Industrial chain analysis method, system and device, storage medium and program product

Abstract

The invention provides an industrial chain analysis method, system and device, a storage medium and a program product, and relates to the technical field of data analysis. The method comprises the steps that when it is detected that to-be-queried information is input into an industrial chain analysis model, intention recognition is conducted on the to-be-queried information, an intention recognition result is obtained, the intention recognition result at least comprises the name of a preset industrial chain, and the industrial chain analysis model is a large language model; the external database is a graph database constructed based on industrial chain data in a vector form; the intention recognition result is decomposed, a sub-query task set is obtained, and each sub-query task corresponds to the name of one industry chain in the intention recognition result; based on the target sub-query task, an analysis result of the target sub-query task is determined, and the target sub-query task is any one sub-query task; and determining an industrial chain analysis result corresponding to the to-be-queried information based on the analysis results of all the target sub-query tasks. According to the invention, the analysis accuracy can be improved.

Description

Technical Field

[0001] The present invention relates to the field of data analysis technology, and in particular to an industrial chain analysis method, system, device, storage medium and program product. Background Art

[0002] Industrial chain analysis is a method of comprehensively sorting out and studying the upstream and downstream links of a specific industry or product. Through a comprehensive analysis of the industrial chain, it can help companies better understand the overall operation of the industry, discover the key links and potential risks in the industrial chain, formulate reasonable development strategies, and improve the competitiveness and sustainable development capabilities of the company.

[0003] Traditional industrial chain analysis methods mainly rely on manual collection and organization of various industrial chain related data. This process is not only time-consuming and laborious, but also limited by personal knowledge and retrieval capabilities. It is often difficult to dig out the deep connections and potential information hidden behind a large amount of data. Especially for those complex industrial chains across industries and fields, it is difficult to comprehensively and accurately construct a complete analysis context with manpower alone.

[0004] At present, although there are analysis technologies based on graph databases, it is difficult to flexibly respond to diverse query requirements during the analysis process, which affects the accuracy of industrial chain analysis. Summary of the invention

[0005] The embodiments of the present invention provide an industrial chain analysis method, system, device, storage medium and program product to solve the problem of inaccurate analysis of current industrial chain analysis methods.

[0006] In a first aspect, an embodiment of the present invention provides an industrial chain analysis method, including:

[0007] When it is detected that the information to be queried is input into the industrial chain analysis model, the intent recognition is performed on the queried information to obtain the intent recognition result, wherein the intent recognition result at least includes the name of a preset industrial chain. The industrial chain analysis model is a large language model, and its external database is a graph database constructed based on the industrial chain data in vector form;

[0008] Decomposing the intent recognition result to obtain a sub-query task set, wherein the sub-query task set includes at least one sub-query task, and each sub-query task corresponds to the name of an industrial chain in the intent recognition result;

[0009] Based on the target sub-query task, determining the analysis result of the target sub-query task, wherein the target sub-query task is any sub-query task;

[0010] Based on the analysis results of all target sub-query tasks, determine the industrial chain analysis result corresponding to the information to be queried.

[0011] In a possible implementation, based on the target sub-query task, determine the analysis result of the target sub-query task, including:

[0012] If it is determined that there is an external knowledge base corresponding to the industrial chain of the target sub-query task, then determine the analysis result of the external knowledge base as the analysis result of the target sub-query task; otherwise, call the agent corresponding to the target sub-query task to analyze the target sub-query task, and determine the analysis result of the agent as the analysis result of the target sub-query task.

[0013] In a possible implementation, based on the target sub-query task, determine the analysis result of the target sub-query task, including:

[0014] Call the agent corresponding to the target sub-query task to analyze the target sub-query task, and determine the analysis result of the agent as the analysis result of the target sub-query task; where each sub-query task corresponds to an agent, and the external database of the agent is the same as the external database of the industrial chain analysis model.

[0015] In a possible implementation, the analysis result of the target sub-query task is stored in the data structure form of a relationship graph;

[0016] Based on the analysis results of all target sub-query tasks, determine the industrial chain analysis result corresponding to the information to be queried, including:

[0017] Perform a union operation on the analysis results of all target sub-query tasks to obtain the industrial chain relationship graph corresponding to the information to be queried.

