Large language model data analysis method and device, computer equipment and storage medium

Abstract

The invention relates to a big language model data analysis method and device, computer equipment, a computer readable storage medium and a computer program product, and is applied to an enterprise resource planning system (ERP) and the like. The method comprises the following steps: acquiring to-be-analyzed question number information, and matching a target data index in a preset data index set according to the to-be-analyzed question number information; matching a corresponding target data entity according to the to-be-analyzed question number information under the condition that the matching of the target data index fails; under the condition that the target data entity is successfully matched, acquiring a data index template, and determining a target data index according to the target data entity and the data index template; and obtaining target data according to the target data index, and obtaining a data analysis result according to the target data. By adopting the method, the intelligent number asking accuracy of the complex form can be improved.

Description

Technical Field

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

[0002] LLM (Large Language Model) is a large-scale natural language processing model trained based on deep learning algorithms. It is trained on massive amounts of text data to understand and generate natural language, and has multiple capabilities such as language understanding, reasoning, and translation. In the field of data processing and analysis, it is often used in scenarios of natural language understanding and natural language generation.

[0003] In the related technical solution, the method of applying the large language model to ask questions is: after obtaining the user's question, the large language model is used to convert the natural language sentence into SQL (Structured Query Language) to obtain data from the data table.

[0004] The advantages of the solutions in related technologies are that they can be quickly built for simple data query scenarios, and the expressions are flexible and easy to understand. However, there are also defects, especially in the ERP (Enterprise Resource Planning) intelligent data query scenario. The ERP system has a large number of complex forms. Directly applying this technology will result in poor recognition accuracy and cannot handle complex data analysis scenarios. Summary of the invention

[0005] Based on this, it is necessary to provide a large language model data analysis method, device, computer equipment, computer-readable storage medium and computer program product that can improve the accuracy of intelligent questions in complex forms in response to the above technical problems.

[0006] In a first aspect, the present application provides a large language model data analysis method, comprising:

[0007] Acquire information of question data to be analyzed, and match target data indicators in a preset data indicator set according to the information of question data to be analyzed;

[0008] In the case where the target data index fails to match, matching the corresponding target data entity according to the question information to be analyzed;

[0009] In the case where the target data entity is matched successfully, a data indicator template is obtained, and a target data indicator is determined according to the target data entity and the data indicator template;

[0010] Target data is acquired according to the target data indicator, and data analysis results are obtained according to the target data.

[0011] In one embodiment, determining the target data indicator according to the target data entity and the data indicator template includes:

[0012] Extracting characteristic fields of the target data entity;

[0013] Generate basic data indicators and enhanced data indicators matching the characteristic fields according to the data indicator template;

[0014] The basic data indicators and the enhanced data indicators are collected to obtain target data indicators.

[0015] In one embodiment, the target data includes basic target data and enhanced target data; and obtaining the target data according to the target data indicator includes:

[0016] Acquire the basic target data according to the basic data indicators; wherein the basic data indicators include any one or more of indicator name, analysis dimension, statistical period, and filtering condition;

[0017] The enhanced target data is obtained by calculation according to the enhanced data indicator and the basic target data.

[0018] In one of the embodiments, when the target data indicator fails to match, after matching the corresponding target data entity according to the question information to be analyzed, the method further includes:

[0019] In the case where the target data entity fails to match, a pre-built analysis table is obtained according to the question information to be analyzed; wherein the pre-built analysis table is pre-built according to the data entity and data model related to the question information to be analyzed;

[0020] Target data is acquired according to the pre-built analysis table, and a data analysis result is obtained according to the target data.

[0021] In one embodiment, before obtaining the information of the question to be analyzed, the method further includes:

[0022] In response to the data analysis instruction, initial question number information is acquired, and target question number information belonging to a preset target field is extracted from the initial question number information;

[0023] Inputting the target question number information into a semantic analysis model to perform question number intent recognition to obtain an intent recognition result;

[0024] The obtaining of the question information to be analyzed includes: obtaining the question information to be analyzed according to the intention recognition result.

[0025] In one embodiment, the method further comprises:

[0026] Inputting the target question information into a semantic analysis model for supplementation and / or splitting;

[0027] Matching an approximate question template according to the target question number information after supplementation and / or splitting;

[0028] In the case where the approximate question template is matched successfully, using the data indicator in the approximate question template as the target data indicator, and performing the step of acquiring the target data according to the target data indicator;

[0029] In the case where the approximate question template matching fails, the step of obtaining the question information to be analyzed according to the intention recognition result is performed.

[0030] In a second aspect, the present application also provides a large language model data analysis device, comprising:

[0031] An information acquisition module is used to acquire the data information to be analyzed, and match the target data indicator in the preset data indicator set according to the data information to be analyzed;

[0032] A data matching module, used for matching the corresponding target data entity according to the data information to be analyzed when the target data indicator fails to match;

[0033] An indicator matching module, used to obtain a data indicator template when the target data entity is successfully matched, and determine a target data indicator according to the target data entity and the data indicator template;

[0034] The data analysis module is used to obtain target data according to the target data indicator and obtain data analysis results according to the target data.

[0035] In a third aspect, the present application further provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the following steps are implemented:

[0036] Acquire information of question data to be analyzed, and match target data indicators in a preset data indicator set according to the information of question data to be analyzed;

[0037] In the case where the target data index fails to match, matching the corresponding target data entity according to the question information to be analyzed;

[0038] In the case where the target data entity is matched successfully, a data indicator template is obtained, and a target data indicator is determined according to the target data entity and the data indicator template;

[0039] Target data is acquired according to the target data indicator, and a data analysis result is obtained according to the target data.

