data analysis based on the target analysis data and the data analysis instruction template included in the target data table model, and obtains the data analysis result corresponding to the data analysis request message. In the data analysis process, the target indicator model and the corresponding data analysis instruction template are obtained through the data analysis requirement information, and the matching target data table model is obtained based on the target indicator model, so as to perform data analysis through the target analysis data and the data analysis instruction template that meet the data analysis indicator included in the target data table model, and various data can be standardized through the target data table model, so as to perform data analysis processing on the standardized data, thereby reducing the difficulty of data analysis and improving the accuracy of data analysis. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the related technologies, the drawings required for use in the embodiments or the related technical descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0019] Figure 1 An application environment diagram of a data analysis method in one embodiment;

[0020] Figure 2 A schematic diagram of a data analysis method in one embodiment;

[0021] Figure 3 A schematic diagram of a process for determining data analysis requirement information in one embodiment;

[0022] Figure 4 A schematic diagram of a process of performing data analysis based on a large model in one embodiment;

[0023] Figure 5 is a flow chart of a data analysis method in yet another embodiment;

[0024] Figure 6 A schematic diagram of a flow chart of data pre-processing in one embodiment;

[0025] Figure 7 A schematic diagram of a process of intent recognition in one embodiment;

[0026] Figure 8 A schematic diagram of comparison of data analysis results in one embodiment;

[0027] Fig. 9 is a structural block diagram of a data analysis device in one embodiment;

[0028] Fig.10 is an internal structure diagram of a computer device in one embodiment;

[0029] Fig.11 FIG. 4 is a diagram showing the internal structure of a computer device in yet another embodiment. DETAILED DESCRIPTION

[0030] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application 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 application and are not used to limit the present application.

[0031] The data analysis method provided in the embodiment of the present application can be applied to Figure 1In the application environment shown. Among them, the terminal 102 communicates with the server 104 through the network. The data storage system can store the data that the server 104 needs to process. The data storage system can be set up separately, integrated on the server 104, or placed on the cloud or other servers. The user can send a data analysis request message to the server 104 through the terminal 102. The server 104 can determine the corresponding data analysis requirement information based on the data analysis request message, and obtain the target indicator model including the data analysis indicator and the data analysis instruction template corresponding to the target indicator model based on the data analysis requirement information. The server 104 can obtain a target data table model that matches the target indicator model, and perform data analysis based on the target analysis data and the data analysis instruction template included in the target data table model to obtain the data analysis result corresponding to the data analysis request message. The server 104 can also return the data analysis result to the terminal 102, so that it can be displayed in the terminal 102.

[0032] In some embodiments, the data analysis method may also be implemented independently by the terminal 102 or the server 104. For example, the terminal 102 may directly perform data analysis based on the data analysis request message configured by the user, thereby obtaining corresponding data analysis results. For another example, the server 104 may directly obtain the data analysis request message from the data storage system to perform data analysis based on the data analysis request message.

[0033] The terminal 102 may be, but is not limited to, various desktop computers, laptop computers, smart phones, tablet computers, IoT devices, and portable wearable devices. The IoT devices may be smart speakers, smart TVs, smart air conditioners, smart vehicle-mounted devices, etc. The portable wearable devices may be smart watches, smart bracelets, head-mounted devices, etc. The server 104 may be implemented as an independent server or a server cluster consisting of multiple servers, or may be a cloud server.

[0034] In an exemplary embodiment, Figure 2 As shown, a data analysis method is provided, which is executed by a computer device, and can be executed by a computer device such as a terminal or a server alone, or by a terminal and a server together. In the embodiment of the present application, the method is applied to Figure 1 The server in the example is used to illustrate, including the following steps 202 to 208. Among them:

[0035] Step 202: Obtain a data analysis request message, and determine corresponding data analysis requirement information based on the data analysis request message.

[0036] Among them, the data analysis request message is a request message for requesting data analysis processing for business data, which can be obtained by sending the user to the server through the terminal. The data analysis demand information is used to characterize the intention of the user to request data analysis through the data analysis request message, such as the business data targeted by the data analysis, the analysis indicators of the data points, and the measurement values ​​of the data analysis, etc. The data analysis demand information can be determined based on the data analysis request message. Exemplarily, the server can receive the data analysis request message sent by the user through the terminal, and the server can perform demand analysis on the data analysis request message to determine the corresponding data analysis demand information. For example, the user can edit a data analysis question, such as "How is the revenue of the sales department this month?" "Why did the revenue drop this month?" The terminal can generate a data analysis request message based on the data analysis question edited by the user, and send the data analysis request message to the server. The server can perform intent recognition based on the data analysis question included in the data analysis request message to determine the corresponding data analysis demand information. For example, for the data analysis question "How is the revenue of the sales department this month?", the corresponding data analysis demand information can be a data analysis of the revenue data within the time range of the sales department that month.

[0037] Step 204: based on the data analysis requirement information, obtain a target indicator model including data analysis indicators and a data analysis instruction template corresponding to the target indicator model.