[0018] In a possible implementation, the external database of the industrial chain analysis model is a knowledge graph formed by converting the preprocessed industrial chain data into vector form and storing the industrial chain data in vector form in the vector library of the graph database.

[0019] In a possible implementation, visually display the industrial chain analysis result corresponding to the information to be queried.

[0020] In a second aspect, an embodiment of the present invention provides an industrial chain analysis device, including:

[0021] An intention recognition module, which is used to recognize the intention of the information to be queried when it is detected that the information to be queried is input into the industrial chain analysis model, and obtain an intention recognition result. At least one name of a pre-set industrial chain is included in the intention recognition result. The industrial chain analysis model is a large language model, and its external database is a graph database constructed based on industrial chain data in vector form;

[0022] An intention decomposition module, which is used to decompose the intention recognition result to obtain a set of sub-query tasks. At least one sub-query task is included in the set of sub-query tasks, and each sub-query task corresponds to the name of an industrial chain in the intention recognition result;

[0023] A first analysis module, which is used to determine the analysis result of the target sub-query task based on the target sub-query task, where the target sub-query task is any one of the sub-query tasks;

[0024] A second analysis module, which is used to determine the industrial chain analysis result corresponding to the information to be queried based on the analysis results of all target sub-query tasks.

[0025] Thirdly, an embodiment of the present invention provides an industrial chain analysis system, including:

[0026] An external database construction module, which is used to collect industrial chain data, convert the collected industrial chain data into a vector format, and store it in the vector library of the graph database;

[0027] An intelligent retrieval and association analysis module, which is used to recognize the intention of the input information to be queried based on the industrial chain analysis model, and obtain an intention recognition result; at least one name of a pre-set industrial chain is included in the intention recognition result, and the external database of the industrial chain analysis model is constructed based on the graph database;

[0028] An intelligent agent task disassembling module, which is used to decompose the intention recognition result based on the intention recognition result into at least one sub-query task, and determine the analysis result of the target sub-query task based on the target sub-query task; each sub-query task corresponds to the name of an industrial chain in the intention recognition result, and the target sub-query task is any one of the sub-query tasks;

[0029] A result integration module, which is used to determine the industrial chain analysis result corresponding to the information to be queried based on the analysis results of all target sub-query tasks;

[0030] A visualization display and interaction module, which is used to visually display the industrial chain analysis result corresponding to the information to be queried.

[0031] Fourthly, an embodiment of the present invention provides an electronic device, including a memory and a processor. The memory stores a computer program, and when the processor executes the computer program, the method in the first aspect above or any possible implementation manner of the first aspect is implemented.

[0032] Fifthly, an embodiment of the present invention provides a computer-readable storage medium, which stores a computer program, and when the computer program is executed by a processor, the method in the first aspect above or any possible implementation manner of the first aspect is implemented.

[0033] Sixthly, an embodiment of the present invention provides a computer program product, including a computer program, and when the computer program is executed by a processor, the method in the first aspect above or any possible implementation manner of the first aspect is implemented.

[0034] In the embodiment of the present invention, first, when it is detected that information to be queried is input into the industrial chain analysis model, intention recognition is performed on the information to be queried to obtain an intention recognition result. Among them, the industrial chain analysis model is a large language model. In order to improve the accuracy of retrieval, a graph database constructed with industrial chain data in vector form is used as its external database, which can improve the retrieval accuracy of associated information. Then, the intention recognition result is decomposed to obtain a sub-query task set, so as to be able to flexibly handle diverse query requirements. Then, based on the target sub-query task, the analysis result of the target sub-query task is determined. Finally, based on the analysis results of all target sub-query tasks, the industrial chain analysis result corresponding to the information to be queried is determined. Through the analysis method provided by the present invention, automated query and association analysis can greatly shorten the time and cost of industrial chain analysis. By using vector form for the external database of the industrial chain analysis model, vectorized matching can improve the accuracy and comprehensiveness of data association. By decomposing the intention recognition result into a sub-query task set and analyzing each sub-query task separately, diverse query requirements can be more flexibly handled. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] Figure 1 is a flowchart of the implementation of the industrial chain analysis method provided by the embodiment of the present invention;

[0036] Figure 2 is a schematic structural diagram of the industrial chain analysis device provided by the embodiment of the present invention;

[0037] Figure 3 is a schematic structural diagram of the industrial chain analysis system provided by the embodiment of the present invention;

[0038] Figure 4 is a schematic diagram of the electronic device provided by the embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0039] The embodiments of the present invention will be described in detail below in conjunction with the accompanying drawings.