[0040] In a fourth aspect, the present application further provides a computer-readable storage medium having a computer program stored thereon, wherein when the computer program is executed by a processor, the following steps are implemented:

[0041] Acquire the question data information to be analyzed, and match the target data indicator in the preset data indicator set according to the question data information to be analyzed;

[0042] In the case where the target data index fails to match, matching the corresponding target data entity according to the question information to be analyzed;

[0043] In the case where the target data entity is matched successfully, a data indicator template is obtained, and a target data indicator is determined according to the target data entity and the data indicator template;

[0044] Target data is acquired according to the target data indicator, and data analysis results are obtained according to the target data.

[0045] In a fifth aspect, the present application further provides a computer program product, including a computer program, which implements the following steps when executed by a processor:

[0046] Acquire the question data information to be analyzed, and match the target data indicator in the preset data indicator set according to the question data information to be analyzed;

[0047] In the case where the target data index fails to match, matching the corresponding target data entity according to the question information to be analyzed;

[0048] In the case where the target data entity is matched successfully, a data indicator template is obtained, and a target data indicator is determined according to the target data entity and the data indicator template;

[0049] Target data is acquired according to the target data indicator, and a data analysis result is obtained according to the target data.

[0050] The above-mentioned large language model data analysis method, device, computer equipment, storage medium and computer program product obtain the question information to be analyzed, and match the target data indicator in the preset data indicator set according to the question information to be analyzed. When the target data indicator fails to match, the corresponding target data entity is matched according to the question information to be analyzed. When the target data entity matches successfully, the data indicator template is obtained, the target data indicator is determined according to the target data entity and the data indicator template, the target data is obtained according to the target data indicator, and the data analysis result is obtained according to the target data, and the indicator matching is combined with the data entity matching, and the indicator direct matching method is adopted to ensure the data acquisition efficiency. At the same time, for the use scenario with a large number of complex forms, when the complexity of the form increases and the target data indicator cannot be directly matched, the target data entity and the data indicator template are used to determine the target data indicator, so as to perform indicator processing on the target object, thereby improving the data acquisition accuracy. Therefore, the above-mentioned method can improve the accuracy of intelligent question counting in the presence of a large number of complex forms. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] 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.

[0052] FIG1 is a diagram of an application environment of a large language model data analysis method according to an embodiment;

[0053] FIG2 is a schematic diagram of a flow chart of a large language model data analysis method according to an embodiment;

[0054] FIG3 is a schematic diagram of a flow chart of a large language model data analysis method in another embodiment;

[0055] FIG4 is a block diagram of a large language model data analysis device in one embodiment;

[0056] FIG5 is a diagram showing the internal structure of a computer device in one embodiment;

[0057] FIG. 6 is a diagram showing the internal structure of a computer device in another embodiment. DETAILED DESCRIPTION

[0058] 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.

[0059] The large language model data analysis method provided in the embodiment of the present application can be applied in the application environment shown in Figure 1. 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 integrated on the server 104, or it can be placed on the cloud or other network servers. The data storage system can be used to store data that needs to be processed, such as initial question information, question information to be analyzed, and can also be used to store preset data such as preset data indicator sets. Among them, the terminal 102 can be, but is not limited to, various personal computers, laptops, smart phones, tablet computers, Internet of Things devices and portable wearable devices. The Internet of Things devices can be smart speakers, smart TVs, smart air conditioners, smart car-mounted devices, etc. Portable wearable devices can be smart watches, smart bracelets, head-mounted devices, etc. The server 104 can be implemented with an independent server or a server cluster consisting of multiple servers.

[0060] The large language model data analysis method provided in the embodiment of the present application can be applied to the data analysis of the large language model intelligent question-and-answer scenario. LLM (Large Language Model) refers to a large language model, which is a large-scale natural language processing model trained based on a deep learning algorithm. It can understand and generate natural language by training on massive text data, and has multiple capabilities such as language understanding, reasoning, and translation. In the field of data processing and analysis, LLM is usually used in scenarios of natural language understanding and natural language generation. For example, when a user enters a question, LLM can understand the intent of the question based on the context, convert unstructured natural language questions into query statements with structured meaning, and support complex data analysis tasks. It can process language in a way that is closer to humans, reduce the dependence of traditional rule systems on fixed expressions, and thus improve flexibility and applicability. When combined with business, LLM usually serves as the "front end" of problem parsing to provide semantic input for subsequent logical queries and data processing.

[0061] This method can be applied to various information management systems, such as enterprise resource planning system (ERP), enterprise management system, financial system, human resource system, supply chain system, etc.

[0062] In an exemplary embodiment, as shown in FIG2 , a large language model data analysis method is provided, which is described by taking the method applied to the server 104 in FIG1 as an example, and includes the following steps S202 to S208 . Among them:

[0063] Step S202: obtaining information of question data to be analyzed, and matching target data indicators in a preset data indicator set according to the information of question data to be analyzed.

[0064] The data information to be analyzed may be directly obtained from the terminal 102 , or may be obtained by the server 104 through data processing of original data obtained from the terminal 102 .

[0065] Furthermore, in this embodiment, the method is applied to the intelligent data query scenario of the enterprise resource planning system as an example. Among them, ERP is a comprehensive management system that integrates different business management modules, aims to improve the utilization efficiency of enterprise resources, reduce operating costs, and optimize various processes within the enterprise. It integrates various resources (such as manpower, material resources, funds, information, etc.) of the enterprise through information technology, thereby realizing the automation, standardization and transparency of enterprise management. Through a unified data platform, operations in various links such as finance, sales, procurement, inventory, and production can be seamlessly connected to ensure real-time sharing and accurate transmission of data.