[0038] Among them, the target indicator model is an indicator model corresponding to the data analysis demand information determined from various indicator models, and the indicator model includes data analysis indicators to perform data analysis on business data through data analysis indicators. Data analysis indicators can be indicators for performing data analysis on business data, such as including but not limited to at least one of various indicators such as indicator name, statistical cycle, analysis dimension, filtering condition, data comparison condition, data prediction condition, etc. Different indicator models may include different data analysis indicators, so that different data analysis demand information can be responded to. The data analysis instruction template can be a template for generating data analysis instructions, and the data analysis instruction template corresponds to the target indicator model, so that different data analysis demand information can be responded to. Through the data analysis instruction template, corresponding data analysis instructions can be generated based on the business data that needs to be analyzed, thereby realizing data analysis processing.

[0039] Optionally, the server can match based on the data analysis requirement information to match the target indicator model and the data analysis instruction template. The target indicator model includes data analysis indicators, and the business data required for data analysis can be obtained based on the data analysis indicators. In some embodiments, the server can match with each pre-configured indicator model based on the data analysis requirement information, thereby determining a matching target indicator model from each indicator model, and further determining the data analysis instruction template corresponding to the target indicator model. In some embodiments, the same indicator model can be configured with at least one data analysis instruction template. When the target indicator model corresponds to multiple data analysis instruction templates, the server can match from multiple data analysis instruction templates corresponding to the target indicator model based on the data analysis requirement information to obtain a data analysis instruction template that meets the data analysis requirement information, thereby generating corresponding data analysis instructions based on the data analysis instruction template for data analysis processing.

[0040] Step 206, obtaining a target data table model that matches the target indicator model, wherein the target data table model includes target analysis data that meets the data analysis indicator.

[0041] Among them, the target data table model is a data table model determined from various pre-configured data table models and matched with the target indicator model. The data table model may include target analysis data that needs to be processed for data analysis. The target analysis data satisfies the data analysis indicators in the target indicator model, so that data analysis is performed on the target analysis data in the target data table model to meet the data analysis demand information.

[0042] Specifically, the server can determine a pre-configured data table model, and match the target indicator model with each data table model, thereby determining a target data table model that matches the target indicator model from each data table model. The target data table model includes target analysis data that needs to be analyzed and meets the data analysis indicators. In some embodiments, the server can pre-process the business data in the business system in advance to process the business data into a data table model. The business data in the data table model can be structured data, such as a single header and a single body structure, thereby reducing the difficulty of understanding various business data. In some embodiments, the server can filter the business data in the original data table of the business system to obtain the business data in the original data table that needs to be analyzed. The server can build a corresponding data table model based on the filtered business data, thereby storing the filtered business data through the standardized structure of the data table model.

[0043] Step 208: Perform data analysis based on the target analysis data and the data analysis instruction template to obtain a data analysis result corresponding to the data analysis request message.

[0044] Among them, the data analysis result is the processing result obtained by performing data analysis processing according to the data analysis request message. Exemplarily, the server can obtain the target analysis data from the target data table model, and perform data analysis based on the target analysis data and the data analysis instruction template, so as to obtain the data analysis result corresponding to the data analysis request message. In some embodiments, the server can perform data analysis on the target analysis data according to the data analysis instruction template to obtain the corresponding data analysis result. For example, the data analysis instruction template may carry a data analysis instruction, and the server can perform data analysis on the target analysis data according to the data analysis instruction carried in the data analysis instruction template to obtain the corresponding data analysis result. For example, the server can input the target analysis data and the data analysis instruction template into a large language model (LLM), so as to construct a data analysis instruction based on the data analysis instruction template through the large language model, perform data analysis on the target analysis data, and output the corresponding data analysis result. In some embodiments, the data analysis instruction template can also define the output format of the data analysis result, such as the data analysis result can be output in a natural language format, so that the data analysis result described in natural language can be obtained. In some embodiments, the data analysis instruction template may further define the output format of the data analysis results as a visual display method, such as a table, chart, etc., so that the data analysis results may be output in a table or chart form.

[0045] In the above data analysis method, based on the data analysis requirement information determined according to the data analysis request message, a target indicator model including data analysis indicators and a data analysis instruction template corresponding to the target indicator model are obtained, a target data table model matching the target indicator model is obtained, and data analysis is performed based on the target analysis data and the data analysis instruction template included in the target data table model to obtain a data analysis result corresponding to the data analysis request message. In the data analysis process, the target indicator model and the corresponding data analysis instruction template are obtained through the data analysis requirement information, and a matching target data table model is obtained based on the target indicator model, so that data analysis is performed through the target analysis data and the data analysis instruction template that meet the data analysis indicators included in the target data table model, and various data can be standardized through the target data table model, so that data analysis processing can be performed on the standardized data, which reduces the difficulty of data analysis and thus improves the accuracy of data analysis.

[0046] In an exemplary embodiment, based on data analysis requirement information, a target indicator model including data analysis indicators and a data analysis instruction template corresponding to the target indicator model are obtained, including: obtaining a pre-built indicator model library, the indicator model library including at least one indicator model, and the indicator model including at least one data analysis indicator; matching the data analysis requirement information with each indicator model in the indicator model library respectively, and determining the target indicator model from the indicator model library according to the matching results; and obtaining the data analysis instruction template corresponding to the target indicator model.

[0047] The indicator model library is pre-built according to actual needs, and the indicator model library may include at least one pre-configured indicator model, and each indicator model may include at least one data analysis indicator. By configuring different data analysis indicators, different indicator models can be built to cope with different data analysis demand information.