[0040] As described in the background art, although there are currently knowledge graphs built based on graph databases, which can improve the ability of data processing and relationship analysis to a certain extent, however, current graph databases focus on using graph structures for relationship traversal, and their adaptability to unstructured data or more complex query requirements is limited.

[0041] In addition, existing information retrieval systems often follow preset and fixed query processes, and it is difficult to flexibly respond to diverse query requirements. Especially for data that has not been associated in advance, traditional systems are often helpless. Moreover, existing industrial chain analysis methods often do not accurately associate industrial chain-related data, and it is difficult to discover deep associations and potential information hidden behind a large amount of data. This will also lead to deviations in analysis results and affect the accuracy of business decisions.

[0042] To solve the problems existing in current industrial chain analysis, the present invention provides an industrial chain analysis method, system, device, storage medium, and program product.

[0043] See Figure 1 , which shows the implementation flowchart of the industrial chain analysis method provided by the embodiments of the present invention, and is described in detail as follows:

[0044] S110. When it is detected that the information to be queried is input into the industrial chain analysis model, perform intent recognition on the information to be queried to obtain an intent recognition result.

[0045] The intent recognition result at least includes the name of at least one pre-set industrial chain, and the external database of the industrial chain analysis model is a graph database built based on industrial chain data in vector form.

[0046] In some embodiments, the external database of the industrial chain analysis model is a knowledge graph formed by converting pre-processed industrial chain data into vector form and storing the industrial chain data converted into vector form in the vector library of the graph database.

[0047] In this embodiment, the industrial chain analysis model is a large language model. When building the external database of the industrial chain analysis model, it is necessary to collect relevant data of the industrial chain from various data sources, such as enterprise databases, industrial dynamics, think tank views, industrial policies, industrial reports, etc. After collecting the industrial chain data, it is also necessary to perform pre-processing work on the collected data, such as cleaning, de-duplication, and standardization.

[0048] After preprocessing the industrial chain data, the preprocessed data is converted into vector form, and the industrial chain data in vector form is stored in the vector library of the Neo4j graph database for subsequent intelligent retrieval and analysis. By adopting advanced vectorization techniques, various types of information related to the industrial chain, such as enterprise databases, industrial dynamics, think tank views, industrial policies, industrial reports, and information of domain experts, etc., are converted into vector form and stored in the vector library of Neo4j, laying a solid foundation for subsequent intelligent retrieval and analysis.

[0049] The industrial chain analysis model is a large language model that can generate natural language text or understand the meaning of language text. In the present invention, by using the industrial chain data in vector form as the knowledge base of the industrial chain analysis model, the inherent parameters of the large language model can be supplemented, the business pertinence of the large language model can be improved, and the effect of fine-tuning the large language model can be achieved. Moreover, different from ordinary knowledge bases, in the present invention, the neo4j vector library is used as the external knowledge base of the artificial intelligence large model, and according to the natural language instruction input by the user, relevant industrial chain data is automatically retrieved and associated. Through vectorized matching, the accuracy and relevance of the retrieval results are enhanced, effectively solving the problem that it is difficult to mine deep associations by traditional methods, improving the retrieval accuracy and comprehensiveness of associated information, and improving the retrieval of indirect data.

[0050] When it is detected that information to be queried is input into the industrial chain analysis model, the industrial chain analysis model will perform intent recognition on the natural language of the information to be queried and obtain the intent recognition result.

[0051] When performing intent recognition, it is necessary to set the roles in the industrial chain analysis model, that is, it is necessary to extract the name of the industrial chain according to the information to be queried input. The name of the industrial chain needs to be restricted, and specifically which industrial chain names to select can be set according to the scenario. Exemplarily, the name of the industrial chain can only be selected from the following words: Xinchuang, integrated circuits, vehicle networking, biomedicine, traditional Chinese medicine, new energy, new materials, high-end equipment, automobiles, new energy vehicles, green petrochemicals, aerospace, and light industry. The industrial chain analysis model will then output the intent recognition result from the set words of the above industrial chain names according to the information to be queried input. The intent recognition result can be output in the form of json array information. For example, the output intent recognition result is {"data": [{"name": "Xinchuang"}, {"name": "high-end equipment"}]}

[0052] S120. Decompose the intent recognition result to obtain a sub-query task set.