[0066] Exemplarily, server 104 can use the ERP digital intelligence base to implement intelligent analysis and intelligent query of LLM data in the following steps. Among them, the ERP digital intelligence base can include an ERP data layer and a semantic model layer. The ERP data layer is used to deploy the ERP data model, indicator platform, and data permissions; the semantic model layer is used to convert natural language questions into query statements that can be recognized by the ERP system, and accurately extract the required data from the ERP database; the vector layer is used to deploy a vector library, which is used to convert natural semantic questions and data items in ERP (such as field names, table names, business terms, etc.) into vector tables, and then quickly find content that matches user questions and answers through vector calculations.

[0067] Among them, the semantic model layer of the ERP digital intelligence base can be used to convert natural language questions into query statements that can be recognized by the ERP system. The semantic model layer includes but is not limited to: ERP professional vocabulary, business confusion word dictionary, business problem template, indicator / dimension definition dictionary, business object dictionary, time dimension analysis and definition, etc., which are pre-organized and pre-built based on ERP common vocabulary, common questions, common indicators and other contents. By building a semantic model layer, the intelligence level and query accuracy of the Question Assistant can be improved, making data acquisition more efficient.

[0068] Accordingly, in the above step S202, the server 104 may use the semantic model layer to parse the acquired query information to be analyzed, obtain a query statement recognizable by the ERP system, and then input it into the indicator platform for matching data indicators.

[0069] Among them, the indicator platform is a pre-built ERP data indicator based on the ERP data model. It includes various preset data indicators for commonly used intelligent calculations, which are used to match accurate data indicator values ​​during intelligent calculations.

[0070] Accordingly, in the above step S202, the server 104 can pre-store a preset data indicator set in the above indicator platform, and then use the ERP data model to match the target data indicator in the preset data indicator set according to the question information to be analyzed. When the target data indicator is successfully matched, the matched target data indicator can be directly used to obtain the target data, and then the data analysis step such as S208 is performed according to the obtained target data.

[0071] Step S204: when the target data index matching fails, matching the corresponding target data entity according to the question information to be analyzed.

[0072] Exemplarily, when there is no target data indicator that can be matched in the preset data indicator set of the indicator platform, and the target data indicator fails to match, the server 104 can use the NL2ORM method to match the corresponding target data entity.

[0073] Among them, NL2ORM is a natural language to object-relational mapping technology that can convert users' natural language questions into query operations on business objects and their attributes. ORM is essentially a way to operate the underlying data model through business objects, which is suitable for scenarios of domain models and object-oriented development. The role of NL2ORM is similar to giving semantics to natural language. By parsing users' questions, it matches specific domain models and generates queries or operations related to business entities. In the application process, NL2ORM pays more attention to operations at the "business object" level: when users ask questions, such as "query the total order amount of a certain customer", NL2ORM first converts the question into the concept of the domain model, such as "customer" and "order". It finds specific data entities (such as "customer table" and "order table") through mapping, and automatically generates standardized indicators, such as total order amount, order quantity, etc. In this process, NL2ORM combines template capabilities to standardize business models and data models, making the parsing of natural language questions more accurate and reducing development complexity.

[0074] Furthermore, the ERP data model in the above-mentioned ERP digital intelligence base may include a domain directory, which may be pre-built or automatically built according to business objects. The domain directory is used to record the business objects existing in each domain and module of ERP, as well as the relationship between domains, modules, and business objects. When adding a new business object, the server 104 may automatically create or manually supplement business information in the domain directory to expand the types and corresponding data of the current candidate data entities.

[0075] For example, in the above step S204, when the target data indicator fails to match, the server 104 can use the ERP data model to match the data information to be analyzed to the corresponding domain directory, and then find the target data entity according to the domain directory.

[0076] Step S206, when the target data entity is matched successfully, a data indicator template is obtained, and a target data indicator is determined according to the target data entity and the data indicator template.

[0077] Exemplarily, the server 104 may extract the characteristic fields of the target data entity; generate basic data indicators and enhanced data indicators that match the characteristic fields according to the data indicator template; and collect the basic data indicators and enhanced data indicators to obtain the target data indicators. Specifically, the server 104 may first select the corresponding data indicator template according to the extracted characteristic fields of the target data entity. Among them, the characteristic fields of the target data entity can be used to characterize the applicable conditions such as the type and field of the data indicator template. Next, the server 104 may classify the fields contained in the selected data indicator template, and the server 104 may identify the field tags of each field in the data indicator template, and use the fields whose field tags are the first indicator tags as the basic data indicators, and use the fields whose field tags are the second indicator tags as the enhanced data indicators. Among them, the data corresponding to the basic data indicators are directly obtained, and the basic data indicators may include any one or more of the indicator name, analysis dimension, statistical period, and filtering condition. The data corresponding to the enhanced data indicators are obtained through calculation, and the enhanced data indicators may include year-on-year value, year-on-year growth rate, and cumulative growth rate, etc.

[0078] Furthermore, in the case where the target data entity fails to match, the server 104 may use the NL2SQL solution to perform data analysis, and the server 104 may obtain a pre-built analysis table according to the data information to be analyzed; obtain the target data according to the pre-built analysis table, and obtain the data analysis result according to the target data. The pre-built analysis table is pre-built according to the data entity and data model related to the data information to be analyzed.

[0079] NL2SQL is a semi-processed data acquisition solution, a natural language to structured query language conversion technology, which is used to directly convert natural language questions into SQL statements in order to query the database and obtain results. It is oriented to the underlying database operations and generates SQL query statements with correct syntax and clear semantics through natural language understanding. Compared with NL2ORM, NL2SQL is more inclined to the data model layer rather than the business object layer. It directly operates the tables and fields of the database and can provide highly flexible query capabilities. Data entities are the most original and detailed data, such as the "sales order" table, which includes many complex fields (such as product number, customer name, product quantity, etc.) and contains a lot of complex information. Unlike data entities, pre-built analysis tables are data tables that have been sorted and processed. They do some pre-processing on the original data and only retain the parts required for analysis, such as "customer name, quantity shipped, amount shipped", etc. It is simpler and easier to calculate and analyze.