[0048] Optionally, the server may query a pre-built indicator model library, the indicator model library includes a pre-built indicator model, and the indicator model includes a corresponding data analysis indicator. In some embodiments, the indicator model may include a component, and the component is used to store different data analysis indicators. By configuring the component, different indicator models can be constructed. The server can match the data analysis demand information with each indicator model in the indicator model library respectively, such as the server can perform text matching on the keywords in the data analysis demand information and the keywords of the data analysis indicators in each indicator model, respectively, to obtain corresponding matching results. The server can determine the target indicator model that matches the data analysis demand information from the indicator model library based on each matching result. For example, the matching result includes the keyword similarity between the keyword of the data analysis demand information and the keyword of the data analysis indicator in the indicator model. The server can determine the indicator model that matches the data analysis demand information based on the keyword similarity, such as determining the indicator model corresponding to the maximum keyword similarity value as the target indicator model that matches the data analysis demand information. For the target indicator model, the server can obtain the corresponding data analysis instruction template, such as the server can query the corresponding data analysis instruction template based on the indicator model identifier of the target indicator model. The corresponding relationship between the data analysis instruction template and the indicator model identifier can be set in advance.

[0049] In this embodiment, the server matches the data analysis demand information with each indicator model in the indicator model library respectively, determines the target indicator model according to the matching results, and obtains the data analysis instruction template corresponding to the target indicator model, so as to obtain the target analysis data based on the target indicator model and then perform data analysis in combination with the data analysis instruction template. Data analysis can be performed using standardized target analysis data, which reduces the difficulty of data analysis and thus improves the accuracy of data analysis.

[0050] In an exemplary embodiment, the indicator model includes a core component and at least one enhanced component, the core component and the enhanced component include data analysis indicators; the data analysis requirement information includes analysis intent categories and annotation content; the data analysis requirement information is matched with each indicator model in the indicator model library, and the target indicator model is determined from the indicator model library based on the matching results, including: matching the annotation content with the model labels of each indicator model in the indicator model library to obtain label matching results; in the indicator model library, the label matching results are represented as matching indicator models and determined as initial indicator models; based on the analysis intent category, the target enhancement component is determined from the enhancement components included in the initial indicator model; based on the core components and target enhancement components of the initial indicator model, the target indicator model is obtained.

[0051] Among them, the indicator model includes core components and enhanced components, the number of enhanced components can be one or more, and the core components and enhanced components each include data analysis indicators. The core component can be a necessary component of the indicator model, and the data analysis indicators included in the core component can be used as the necessary data analysis indicators of the indicator model; and the enhanced component can be an additional component of the indicator model, and the data analysis indicators included in the enhanced component can be used as the additional data analysis indicators of the indicator model. The data analysis indicators in each indicator model can include necessary data analysis indicators and additional data analysis indicators. The analysis intention category is used to characterize the demand category of data analysis, such as including but not limited to at least one of various categories including current situation analysis, cause analysis, early warning analysis, and predictive analysis; the annotation content can be the content obtained by annotation based on the data analysis request message, such as including but not limited to at least one of various annotation contents including time, scope, object, association, cause, etc.

[0052] The model label is a label configured for the indicator model and is used for indicator model matching. Specifically, the indicator model can be screened by matching the model label with the annotation content in the data analysis requirement information. The label matching result is the matching result of the model label of each indicator model in the indicator model library and the annotation content in the data analysis requirement information. The initial indicator model is an indicator model screened from the indicator model library based on the annotation content. The target enhancement component is an enhancement component determined from the various enhancement components of the initial indicator model that matches the analysis intent category. The target indicator model is an indicator model determined from the indicator model library that matches the data analysis requirement information, which can be obtained by combining the core components of the initial indicator model and the target enhancement components.

[0053] Exemplarily, each indicator model in the indicator model library includes necessary core components and optional enhanced components. The server can determine the analysis intention category and the annotation content based on the data analysis request message, wherein the analysis intention category can be obtained by the server performing intent analysis on the data analysis request message, such as performing intent analysis on the data analysis request message through a pre-trained intent classification model, thereby obtaining the analysis intention category; for the annotation content, the server can perform word segmentation based on the data analysis request message, and annotate according to the word segmentation result, thereby obtaining the corresponding annotation content, and the server can obtain data analysis demand information based on the analysis intention category and the annotation content. The server can match the annotated content with each indicator model in the indicator model library respectively, such as matching the annotated content with the model label of each indicator model, thereby obtaining a label matching result. The server can determine the indicator model that matches the annotated content from the indicator model library based on the label matching result, and use the matched indicator model as the initial indicator model. The server can screen the enhanced components included in the initial indicator model based on the analysis intention category to determine the target enhanced component corresponding to the analysis intention category from each enhanced component. The server can combine the core component and the target enhanced component of the initial indicator model to obtain the target indicator model.

[0054] In this embodiment, the server determines the initial indicator model from the indicator model library based on the annotation content in the data analysis requirement information and the model labels of each indicator model, determines the target enhancement component from the enhancement components included in the initial indicator model based on the analysis intent category in the data analysis requirement information, and obtains the target indicator model based on the core components and the target enhancement components of the initial indicator model, so that the corresponding target indicator model can be configured based on the analysis intent category and the annotation content to ensure the targetedness of the target indicator model, so that after obtaining the target analysis data based on the target indicator model, data analysis can be performed in combination with the data analysis instruction template, and data analysis processing can be performed using standardized target analysis data, which reduces the difficulty of data analysis and improves the accuracy of data analysis.