[0053] The intention recognition result can be decomposed through designed prompt instructions. Since the intention recognition result includes at least the name of a pre-set industrial chain, the intention recognition result can be decomposed according to the number of industrial chain names included in the intention recognition result to obtain a sub-query task set. The sub-query task set includes at least one sub-query task, and each sub-query task corresponds to the name of an industrial chain in the intention recognition result. The number of all sub-query tasks in the sub-query task set is the same as the number of industrial chain names included in the intention recognition result.

[0054] Still taking the intention recognition result {"data": [{"name": "IT application innovation"}, {"name": "high-end equipment"}]} obtained in S110 as an example for illustration. It can be seen from the intention recognition result that it includes the names of 2 industrial chains, namely "IT application innovation" and "high-end equipment". Therefore, it can be decomposed into 2 sub-query tasks according to the number of industrial chain names in the intention recognition result, and the sub-query task set also includes 2 sub-query tasks.

[0055] By decomposing the intention recognition result, it is possible to more flexibly respond to diverse query requirements and further improve the accuracy of query results.

[0056] S130. Based on the target sub-query task, determine the analysis result of the target sub-query task.

[0057] Among them, the target sub-query task is any one sub-query task.

[0058] In some embodiments, if it is determined that the industrial chain corresponding to the target sub-query task has an external knowledge base corresponding to the industrial chain, the analysis result of the external knowledge base is determined as the analysis result of the target sub-query task. Otherwise, call the agent corresponding to the target sub-query task to analyze the target sub-query task, and determine the analysis result of the agent as the analysis result of the target sub-query task.

[0059] Among them, each sub-query task corresponds to an agent, and the external database of the agent is the same as the external database of the industrial chain analysis model.

[0060] When it is determined that the industrial chain corresponding to the target sub-query task has an external knowledge base corresponding to the industrial chain, it can be directly called without re-analyzing in the large database, which can improve the analysis efficiency. However, when the external knowledge base corresponding to the target sub-query task cannot be queried, the agent corresponding to the target sub-query task can be directly called to analyze the target sub-query task.

[0061] In some embodiments, the intelligent agent corresponding to the target sub-query task can be directly called to analyze the target sub-query task, and the analysis result of the intelligent agent can be determined as the analysis result of the target sub-query task.

[0062] In this embodiment, by setting a sub-query task for the name of each industrial chain and setting a corresponding intelligent agent for the name of each industrial chain, the system can flexibly respond to diverse query requirements and obtain analysis results more quickly and accurately.

[0063] In some embodiments, the intelligent agent is pre-encapsulated based on the pre-set roles in the industrial chain analysis model, and each industrial chain name included in the pre-set roles corresponds to an intelligent agent.

[0064] For the sake of easy understanding, the example in the above steps is still used to continue the explanation. Taking the name of the "IT application innovation" industrial chain as an example, the construction process of the intelligent agent corresponding to "IT application innovation" is introduced.

[0065] The policy prompt words for [IT application innovation] are designed as follows:

[0066] 1. Input information: What are the policies corresponding to the [IT application innovation] industrial chain?

[0067] 2. Requirement description: IT application innovation is one entity, the policy name is another entity, and the relationship name is [policy].

[0068] 3. Return format: {links: [{"sourceid": "IT application innovation", "targetid": "policy name 1", "label": "policy"}, {"sourceid": "IT application innovation", "targetid": "policy name 2", "label": "policy"}]}. It should be noted that the policy name must be a policy existing in the knowledge base.

[0069] The enterprise prompt words for [IT application innovation] are designed as follows:

[0070] 1. Input information: What are the enterprises associated with the [IT application innovation] industrial chain?

[0071] 2. Requirement description: [IT application innovation] is one entity, the enterprise name is another entity, and the relationship name is [enterprise]

[0072] 3. Return format: {links: [{"sourceid": "IT application innovation", "targetid": "enterprise name 1", "label": "enterprise"}, {"sourceid": "IT application innovation", "targetid": "enterprise name 2", "label": "enterprise"}]}.