[0080] For example, the server 104 can use NL2SQL technology to directly query the "pre-processed" analysis table. The analysis table has integrated and simplified the complex data, and only contains the required parts, such as the "amount shipped" field, calculation formulas, etc. By pre-processing the complex original data (such as sales orders, sales shipping orders) into "analysis tables", it is possible to avoid processing a large amount of irrelevant data during query, reduce the complexity of the query, and improve the accuracy.

[0081] Furthermore, the ERP data model in the above ERP digital intelligence base can also include an analysis layer and a source layer. The analysis layer can pre-build data analysis tables based on the data entities of business objects, clean and process the business object entities in the fields of finance, supply chain, production, etc., to obtain the above pre-built analysis tables; the source layer can be used to align the pre-built analysis tables with the data entities, making it easier for users to understand.

[0082] Exemplarily, in the event that the target data entity match fails, the server 104 may use the ERP data model to retrieve the information of the question to be analyzed at the analysis layer, obtain the corresponding pre-built analysis table, thereby obtaining the target data based on the pre-built analysis table, and perform the following data analysis step S208 on the target data to obtain the data analysis result.

[0083] Furthermore, the above analysis table can be pre-constructed into a single-header, single-body structure, and the content that needs to be statistically processed is pre-counted in the analysis table. For example, the data entity sales order is a complex multi-header, multi-body structure, which may contain more than 500 fields of information, and the data entity sales delivery order is also a complex multi-header, multi-body structure, which may contain more than 600 fields of information. The information currently to be analyzed is the amount of goods shipped out of the latest sales order of a certain user. When the information cannot be directly obtained from the data entity sales order and the sales delivery order, the server 104 can pre-process and construct the data entity sales order and the sales delivery order through the pre-constructed analysis table, first integrate the two data entity tables into a single-header, single-body data analysis table, and non-analysis-related data does not enter the data analysis table. This design reduces the complexity of the analysis table, and then pre-constructs the delivery identification, the quantity shipped out, the amount shipped out field and the calculation logic. When NL2SQL needs to obtain the amount of goods shipped out, server 104 can directly obtain it through a pre-built analysis table. When it needs to obtain the amount of goods not shipped out, server 104 can generate code through NL2SQL to obtain the order amount and the amount of goods shipped out from the pre-analysis table, and then calculate the amount of goods not shipped out.

[0084] Step S208, acquiring target data according to the target data indicator, and obtaining data analysis results according to the target data.

[0085] The target data may include basic target data and enhanced target data.

[0086] Exemplarily, the server 104 can obtain basic target data based on the basic data indicators, and calculate enhanced target data based on the enhanced data indicators and the basic target data. Among them, the basic target data may include multiple aspects such as indicator name, analysis dimension, statistical period, filtering conditions, etc. Specifically, the indicator name represents the specific business indicator to be analyzed; the analysis dimension determines the angles from which the data is disassembled, such as by time, by region, by department, etc.; the statistical period defines the time range of the data, such as day, month, year, etc.; and the filtering condition limits the scope of the data or the screening criteria according to user needs. Enhanced target data refers to the addition of more analysis dimensions or calculation methods on the basis of basic data indicators to generate more in-depth and comprehensive data. Enhanced target data may include multiple aspects such as year-on-year value, year-on-year growth rate and cumulative growth rate.

[0087] Further, taking the enhanced target data as a year-on-year value as an example, the year-on-year value is used to compare the data changes at two time points or time periods. The server 104 can obtain the data to be calculated according to the indicator name, statistical period and filtering condition, and calculate the obtained data to be calculated according to the calculation method corresponding to the year-on-year value to obtain the year-on-year value.

[0088] Exemplarily, after acquiring these basic target data, the server 104 may first select the data range, and determine the two time periods to be analyzed year-on-year according to the statistical period and filtering conditions defined by the user, wherein the indicator name may be inventory change data, the statistical period may be one month, and the filtering condition may be that the interval between the two time periods is one year. For example, the server 104 may select the inventory change data of the current month to compare with the inventory change data of the same month last year. After determining the time period, the server 104 may extract the inventory change data within the two time periods according to the basic data indicators. For example, the server 104 may extract the inventory change amount of the current month and the same month last year. Next, the server 104 may calculate the year-on-year value according to the year-on-year value calculation formula, and the year-on-year value calculation formula may be (current period value−last year value) / last year value×100%, so as to calculate the growth percentage between the two time periods and measure the change of the data. Through the calculation of the year-on-year value, the result obtained by the server 104 is the enhanced target data. These data provide a clear perspective to help analyze the trends and changes between current data and historical data, and further support decision-making.

[0089] Furthermore, taking the enhanced target data as the year-on-year growth rate as an example, the year-on-year growth rate can be used to measure the degree of change of a certain indicator between the current cycle and the same cycle of the previous year, so as to evaluate the annual change trend. The server 104 can obtain the data to be calculated according to the indicator name, statistical cycle and filtering conditions, and calculate the obtained data to be calculated according to the calculation method corresponding to the year-on-year growth rate to obtain the year-on-year growth rate. Among them, the indicator name can be the inventory quantity, the statistical cycle can be one month, and the filtering condition can be that the time interval between the two time periods is one year. Next, the server 104 can calculate the year-on-year growth rate according to the year-on-year growth rate formula: (current cycle value-previous cycle value) / previous cycle value×100%.

[0090] Furthermore, taking the cumulative growth rate as an example of enhanced target data, the cumulative growth rate is a measure of the overall growth rate of a certain indicator from the starting value to the current value within a specific time period, which can be used to evaluate long-term trends. The server 104 can obtain the data to be calculated according to the indicator name and the statistical period, and calculate the obtained data to be calculated according to the calculation method corresponding to the cumulative growth rate to obtain the cumulative growth rate. Among them, the indicator name can be the number of products, the statistical period can be one year, the number of products at the beginning of the year is used as the initial value, and the number of products at the end of the year is used as the current value. Next, the server 104 can calculate the cumulative growth rate according to the cumulative growth rate formula: (current value-initial value) / initial value×100%.