[0055] In an exemplary embodiment, the data analysis method also includes: obtaining a pre-built data table model library, the data table model library includes at least one data table model, and the data table model includes structured data; obtaining at least one data analysis indicator, and configuring each data table model in the data table model library according to the data analysis indicator to obtain an indicator model that meets different data analysis indicators.

[0056] The data table model library is pre-built according to actual needs, and the data table model library may include at least one data table model. Each data table model may include structured data, and the structured data is obtained by structuring the business data. Different data table models can be built by configuring different structured data. For example, the data table model may be a single header and single body structure, so that the corresponding business data is recorded through the single header and single body structure.

[0057] Exemplarily, the server can query a pre-built data table model library and obtain data analysis indicators, and configure the data table models in the data table model library according to the data analysis indicators. For example, corresponding structured data can be added to the data table model based on the needs of the data analysis indicators, thereby obtaining different indicator models, and different indicator models can include different data analysis indicators. In some embodiments, the indicator model can include core components and enhanced components. When the data table model is output, it must meet the structural requirements of the core components, that is, it needs to be configured for the data table model according to the core component structure "indicator name, statistical period, analysis dimension, and filter condition"; the enhanced component can include year-on-year, month-on-month, target, attribution and other data analysis indicators, and the corresponding structured data needs to be configured in the data table model.

[0058] In this embodiment, the server configures the data table model in the data table model library based on the data analysis indicators to build an indicator model including different data analysis indicators. Therefore, after determining the target indicator model, the server can obtain the target analysis data based on the target indicator model and then perform data analysis in combination with the data analysis instruction template. Data analysis processing can be performed using standardized target analysis data, which reduces the difficulty of data analysis and thus improves the accuracy of data analysis.

[0059] In an exemplary embodiment, obtaining a target data table model that matches a target indicator model includes: obtaining a pre-built data table model library, the data table model library including at least one data table model, the data table model including structured data; and matching the target data table model from the data table model library according to the target indicator model.

[0060] Among them, the data table model library may include at least one data table model, each data table model may include structured data, and the structured data is obtained by structuring the business data. By configuring different structured data, different data table models can be constructed. Specifically, the server can match the data table model library based on the target indicator model, such as matching the target indicator model with each data table model in the data table model library, so as to obtain a matching target data table model. In some embodiments, the server can match the target indicator model with the data analysis indicators included in each data table model to obtain a matching result, and the server can determine the data table model corresponding to the matching result that matches the same as the target data table model that matches the target indicator model.

[0061] In this embodiment, the server matches the target data table model from the data table model library based on the target indicator model, so as to obtain the target analysis data based on the target indicator model and then perform data analysis in combination with the data analysis instruction template. Data analysis processing can be performed through standardized target analysis data, which reduces the difficulty of data analysis and thus improves the accuracy of data analysis.

[0062] In an exemplary embodiment, the data analysis method further includes: acquiring various data tables, and performing structured configuration on the business data included in the data tables to obtain at least one data table model.

[0063] The data table may be a data structure for storing business data in a database. Specifically, the server may obtain various data tables including business data, and perform structured configuration on the business data included in the data table, such as configuring the business data in the data table as a data structure of a single header and a single body, thereby obtaining a corresponding data table model. According to different structured configuration methods, different data table models may be obtained based on each business data configuration in the data table.

[0064] In this embodiment, the server can perform structured configuration on the business data in the data table to construct a data table model, so that the data table model can be used to perform standardized processing on the business data so that data analysis can be performed based on the standardized data, thereby reducing the difficulty of data analysis and improving the accuracy of data analysis.

[0065] In an exemplary embodiment, Figure 3 As shown, the process of determining the data analysis requirement information, that is, determining the corresponding data analysis requirement information based on the data analysis request message, includes:

[0066] Step 302: perform intent classification according to the data analysis request message to obtain the analysis intent category corresponding to the data analysis request message.

[0067] Among them, the analysis intent category is used to characterize the demand category of data analysis, which is obtained by performing intent classification processing on the data analysis request message. For example, the server can perform intent classification based on the data analysis request message, such as performing intent analysis on the data analysis request message through a pre-trained intent classification model, thereby obtaining the analysis intent category corresponding to the data analysis request message.

[0068] Step 304: perform word segmentation on the data analysis request message to obtain a word segmentation result.

[0069] Optionally, the server may perform word segmentation based on the data analysis request message to divide the data analysis request message into different fields to obtain corresponding word segmentation results. For example, for the data analysis request message "How is the revenue of sales department one this month", after word segmentation, the word segmentation results obtained may include "this month", "sales department one", "revenue", and "how".

[0070] Step 306: annotate the word segmentation result to obtain the annotated content corresponding to the word segmentation result.

[0071] Specifically, the server may perform annotation based on the word segmentation result, and may specifically perform annotation on each field in the word segmentation result to obtain corresponding annotation content. For example, the obtained annotation content may include "time", "scope", "object" and "reason".

[0072] Step 308: Based on the analysis intent category and the annotation content, obtain the data analysis requirement information corresponding to the data analysis request message.