[0073] For experts in [information technology innovation], innovation institutions, industry reports, industry trends, and industry links, there are corresponding similar prompts and designs for each relevant subtask, which will not be elaborated here.

[0074] Based on the above, an agent corresponding to the name of the industrial chain of "information technology innovation" can be constructed.

[0075] In some embodiments, the analysis results returned by the agent corresponding to each sub-query task are stored in the data structure of a relational graph.

[0076] In this embodiment, the analysis results of each agent can be pushed to the message server Kafka in the data structure of a relational graph.

[0077] S140. Based on the analysis results of all target sub-query tasks, determine the industrial chain analysis result corresponding to the information to be queried.

[0078] After obtaining the analysis results of each target sub-query task, a union operation can be performed on the analysis results of all target sub-query tasks, and the industrial chain relational graph corresponding to the information to be queried can be automatically spliced according to the node relationship.

[0079] In addition, when the analysis results of each agent are pushed to the message server Kafka in the data structure of a relational graph, these sub-task results can be obtained from Kafka, and a complete industrial chain relational graph can be automatically spliced according to the node relationship for business personnel to adjust and improve.

[0080] In some embodiments, the industrial chain analysis result corresponding to the information to be queried can also be visually displayed. By providing an intuitive visual interface, the industrial chain relational graph is displayed, and business personnel are supported to perform dynamic adjustment and improvement. At the same time, rich interactive functions are provided to facilitate business personnel to deeply understand the industrial chain structure and discover potential risks and opportunities.

[0081] In the embodiments of the present invention, first, when it is detected that the information to be queried is input into the industrial chain analysis model, the intention of the information to be queried is recognized to obtain an intention recognition result. Among them, the industrial chain analysis model is a large language model. To improve the accuracy of retrieval, a graph database constructed with industrial chain data in vector form is used as its external database, which can improve the retrieval accuracy of associated information. Then, the intention recognition result is decomposed to obtain a sub-query task set, so as to be able to flexibly handle diverse query requirements. Then, based on the target sub-query task, the analysis result of the target sub-query task is determined. Finally, based on the analysis results of all target sub-query tasks, the industrial chain analysis result corresponding to the information to be queried is determined. Through the analysis method provided by the present invention, automated query and correlation analysis can greatly shorten the time and cost of industrial chain analysis. By using the vector form for the external database of the industrial chain analysis model, vectorized matching can improve the accuracy and comprehensiveness of data association. By decomposing the intention recognition result into a sub-query task set and analyzing each sub-query task separately, diverse query requirements can be handled more flexibly.

[0082] It should be understood that the magnitudes of the sequence numbers of the steps in the above embodiments do not mean the order of execution is prior or posterior. The execution order of each process should be determined according to its function and internal logic, and should not constitute any limitation to the implementation process of the embodiments of the present invention.

[0083] The following are the device embodiments of the present invention. For the details not described in detail therein, reference can be made to the corresponding method embodiments above.

[0084] Figure 2 The structural schematic diagram of the industrial chain analysis device provided by the embodiments of the present invention is shown. For the convenience of description, only the parts related to the embodiments of the present invention are shown and are described in detail as follows:

[0085] As Figure 2 shown, an industrial chain analysis device 200 includes:

[0086] An intention recognition module 210, configured to recognize the intention of the information to be queried to obtain an intention recognition result when it is detected that the information to be queried is input into the industrial chain analysis model. Among them, the intention recognition result includes at least the name of a preset industrial chain, and the external database of the industrial chain analysis model is a vector library constructed based on a graph database;

[0087] An intention decomposition module 220, configured to decompose the intention recognition result to obtain a sub-query task set. Among them, the sub-query task set includes at least one sub-query task, and each sub-query task corresponds to the name of an industrial chain in the intention recognition result;

[0088] The first analysis module 230 is configured to determine the analysis result of a target sub-query task based on the target sub-query task, where the target sub-query task is any one sub-query task;

[0089] The second analysis module 240 is configured to determine the industrial chain analysis result corresponding to the information to be queried based on the analysis results of all target sub-query tasks.