[0091] Exemplarily, the server 104 can use the ERP data model, find the corresponding data entity by matching the domain directory of the ERP data model, introduce the data entity into the above-mentioned indicator platform, and the indicator platform provides an indicator model for processing and standardizing data output. Among them, the indicator model includes core components and enhanced components, the core components are used to obtain basic target data, and the enhanced components are used to obtain enhanced target data. After the data entity is introduced into the indicator platform, the server 104 can first match the fields of the data entity with the core components for output, which may include indicator names, analysis dimensions, statistical cycles, filtering conditions, etc.; other fields match the output of the enhanced components, such as ratios, etc., so as to obtain the data results of each data indicator as the target data, and then obtain the data analysis results according to the target data. This method applies the data standardization and template analysis capabilities of the indicator platform, and constructs the analysis data of the business entity according to the indicator template, so that the business objects of ERP can support a wide range of questions, and at the same time, temporary indicator processing can be performed on the business objects.

[0092] Exemplarily, the server 104 can establish an indicator platform according to the above usage scenario, so as to provide data analysis services to the terminal 102. In the indicator platform, a preset data indicator set for indicator query is preset, and an indicator directory is generated accordingly. When the data information to be analyzed is obtained, the server 104 can match the data content to be queried with the indicator directory, dimensions, etc. of the indicator platform. If the match is successful, the corresponding data indicator is obtained and called. Since the indicator platform presets the commonly used data indicators based on ERP, the server 104 can accurately and quickly match the commonly used data indicators.

[0093] Furthermore, during the data analysis process, the server 104 can pass the target data into the expert analysis framework. The expert analysis framework provides expert analysis templates in the ERP business field, and by calling expert analysis templates of different problem types, expert analysis suggestions are formed to finally obtain data analysis results. Exemplarily, the server 104 can receive target data from different data sources according to target data indicators, and after pre-processing such as data cleaning, pass the target data to the expert analysis framework in a format that meets the requirements of the expert analysis template. Among them, the expert analysis framework is used to process data analysis in the ERP business field. Multiple expert analysis templates are integrated within the framework. These templates are formulated by domain experts based on actual needs and experience, and can provide systematic analysis methods and solutions for specific business problems. Each expert analysis template has clear goals and analysis methods for different problem types, such as inventory management, data report analysis, etc.

[0094] In actual operation, the server 104 can select an appropriate expert analysis template to call based on the received target data, the nature of the problem and the different business needs. For example, if the target data involves inventory change trend analysis, the server 104 can select an expert analysis template related to inventory changes. When the corresponding expert analysis template is called, the expert analysis framework performs specific calculations and data analysis tasks according to the provisions of the template. Finally, the server 104 can feed back the analysis results and suggestions obtained from the expert analysis framework to relevant personnel or systems to support a more efficient decision-making process. This process ensures that the data in the ERP system can be fully utilized, and the generated analysis suggestions can provide reliable decision support for the enterprise's operational optimization.

[0095] In the above-mentioned large language model data analysis method, by obtaining the information of the question to be analyzed, and matching the target data indicator in the preset data indicator set according to the question to be analyzed information, when the target data indicator fails to match, the corresponding target data entity is matched according to the question to be analyzed information, and when the target data entity is matched successfully, the data indicator template is obtained, the target data indicator is determined according to the target data entity and the data indicator template, the target data is obtained according to the target data indicator, and the data analysis result is obtained according to the target data, the indicator matching is combined with the data entity matching, and the indicator direct matching method is adopted to ensure the data acquisition efficiency. At the same time, for the use scenarios with a large number of complex forms, when the complexity of the form increases and the target data indicator cannot be directly matched, the target data entity and the data indicator template are used to determine the target data indicator, so as to perform indicator processing on the target object, thereby improving the accuracy of data acquisition, so as to improve the accuracy of intelligent question counting in the presence of a large number of complex forms.

[0096] In an exemplary embodiment, as shown in FIG3 , before step S202 , the above method may further include steps S302 to S304 . Among them:

[0097] Step S302, in response to the data analysis instruction, obtaining initial question information, and extracting target question information belonging to a preset target field from the initial question information.

[0098] Exemplarily, the server 104 may receive data analysis instructions and initial question information through the terminal 102. Specifically, the server 104 may receive and obtain questions sent by the user in voice or text through the mobile terminal, PC terminal intelligent question number, and intelligent assistant portal. Next, the server 104 may determine whether the obtained user question is related to the ERP field through the field identification of the vector layer ERP professional vocabulary; if it is a non-ERP field key question, it is matched to the common sense question answering; if it is a ERP field related question, it is extracted as the target question number information of the preset target field.

[0099] Step S304: input the target question number information into the semantic analysis model to perform question number intent recognition to obtain an intent recognition result.

[0100] For example, in the ERP system, the questions raised by users may often have missing information or incomplete descriptions. The server 104 can supplement the user input through the vector layer to fill in the key information that the user often omits in the description. The vector layer understands the context of the user's question, thereby inferring the missing part, or automatically completing the description of the question through other information in the conversation, ensuring that the system can obtain the complete context and data requirements so as to more accurately process subsequent requests. The server 104 can determine whether there is a need to split the problem. Some complex questions may contain multiple sub-questions. The system analyzes the structure of the user's question and identifies which parts can be broken down into independent sub-questions for separate processing. After completing and splitting the problem, the server 104 can enter the user intention recognition stage, and through semantic analysis of the completed or split questions, identify what the user's actual needs are, and clarify what kind of answer or result the user wants to obtain. According to the identified intention, the system will call the relevant data module, relying on the data of various business modules (such as sales, inventory, finance, etc.) in the ERP system to perform necessary data query and processing.