[0073] Exemplarily, the server may obtain data analysis requirement information based on the analysis intention category and the annotation content. For example, the server may combine the analysis intention category and the annotation content to obtain the data analysis requirement information corresponding to the data analysis request message.

[0074] In this embodiment, the server performs intent classification on the data analysis request message to obtain an analysis intent category, annotates the word segmentation result of the data analysis request message to determine the annotated content, and comprehensively analyzes the intent category and the annotated content to obtain data analysis requirement information, so as to determine the target indicator model and data analysis instruction template based on the data analysis requirement information to perform data analysis, thereby ensuring the accuracy of the data analysis.

[0075] In an exemplary embodiment, data analysis is performed based on target analysis data and a data analysis instruction template to obtain a data analysis result corresponding to a data analysis request message, including: generating model query information based on the target analysis data and the data analysis instruction template; inputting the model query information into a pre-trained large language model for data analysis to obtain a data analysis result output by the large language model and corresponding to the data analysis request message.

[0076] The model question information can be used as input of the large language model, and is specifically generated according to the target analysis data and the data analysis instruction template. Optionally, the server can generate the model question information based on the target analysis data and the data analysis instruction template, and input the model question information into the pre-trained large language model, so that the large language model performs data analysis on the model question information, and the large language model outputs the corresponding data analysis results.

[0077] In this embodiment, the server generates model question information based on the target analysis data and the data analysis instruction template, and performs data analysis on the model question information through a large language model. The large language model can be used to perform data analysis and processing on standardized data, which reduces the difficulty of data analysis and thus improves the accuracy of data analysis.

[0078] The present application also provides an application scenario, which applies the above data analysis method. Specifically, the application of the data analysis method in the application scenario is as follows:

[0079] This application scenario involves an LLM question scenario for form data, in which the data structures of different forms are differentiated. After identifying the user's intention, traditional methods often go directly to the data form to obtain the data results and return relevant content. The large model also needs to be adjusted one by one for different data conditions through fine-tuning methods such as prompt words. However, the direct matching and acquisition of data tables for different differentiated forms also have very different effects, and the data returned for complex forms is very poor. Moreover, the large model is not good at data analysis and statistics, and it is difficult to directly obtain data results based on existing data forms. In addition, under the traditional solution, only limited instruction templates of the large model can be applied, and the scope of application is limited, making it difficult to apply to a wider range of scenarios. In addition, there will be data tables and data structures of various formats in practical application scenarios. The existing solution can only perform shallow interpretation and analysis, and cannot support a wider range of LLM data analysis scenarios and form deeper insight analysis, resulting in limited accuracy of data analysis. Such as Figure 4As shown in the figure, in actual application scenarios, the business data that needs to be analyzed may include various data structures such as numerical data, pie charts, and bar charts. When performing data analysis based on LLM, it includes direct reading of business data and data interpretation. Among them, the LLM model is an artificial intelligence model that can process and generate natural language. It usually has a large number of parameters and powerful computing power. It can learn and understand a large amount of text data, and can generate natural and fluent text to answer questions, conduct conversations, and complete various language tasks.

[0080] In the data analysis method provided in the embodiment of the present application, accurate data analysis and processing is performed based on LLM, which can output in-depth and intelligent analysis for different types of data structure tables and can be reused in a wider range of data analysis scenarios. The data analysis method provided in the embodiment of the present application is designed and created for the LLM data analysis method of differentiated form data, improves the user intention recognition and acquisition of relevant data links, and adds a data processing center link. Through the design and creation of a comprehensive solution, it can be widely adapted to LLM data analysis scenarios, and the results are more accurate.

[0081] Specifically, Figure 5 As shown, the server can receive user questions, which can be used as data analysis request messages. The server can understand the user questions and identify the user's intentions based on the user questions, thereby obtaining the corresponding data analysis demand information. The server can match in the data processing center based on the data analysis demand information to determine the target indicator model including the data analysis indicators from the various indicator models of the data processing center, and further obtain the data analysis instruction template corresponding to the target indicator model. The server can obtain a target data table model that matches the target indicator model, and input the target analysis data and the data analysis instruction template in the target data table model into the large model for data analysis to produce a reply result fed back to the user, that is, to obtain the data analysis result. Among them, the server can pre-process the business data in the database through the computing service in advance to construct the indicator model and data table model of the data processing center, and pre-train the corresponding data analysis instruction template for the large model.

[0082] Specifically, the data analysis method provided in the embodiment of the present application adds a data processing center link, through which differentiated data forms can be applied and matched; and based on the indicator model of the data processing center, the user intention recognition link, matching and data acquisition link are transformed. Among them, the data processing center includes three parts: indicator model, data table model and computing service; the data processing center is connected with various differentiated forms of the underlying database. Its main function is to process and organize differentiated data tables to output them according to the assemblable structure of the indicator model, which can respond to various types of questions and intentions, that is, corresponding to various different data analysis demand information.