[0090] In a possible implementation manner, the first analysis module 230 is configured to, if it is determined that there is an external knowledge base corresponding to the industrial chain of the target sub-query task, determine the analysis result of the external knowledge base as the analysis result of the target sub-query task; otherwise, call the agent corresponding to the target sub-query task to analyze the target sub-query task, and determine the analysis result of the agent as the analysis result of the target sub-query task.

[0091] In a possible implementation manner, the first analysis module 230 is configured to call the agent corresponding to the target sub-query task to analyze the target sub-query task, and determine the analysis result of the agent as the analysis result of the target sub-query task; where each sub-query task corresponds to an agent, and the external database of the agent is the same as the external database of the industrial chain analysis model.

[0092] In a possible implementation manner, the analysis result of the target sub-query task is stored in the form of a data structure of a relationship graph;

[0093] The second analysis module 240 is configured to perform a union operation on the analysis results of all target sub-query tasks to obtain the industrial chain relationship graph corresponding to the information to be queried.

[0094] In a possible implementation manner, the external database of the industrial chain analysis model is a knowledge graph formed by converting the preprocessed industrial chain data into a vector form and storing the industrial chain data converted into a vector form in the vector library of the graph database.

[0095] In a possible implementation manner, the second analysis module 240 is configured to perform a visual display on the industrial chain analysis result corresponding to the information to be queried.

[0096] In addition, the present invention also provides an industrial chain analysis system. For the sake of convenience of description, only the parts related to the embodiments of the present invention are shown and are described in detail as follows:

[0097] Figure 3 Fig. shows an industrial chain analysis system provided by an embodiment of the present invention. The industrial chain analysis system 300 includes:

[0098] The external database construction module 310 is used to collect industrial chain data, convert the collected industrial chain data into vector format, and store it in the vector library of the graph database;

[0099] The intelligent retrieval and correlation analysis module 320 is used to perform intent recognition on the input information to be queried based on the industrial chain analysis model to obtain an intent recognition result; wherein, the intent recognition result includes at least the name of a pre-set industrial chain, and the external database of the industrial chain analysis model is constructed based on the graph database;

[0100] The intelligent agent task decomposition module 330 is used to decompose the intent recognition result based on the intent recognition result into at least one sub-query task, and determine the analysis result of the target sub-query task based on the target sub-query task; wherein, each sub-query task corresponds to the name of an industrial chain in the intent recognition result, and the target sub-query task is any one of the sub-query tasks;

[0101] The result integration module 340 is used to determine the industrial chain analysis result corresponding to the information to be queried based on the analysis results of all target sub-query tasks;

[0102] The visualization display and interaction module 350 is used to visually display the industrial chain analysis result corresponding to the information to be queried.

[0103] In a possible implementation manner, the external database construction module 310 includes a data collection and preprocessing module and a data vectorization and storage module. The data collection and preprocessing module is used to collect industrial chain-related data from various data sources, such as enterprise databases, industry dynamics, think tank views, industrial policies, industrial reports, etc., and perform preprocessing work such as cleaning, deduplication, and standardization. The data vectorization and storage module: converts the preprocessed data into vector form and stores it in the vector library of the Neo4j graph database for subsequent intelligent retrieval and analysis.

[0104] In addition, the intelligent retrieval and correlation analysis module 320 in this embodiment has the same function as the intent recognition module 210 in the second aspect, and the intelligent agent task decomposition module 330 has the same functions as the intent decomposition module 220 and the first analysis module 230 in the second aspect, and will not be elaborated here.

[0105] Figure 4 It is a schematic diagram of an electronic device provided by an embodiment of the present invention. As Figure 4 shown, the electronic device 4 in this embodiment includes: a processor 40 and a memory 41. The memory 41 stores a computer program 42. When the processor 40 executes the computer program 42, the steps in the above-mentioned various method embodiments are implemented. Or, when the processor 40 executes the computer program 42, the functions of each module / unit in the above-mentioned various device embodiments are implemented.

[0106] Exemplarily, the computer program 42 can be divided into one or more modules / units, which are stored in the memory 41 and executed by the processor 40 to implement the present invention. The one or more modules / units can be a series of computer program instruction segments capable of performing specific functions, and these instruction segments are used to describe the execution process of the computer program 42 in the electronic device 4.