[0101] Correspondingly, step S202 may include: step S306, obtaining question information to be analyzed according to the intention recognition result.

[0102] Exemplarily, the server 104 can directly use the intent recognition results as question information to be analyzed, or classify the intent recognition results to obtain different types of question information to be analyzed to distinguish the corresponding analysis models, thereby improving data analysis efficiency. In addition, the server 104 can also set priorities for different types of intent recognition results based on the content of the intent recognition results, thereby further improving data analysis efficiency and flexibility.

[0103] In an exemplary embodiment, during the process of question intent recognition, the semantic analysis model can be used to supplement and / or split the target question information. The server 104 can also input the target question information into the semantic analysis model for supplementation and / or splitting, and match the approximate question template according to the supplemented and / or split target question information; if the approximate question template matches successfully, use the data index in the approximate question template as the target data index, and perform the step of obtaining the target data according to the target data index; if the approximate question template matches unsuccessfully, perform the step of obtaining the question information to be analyzed according to the intent recognition result.

[0104] Furthermore, the above-mentioned ERP digital intelligence base can also include a vector layer, which can dynamically maintain a vector library. The server 104 can use the vector library to convert natural semantic questions and data items in the ERP (such as field names, table names, business terms, etc.) into vector tables, and then quickly find content that matches user questions and answers through vector calculations, so that it can more quickly complete descriptions that users often omit, and determine whether the question needs to be split, and then make a supplementary description of the question or split multiple questions.

[0105] In another exemplary embodiment, in response to the data analysis instruction, the server 104 obtains the initial question information, extracts the target question information belonging to the preset target field in the initial question information, inputs the target question information into the semantic analysis model to supplement and / or split the target question information, and then performs question intent recognition to obtain intent recognition results; matches the approximate question template according to the supplemented and / or split target question information; if the approximate question template matches successfully, uses the data index in the approximate question template as the target data index, and executes to obtain the target data according to the target data index. If the approximate question template matches unsuccessfully, the question information to be analyzed is obtained according to the intent recognition result.

[0106] Next, the server 104 obtains the question information to be analyzed, and matches the target data indicator in the preset data indicator set according to the question information to be analyzed. If the target data indicator fails to match, the server 104 matches the corresponding target data entity according to the question information to be analyzed.

[0107] In the case where the target data entity matching fails, a pre-built analysis table pre-built according to the data entity and data model related to the question information to be analyzed is obtained according to the question information to be analyzed, the target data is obtained according to the pre-built analysis table, and the data analysis result is obtained according to the target data.

[0108] Accordingly, when the target data entity is matched successfully, a data indicator template is obtained, characteristic fields of the target data entity are extracted, basic data indicators and enhanced data indicators matching the characteristic fields are generated according to the data indicator template, and the basic data indicators and enhanced data indicators are aggregated to obtain the target data indicators.

[0109] Next, the server 104 obtains basic target data according to the basic data indicators, wherein the basic data indicators include any one or more of the indicator name, analysis dimension, statistical period, and filtering condition, and finally calculates the enhanced target data according to the enhanced data indicators and the basic target data, and obtains the data analysis result according to the target data.

[0110] 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.

[0111] Based on the same inventive concept, the embodiment of the present application also provides a large language model data analysis device for implementing the large language model 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 large language model data analysis device embodiments provided below can refer to the limitations of the large language model data analysis method above, and will not be repeated here.

[0112] In an exemplary embodiment, as shown in FIG4 , a large language model data analysis device is provided, including: an information acquisition module 402, a data matching module 404, an indicator matching module 406 and a data analysis module 408, wherein:

[0113] The information acquisition module 402 is used to acquire the data information to be analyzed, and match the target data index in the preset data index set according to the data information to be analyzed;

[0114] The data matching module 404 is used to match the corresponding target data entity according to the data information to be analyzed when the target data index matching fails;

[0115] The indicator matching module 406 is used to obtain the data indicator template when the target data entity is successfully matched, and determine the target data indicator according to the target data entity and the data indicator template;

[0116] The data analysis module 408 is used to obtain target data according to the target data indicator and obtain data analysis results according to the target data.

[0117] In one embodiment, the indicator matching module 406 is specifically used to: extract the characteristic fields of the target data entity; generate basic data indicators and enhanced data indicators that match the characteristic fields according to the data indicator template; and collect the basic data indicators and enhanced data indicators to obtain the target data indicators.

[0118] In one embodiment, the target data includes basic target data and enhanced target data; the data analysis module 408 is specifically used to: obtain basic target data according to basic data indicators; wherein the basic data indicators include any one or more of indicator name, analysis dimension, statistical period, and filtering condition; and calculate the enhanced target data according to the enhanced data indicators and the basic target data.

[0119] In one embodiment, the device further comprises:

[0120] An analysis table acquisition unit, used to acquire a pre-built analysis table according to the question information to be analyzed when the target data entity fails to match; wherein the pre-built analysis table is pre-built according to the data entity and data model related to the question information to be analyzed;

[0121] The data analysis unit is used to obtain target data according to a pre-built analysis table and obtain data analysis results according to the target data.

[0122] In one embodiment, the device further comprises:

[0123] An initial information acquisition unit, configured to acquire initial question information in response to a data analysis instruction, and extract target question information belonging to a preset target field from the initial question information;

[0124] An intention recognition unit is used to input the target question information into the semantic analysis model to perform question intention recognition and obtain an intention recognition result;

[0125] Accordingly, the information acquisition module 402 is specifically used to obtain the question information to be analyzed according to the intention recognition result.