[0083] Furthermore, if Figure 6 As shown in the figure, in the data processing center, firstly, pre-processing is performed on various differentiated data tables in the database to obtain a data table model. The data table model is a single header and single body structure. The design of the data table model can effectively reduce the difficulty of understanding the large model. The pre-processing process is to extract the analytical data of the original data table, and process and count it through the computing service, and store it according to the data table model structure. The non-analytical data in the original data table is not retained, that is, it is not stored in the data table model. Furthermore, the data table model is processed again by the indicator model and output according to the indicator model structure. The indicator model includes core components and enhanced components; when the data table model is output, it must meet the structural requirements of the core components, and it needs to be output according to the core component structure "indicator name, statistical period, analysis dimension, and filtering condition"; when other data content is output, it is output according to the enhanced components, which include year-on-year, quarter-on-quarter, target, attribution, etc., so that the indicator model has scalability. The enhanced components depend on the core components. On the basis of the core components, the enhanced components can be assembled, that is, the core components + enhanced components (which can be assembled), to match different user intentions to obtain target analysis data for data analysis and processing. After the indicator model is built, the business data in the data table is pre-processed and pre-trained with a large model to form a data analysis instruction template. After the corresponding data analysis requirement information is determined based on the data analysis request message, such as obtaining the user's question to identify the user's intention, the corresponding data analysis instruction template is matched according to the intention classification to quickly realize high-quality organized natural language questions, thereby realizing fast and accurate data analysis and processing.

[0084] In the user intention recognition process, such as Figure 7As shown, the intent classification link is added according to the structure of the data template, including but not limited to current situation analysis, cause analysis, early warning analysis, and predictive analysis. When different user intent classifications are identified, the core components of the indicator template + the assembleable enhanced components are used to match and obtain data, so that it can adapt to and support a wider range of question scenarios. For example, when the user intent is cause analysis, the core components of the indicator model + the "attribution" enhanced components are obtained and called, so that it can be applied to a wider range of user intent recognition to support richer question scenarios. In addition, after obtaining the data analysis request message, that is, after obtaining the user question, the user question can be segmented and then annotated, and the annotated content is matched with the components of the indicator model, so that the annotated content is more effective and data can be obtained more accurately. In the matching & acquisition data processing, the incoming annotations and intent classifications are judged at the same time, and matched with the indicator library and indicator model of the data processing center. If matched, the corresponding target analysis data is obtained, otherwise the target analysis data is not obtained; when the user intent recognition link identifies the need for computing services, the corresponding data components and computing services are called, and after data processing, the data is returned to the large model for data analysis and processing by the large model.

[0085] In the data analysis method provided in the embodiment of the present application, after receiving the user question, the server understands the user question, supplements the user question description or disassembles the user question; the server then identifies the user intent, specifically according to the indicator model core component + the assembleable enhancement component solution to respond to a wide range of user intents, so that the user question can more accurately match the preset data template. Furthermore, the server will match the indicator library and data template of the data processing center according to the incoming user intent, annotation and intent classification, and directly return the target analysis data if it matches; when the intent needs to process data, the corresponding components and technical services of the data template are called to process and return the corresponding target analysis data. If none of the above can be matched, a result of no data is returned. After the large model obtains the corresponding target analysis data, it combines the data analysis instruction template to perform data analysis to obtain the corresponding data analysis results, and specifically obtains the corresponding data conclusions.

[0086] In traditional data analysis and processing, it is often impossible to accurately obtain data results for questions in different forms; often, only the analysis of the current data status can be done. The data analysis method provided in the embodiment of the present application standardizes the data through the design of the data processing center "the data table is pre-processed into a data table model, and then processed into an indicator model". The design of the indicator model "core component + assemblable enhancement component" enhances the data combination capability, which can respond to different user data analysis intentions, thereby achieving high-quality and more extensive intelligent data analysis of differentiated data tables. Furthermore, the data analysis method provided in the embodiment of the present application optimizes and improves the components of the data template, so that the data content of the data processing center can be more accurately identified and matched. In addition, the data analysis method provided in the embodiment of the present application combines user intention classification and data template components to preset different types of data analysis instruction templates, so that it can adapt to a wider range of scenarios.

[0087] The data analysis method provided in the embodiment of the present application makes LLM-based data analysis more accurate and data acquisition faster through improvements in the design of the data processing center and user intent recognition. Moreover, the data processing center provides the design of data table models and indicator models, so that differentiated data tables can be more standardized, reducing the difficulty of understanding large models. The indicator model provides a "core component + assembleable enhancement component" design, which can be widely adapted to different user intentions, making the application scenarios of this technical solution wider and easier to promote. In addition, Figure 8 As shown, in the traditional data analysis processing results, LLM can only perform current status analysis on the data appearance; while the data analysis method provided in the embodiment of the present application can achieve insights into data anomalies, analysis of the causes of data problems, data growth trend analysis, etc. through LLM, making the data analysis based on LLM more accurate and intelligent.

[0088] It should be understood that, although the various steps in the flowcharts involved in the above-mentioned embodiments are displayed in sequence according to the indication of the arrows, these steps are not necessarily executed in sequence according to the order indicated by the arrows. Unless there is a clear explanation in this article, the execution of these steps does not have a strict order restriction, and these steps can be executed in other orders. Moreover, at least a part of the steps in the flowcharts involved in the above-mentioned embodiments can include multiple steps or multiple stages, and these steps or stages are not necessarily executed at the same time, but can be executed at different times, and the execution order of these steps or stages is not necessarily to be carried out in sequence, but can be executed in turn or alternately with other steps or at least a part of the steps or stages in other steps.