[0126] In one of the embodiments, the device is also used to: input the target question information into the semantic analysis model for supplementation and / or splitting, and match the approximate question template based on the supplemented and / or split target question information; when the approximate question template matches successfully, use the data indicator in the approximate question template as the target data indicator, and execute the step of obtaining the target data according to the target data indicator; when the approximate question template matches unsuccessfully, execute the step of obtaining the question information to be analyzed according to the intention recognition result.

[0127] Each module in the above-mentioned large language model data analysis device can be implemented in whole or in part by software, hardware and a combination thereof. Each of the above-mentioned modules can be embedded in or independent of the processor in the computer device in the form of hardware, or can be stored in the memory in the computer device in the form of software, so that the processor can call and execute the operations corresponding to each of the above modules.

[0128] In an exemplary embodiment, a computer device is provided, which may be a server, and its internal structure diagram may be shown in FIG5. The computer device includes a processor, a memory, an input / output interface (Input / Output, referred to as I / O) and a communication interface. 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. 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 data to be processed, such as initial question information and question information to be analyzed, and can also be used to store preset data such as a preset data indicator set. 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 large language model data analysis method is implemented.

[0129] In an exemplary embodiment, a computer device is provided, which may be a terminal, and its internal structure diagram may be shown in FIG6. 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 via a system bus, and the communication interface, the display unit, and the input device are connected to the system bus via 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 may 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 large language model data analysis method is implemented. The display unit of the computer device is used to form a visually visible picture, which may 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, and the input device of the computer device can be a touch layer covering the display screen, or a button, trackball or touchpad set on the computer device shell, or an external keyboard, touchpad or mouse.

[0130] Those skilled in the art will understand that the structure shown in FIG. 6 is merely a block diagram of a partial 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 shown in the figure, or combine certain components, or have a different arrangement of components.

[0131] In an exemplary embodiment, a computer device is provided, including a memory and a processor, wherein a computer program is stored in the memory, and the processor implements the following steps when executing the computer program: obtaining question information to be analyzed, and matching target data indicators in a preset data indicator set according to the question information to be analyzed; in the case where the target data indicator fails to match, matching the corresponding target data entity according to the question information to be analyzed; in the case where the target data entity matches successfully, obtaining a data indicator template, determining the target data indicator according to the target data entity and the data indicator template; obtaining target data according to the target data indicator, and obtaining a data analysis result according to the target data.

[0132] In one embodiment, when the processor executes the computer program, the following steps are also implemented: extracting characteristic fields of the target data entity; generating basic data indicators and enhanced data indicators that match the characteristic fields according to the data indicator template; and aggregating the basic data indicators and enhanced data indicators to obtain the target data indicators.

[0133] In one embodiment, when the processor executes the computer program, the following steps are also implemented: basic target data is obtained according to basic data indicators; wherein the basic data indicators include any one or more of indicator name, analysis dimension, statistical period, and filtering condition; and enhanced target data is calculated according to enhanced data indicators and basic target data.

[0134] In one embodiment, the processor further implements the following steps when executing the computer program: in the event that the target data entity fails to match, obtaining a pre-built analysis table based on the question information to be analyzed; wherein the pre-built analysis table is pre-built based on the data entity and data model related to the question information to be analyzed; obtaining the target data based on the pre-built analysis table, and obtaining a data analysis result based on the target data.

[0135] In one embodiment, when the processor executes the computer program, the following steps are also implemented: in response to the data analysis instruction, initial question information is obtained, and target question information belonging to a preset target field is extracted from the initial question information; the target question information is input into the semantic analysis model for question intent recognition to obtain an intent recognition result; and the question information to be analyzed is obtained based on the intent recognition result.

[0136] In one embodiment, when the processor executes the computer program, the following steps are also implemented: inputting the target question information into the semantic analysis model for supplementation and / or splitting, and matching the approximate question template according to the supplemented and / or split target question information; when the approximate question template matches successfully, using the data indicator in the approximate question template as the target data indicator, and executing the step of obtaining the target data according to the target data indicator; when the approximate question template matches unsuccessfully, executing the step of obtaining the question information to be analyzed according to the intention recognition result.

[0137] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the following steps are implemented: obtaining question information to be analyzed, and matching target data indicators in a preset data indicator set according to the question information to be analyzed; in the case where the target data indicator fails to match, matching the corresponding target data entity according to the question information to be analyzed; in the case where the target data entity is matched successfully, obtaining a data indicator template, determining the target data indicator according to the target data entity and the data indicator template; obtaining target data according to the target data indicator, and obtaining a data analysis result according to the target data.

[0138] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented: extracting characteristic fields of the target data entity; generating basic data indicators and enhanced data indicators that match the characteristic fields according to a data indicator template; and aggregating the basic data indicators and enhanced data indicators to obtain target data indicators.

[0139] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented: basic target data is obtained based on basic data indicators; wherein the basic data indicators include any one or more of indicator name, analysis dimension, statistical period, and filtering condition; enhanced target data is calculated based on enhanced data indicators and basic target data.

[0140] In one embodiment, when the computer program is executed by the processor, the following steps are also implemented: in the event that the target data entity fails to match, a pre-built analysis table is obtained based on the question information to be analyzed; wherein the pre-built analysis table is pre-built based on the data entity and data model related to the question information to be analyzed; the target data is obtained based on the pre-built analysis table, and a data analysis result is obtained based on the target data.

[0141] In one embodiment, when the computer program is executed by the processor, the following steps are also implemented: in response to the data analysis instruction, initial question information is obtained, and target question information belonging to a preset target field is extracted from the initial question information; the target question information is input into the semantic analysis model for question intent recognition to obtain an intent recognition result; and the question information to be analyzed is obtained based on the intent recognition result.