[0089] Based on the same inventive concept, the embodiment of the present application also provides a data analysis device for implementing the data analysis method involved above. The implementation scheme for solving the problem provided by the device is similar to the implementation scheme recorded in the above method, so the specific limitations in one or more data analysis device embodiments provided below can refer to the limitations on the data analysis method above, and will not be repeated here.

[0090] In an exemplary embodiment, Fig. 9 As shown, a data analysis device 900 is provided, including: a demand determination module 902, an indicator model acquisition module 904, a data table model acquisition module 906 and an analysis result acquisition module 908, wherein:

[0091] The demand determination module 902 is used to obtain a data analysis request message and determine corresponding data analysis demand information based on the data analysis request message;

[0092] The indicator model acquisition module 904 is used to acquire a target indicator model including data analysis indicators and a data analysis instruction template corresponding to the target indicator model based on the data analysis requirement information;

[0093] A data table model obtaining module 906 is used to obtain a target data table model that matches the target indicator model, wherein the target data table model includes target analysis data that meets the data analysis indicator;

[0094] The analysis result obtaining module 908 is used to perform data analysis based on the target analysis data and the data analysis instruction template to obtain the data analysis result corresponding to the data analysis request message.

[0095] In an exemplary embodiment, the indicator model acquisition module 904 is also used to obtain a pre-built indicator model library, the indicator model library includes at least one indicator model, and the indicator model includes at least one data analysis indicator; the data analysis requirement information is matched with each indicator model in the indicator model library, and the target indicator model is determined from the indicator model library according to the matching results; and the data analysis instruction template corresponding to the target indicator model is obtained.

[0096] In an exemplary embodiment, the indicator model includes a core component and at least one enhanced component, and the core component and the enhanced component include data analysis indicators; the data analysis requirement information includes analysis intent categories and annotation content; the indicator model acquisition module 904 is also used to match the annotation content with the model labels of each indicator model in the indicator model library to obtain a label matching result; in the indicator model library, the label matching result is represented as a matching indicator model and determined as an initial indicator model; based on the analysis intent category, the target enhancement component is determined from the enhancement components included in the initial indicator model; based on the core component and the target enhancement component of the initial indicator model, the target indicator model is obtained.

[0097] In an exemplary embodiment, it also includes an indicator model configuration module, which is used to obtain a pre-built data table model library, the data table model library includes at least one data table model, and the data table model includes structured data; obtain at least one data analysis indicator, and configure each data table model in the data table model library according to the data analysis indicator to obtain an indicator model including different data analysis indicators.

[0098] In an exemplary embodiment, the data table model acquisition module 906 is also used to obtain a pre-built data table model library, the data table model library includes at least one data table model, and the data table model includes structured data; according to the target indicator model, the target data table model is matched from the data table model library.

[0099] In an exemplary embodiment, it also includes a structured configuration module for acquiring various data tables and performing structured configuration on the business data included in the data tables to obtain at least one data table model.

[0100] In an exemplary embodiment, the demand determination module 902 is also used to perform intent classification according to the data analysis request message to obtain the analysis intent category corresponding to the data analysis request message; perform word segmentation on the data analysis request message to obtain the word segmentation result; perform annotation on the word segmentation result to obtain the annotation content corresponding to the word segmentation result; and obtain the data analysis demand information corresponding to the data analysis request message based on the analysis intent category and the annotation content.

[0101] In an exemplary embodiment, the analysis result acquisition module 908 is also used to generate model question information based on the target analysis data and the data analysis instruction template; the model question information is input into the pre-trained large language model for data analysis, and the data analysis result corresponding to the data analysis request message output by the large language model is obtained.

[0102] Each module in the above data analysis device can be implemented in whole or in part by software, hardware, or a combination thereof. Each module can be embedded in or independent of a processor in a computer device in the form of hardware, or can be stored in a memory in a computer device in the form of software, so that the processor can call and execute operations corresponding to each module.

[0103] In an exemplary embodiment, a computer device is provided. The computer device may be a server, and its internal structure diagram may be as shown in FIG. Fig.10 As shown. The computer device includes a processor, a memory, an input / output interface (Input / Output, referred to as I / O) and a communication interface. Among them, the processor, the memory and the input / output interface are connected through a system bus, and the communication interface is connected to the system bus through the input / output interface. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store various data involved in the data analysis method. The input / output interface of the computer device is used to exchange information between the processor and an external device. The communication interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, a data analysis method is implemented.

[0104] In one embodiment, a computer device is provided. The computer device may be a terminal, and its internal structure diagram may be as follows: Fig.11As shown. The computer device includes a processor, a memory, an input / output interface, a communication interface, a display unit and an input device. The processor, the memory and the input / output interface are connected through a system bus, and the communication interface, the display unit and the input device are connected to the system bus through the input / output interface. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The input / output interface of the computer device is used to exchange information between the processor and an external device. The communication interface of the computer device is used to communicate with an external terminal in a wired or wireless manner, and the wireless manner can be implemented through WIFI, a mobile cellular network, NFC (near field communication) or other technologies. When the computer program is executed by the processor, a data analysis method is implemented. The display unit of the computer device is used to form a visually visible image, and can be a display screen, a projection device or a virtual reality imaging device. The display screen can be a liquid crystal display screen or an electronic ink display screen. The input device of the computer device can be a touch layer covered on the display screen, or a button, trackball or touchpad set on the computer device casing, or an external keyboard, touchpad or mouse, etc.

[0105] Those skilled in the art will understand that Fig.10 or Fig.11 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0106] In one embodiment, a computer device is further provided, including a memory and a processor, wherein a computer program is stored in the memory, and the processor implements the steps in the above method embodiments when executing the computer program.

[0107] In one embodiment, a computer-readable storage medium is provided, storing a computer program, which implements the steps in the above method embodiments when executed by a processor.

[0108] In one embodiment, a computer program product is provided, including a computer program, which implements the steps in the above method embodiments when executed by a processor.

[0109] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with relevant regulations.

[0110] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to the memory, database or other medium used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetoresistive random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. As an illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM). The database involved in each embodiment provided in this application may include at least one of a relational database and a non-relational database. Non-relational databases may include distributed databases based on blockchains, etc., but are not limited to this. The processor involved in each embodiment provided in this application may be a general-purpose processor, a central processing unit, a graphics processor, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, etc., but are not limited to this.

[0111] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0112] The above-described embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be construed as limiting the scope of the present application. It should be noted that, for a person of ordinary skill in the art, several modifications and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the attached claims.

Claims

1. A data analysis method, characterized in that: The method comprises: Obtaining a data analysis request message, and determining corresponding data analysis requirement information based on the data analysis request message; Based on the data analysis requirement information, obtaining a target indicator model including data analysis indicators and a data analysis instruction template corresponding to the target indicator model; Obtaining a target data table model that matches the target indicator model, wherein the target data table model includes target analysis data that meets the data analysis indicator; Data analysis is performed based on the target analysis data and the data analysis instruction template to obtain a data analysis result corresponding to the data analysis request message.

2. The method according to claim 1, characterized in that The step of acquiring a target indicator model including data analysis indicators and a data analysis instruction template corresponding to the target indicator model based on the data analysis requirement information includes: Acquire a pre-built indicator model library, wherein the indicator model library includes at least one indicator model, and the indicator model includes at least one data analysis indicator; Matching the data analysis requirement information with each indicator model in the indicator model library respectively, and determining a target indicator model from the indicator model library according to the obtained matching results; Obtain the data analysis instruction template corresponding to the target indicator model.

3. The method according to claim 2, characterized in that The indicator model includes a core component and at least one enhanced component, wherein the core component and the enhanced component include data analysis indicators; the data analysis requirement information includes analysis intent categories and annotation content; The step of matching the data analysis requirement information with each indicator model in the indicator model library respectively, and determining a target indicator model from the indicator model library according to the obtained matching results, comprises: Matching the annotation content with the model labels of each indicator model in the indicator model library to obtain a label matching result; In the indicator model library, the label matching result is represented as a matched indicator model and determined as an initial indicator model; Based on the analysis intention category, determining a target enhancement component from the enhancement components included in the initial indicator model; A target indicator model is obtained according to the core components of the initial indicator model and the target enhancement components.

4. The method according to claim 2, characterized in that: The method further comprises: Acquire a pre-built data table model library, wherein the data table model library includes at least one data table model, and the data table model includes structured data; At least one data analysis indicator is obtained, and each of the data table models in the data table model library is configured according to the data analysis indicator to obtain an indicator model including different data analysis indicators.

5. The method according to claim 1, characterized in that The step of obtaining a target data table model that matches the target indicator model includes: Acquire a pre-built data table model library, wherein the data table model library includes at least one data table model, and the data table model includes structured data; According to the target indicator model, a target data table model is matched from the data table model library.

6. The method according to claim 5, characterized in that The method further comprises: Various data tables are obtained, and business data included in the data tables are structured to obtain at least one data table model.

7. The method according to claim 1, characterized in that The determining corresponding data analysis requirement information based on the data analysis request message includes: Performing intent classification according to the data analysis request message to obtain an analysis intent category corresponding to the data analysis request message; Performing word segmentation on the data analysis request message to obtain a word segmentation result; Annotate the word segmentation result to obtain the annotation content corresponding to the word segmentation result; Based on the analysis intention category and the marked content, data analysis requirement information corresponding to the data analysis request message is obtained.

8. The method according to any one of claims 1 to 7, characterized in that: The performing data analysis based on the target analysis data and the data analysis instruction template to obtain a data analysis result corresponding to the data analysis request message includes: Generate model question information based on the target analysis data and the data analysis instruction template; The model question information is input into a pre-trained large language model for data analysis, and a data analysis result corresponding to the data analysis request message output by the large language model is obtained.

9. A data analysis device, characterized in that: The device comprises: A demand determination module, used to obtain a data analysis request message and determine corresponding data analysis demand information based on the data analysis request message; An indicator model acquisition module, used to acquire a target indicator model including data analysis indicators and a data analysis instruction template corresponding to the target indicator model based on the data analysis requirement information; A data table model acquisition module, used to obtain a target data table model that matches the target indicator model, wherein the target data table model includes target analysis data that meets the data analysis indicator; The analysis result obtaining module is used to perform data analysis based on the target analysis data and the data analysis instruction template to obtain the data analysis result corresponding to the data analysis request message.

10. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 8 are implemented.

11. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 8 are implemented.

12. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 8 are implemented.

Images