[0142] In one embodiment, when the computer program is executed by the processor, the following steps are also implemented: inputting the target question information into the semantic analysis model for supplementation and / or splitting, and matching the approximate question template according to the supplemented and / or split target question information; when the approximate question template matches successfully, using the data indicator in the approximate question template as the target data indicator, and executing the step of obtaining the target data according to the target data indicator; when the approximate question template matches unsuccessfully, executing the step of obtaining the question information to be analyzed according to the intention recognition result.

[0143] In one embodiment, a computer program product is provided, including a computer program, which implements the following steps when executed by a processor: obtaining information of question data to be analyzed, and matching a target data indicator in a preset data indicator set according to the question data to be analyzed; in the case where the target data indicator fails to match, matching the corresponding target data entity according to the question data to be analyzed; in the case where the target data entity is successfully matched, obtaining a data indicator template, determining the target data indicator according to the target data entity and the data indicator template; obtaining target data according to the target data indicator, and obtaining a data analysis result according to the target data.

[0144] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented: extracting characteristic fields of the target data entity; generating basic data indicators and enhanced data indicators that match the characteristic fields according to a data indicator template; and aggregating the basic data indicators and enhanced data indicators to obtain target data indicators.

[0145] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented: basic target data is obtained based on basic data indicators; wherein the basic data indicators include any one or more of indicator name, analysis dimension, statistical period, and filtering condition; enhanced target data is calculated based on enhanced data indicators and basic target data.

[0146] In one embodiment, when the computer program is executed by the processor, the following steps are also implemented: in the event that the target data entity fails to match, a pre-built analysis table is obtained based on the question information to be analyzed; wherein the pre-built analysis table is pre-built based on the data entity and data model related to the question information to be analyzed; the target data is obtained based on the pre-built analysis table, and a data analysis result is obtained based on the target data.

[0147] In one embodiment, when the computer program is executed by the processor, the following steps are also implemented: in response to the data analysis instruction, initial question information is obtained, and target question information belonging to a preset target field is extracted from the initial question information; the target question information is input into the semantic analysis model for question intent recognition to obtain an intent recognition result; and the question information to be analyzed is obtained based on the intent recognition result.

[0148] In one embodiment, when the computer program is executed by the processor, the following steps are also implemented: inputting the target question information into the semantic analysis model for supplementation and / or splitting, and matching the approximate question template according to the supplemented and / or split target question information; when the approximate question template matches successfully, using the data indicator in the approximate question template as the target data indicator, and executing the step of obtaining the target data according to the target data indicator; when the approximate question template matches unsuccessfully, executing the step of obtaining the question information to be analyzed according to the intention recognition result.

[0149] 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.

[0150] 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.

[0151] 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.

[0152] 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 understood as limiting the scope of the present application. It should be pointed out that, for a person of ordinary skill in the art, several variations 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 large language model data analysis method, characterized in that: The method comprises: Acquire information of question data to be analyzed, and match target data indicators in a preset data indicator set according to the information of question data to be analyzed; In the case where the target data index fails to match, matching the corresponding target data entity according to the question information to be analyzed; In the case where the target data entity is matched successfully, a data indicator template is obtained, and a target data indicator is determined according to the target data entity and the data indicator template; Target data is acquired according to the target data indicator, and data analysis results are obtained according to the target data.

2. The method according to claim 1, characterized in that Determining the target data indicator according to the target data entity and the data indicator template includes: Extracting characteristic fields of the target data entity; Generate basic data indicators and enhanced data indicators matching the characteristic fields according to the data indicator template; The basic data indicators and the enhanced data indicators are collected to obtain target data indicators.

3. The method according to claim 2, characterized in that The target data includes basic target data and enhanced target data; the step of acquiring the target data according to the target data indicator includes: Acquire the basic target data according to the basic data indicators; wherein the basic data indicators include any one or more of indicator name, analysis dimension, statistical period, and filtering condition; The enhanced target data is obtained by calculation according to the enhanced data indicator and the basic target data.

4. The method according to claim 1, characterized in that: In the case where the target data indicator fails to match, after matching the corresponding target data entity according to the question information to be analyzed, the method further includes: In the case where the target data entity fails to match, a pre-built analysis table is obtained according to the question information to be analyzed; wherein the pre-built analysis table is pre-built according to the data entity and data model related to the question information to be analyzed; Target data is acquired according to the pre-built analysis table, and a data analysis result is obtained according to the target data.

5. The method according to any one of claims 1 to 4, characterized in that: Before obtaining the information of the question to be analyzed, the method further includes: In response to the data analysis instruction, initial question number information is acquired, and target question number information belonging to a preset target field is extracted from the initial question number information; Inputting the target question number information into a semantic analysis model to perform question number intent recognition to obtain an intent recognition result; The obtaining of the question information to be analyzed includes: obtaining the question information to be analyzed according to the intention recognition result.

6. The method according to claim 5, characterized in that The method further comprises: Inputting the target question information into a semantic analysis model for supplementation and / or splitting; Matching an approximate question template according to the target question number information after supplementation and / or splitting; In the case where the approximate question template is matched successfully, using the data indicator in the approximate question template as the target data indicator, and performing the step of acquiring the target data according to the target data indicator; In the case where the approximate question template matching fails, the step of obtaining the question information to be analyzed according to the intention recognition result is performed.

7. A large language model data analysis device, characterized in that: The device comprises: An information acquisition module is used to acquire the data information to be analyzed, and match the target data indicator in the preset data indicator set according to the data information to be analyzed; A data matching module, used for matching the corresponding target data entity according to the data information to be analyzed when the target data indicator fails to match; An indicator matching module, used to obtain a data indicator template when the target data entity is successfully matched, and determine a target data indicator according to the target data entity and the data indicator template; The data analysis module is used to obtain target data according to the target data indicator and obtain data analysis results according to the target data.

8. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 6 are implemented.

9. 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 6 are implemented.

10. 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 6 are implemented.

Images