[0107] The electronic device 4 may include, but is not limited to, a processor 40 and a memory 41. Those skilled in the art can understand that Figure 4 merely being examples of the electronic device 4, they do not constitute a limitation on the electronic device 4. It may include more or fewer components than shown in the figure, or combine certain components, or have different components. For example, the electronic device 4 may also include input / output devices, network access devices, buses, etc.

[0108] The processor 40 may be a central processing unit (CPU), or may also be other general-purpose processors, digital signal processors (DSPs), application specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or the processor may also be any conventional processor, etc.

[0109] The memory 41 may be an internal storage unit of the electronic device 4, such as the hard disk or memory of the electronic device 4. The memory 41 may also be an external storage device of the electronic device 4, such as a plug-in hard disk, a smart media card (SMC), a secure digital (SD) card, a flash card, etc. equipped on the electronic device 4. Further, the memory 41 may also include both the internal storage unit and the external storage device of the electronic device 4. The memory 41 is used to store the computer program 42 and other programs and data required by the electronic device 4. The memory 41 may also be used to temporarily store data that has been output or is to be output.

[0110] For the convenience and brevity of description, only the above-mentioned division of each functional module / unit is used as an example for illustration. In actual applications, the above functions can be assigned to different functional modules / units according to needs. The above modules / units can be implemented in the form of hardware, software, or a combination of hardware and software.

[0111] An embodiment of the present invention further provides a computer-readable storage medium, which stores a computer program. When the computer program is executed by a processor, the methods in the above method embodiments are implemented.

[0112] An embodiment of the present invention further provides a computer program product, including a computer program. When the computer program is executed by a processor, the methods in the above method embodiments are implemented.

[0113] Among them, the computer program includes computer program code, and the computer program code can be in the form of source code, object code, executable file, or some intermediate form, etc. The computer-readable medium can include: any entity or device capable of carrying the computer program code, recording medium, USB flash drive, mobile hard disk, magnetic disk, optical disk, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signal, telecommunication signal, and software distribution medium, etc.

[0114] In the above embodiments, the descriptions of each embodiment have their own focuses. For the parts not detailed or recorded in a certain embodiment, reference can be made to the relevant descriptions of other embodiments. Without special instructions and logical conflicts, the terms and / or descriptions between different embodiments are consistent and can be mutually referred to. The technical features in different embodiments can be combined to form new embodiments according to their internal logical relationships.

[0115] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that: they can still modify the technical solutions recorded in the foregoing embodiments, or perform equivalent replacements on some of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be included in the protection scope of the present invention.

Claims

1. A method for analyzing an industrial chain, characterized in that: include: When it is detected that information to be queried is input into the industrial chain analysis model, the information to be queried is subjected to intent recognition to obtain an intent recognition result, wherein the intent recognition result includes at least a pre-set name of an industrial chain; the industrial chain analysis model is a large language model, and its external database is a graph database constructed based on industrial chain data in vector form; Decomposing the intention recognition result to obtain a sub-query task set, wherein the sub-query task set includes at least one sub-query task, and each sub-query task corresponds to the name of an industrial chain in the intention recognition result; Based on the target sub-query task, determining the analysis result of the target sub-query task, wherein the target sub-query task is any sub-query task; Based on the analysis results of all the target sub-query tasks, the industrial chain analysis result corresponding to the information to be queried is determined.

2. The industrial chain analysis method according to claim 1, characterized in that: The determining, based on the target sub-query task, an analysis result of the target sub-query task includes: If it is determined that the industrial chain corresponding to the target sub-query task is equipped with an external knowledge base corresponding to the industrial chain, the analysis result of the external knowledge base is determined as the analysis result of the target sub-query task; otherwise, the intelligent agent corresponding to the target sub-query task is called to analyze the target sub-query task, and the analysis result of the intelligent agent is determined as the analysis result of the target sub-query task; wherein each sub-query task corresponds to an intelligent agent, and the external database of the intelligent agent is the same as the external database of the industrial chain analysis model.

3. The industrial chain analysis method according to claim 1, characterized in that: The determining, based on the target sub-query task, an analysis result of the target sub-query task includes: Call the intelligent agent corresponding to the target sub-query task to analyze the target sub-query task, and determine the analysis result of the intelligent agent as the analysis result of the target sub-query task; wherein each sub-query task corresponds to an intelligent agent, and the external database of the intelligent agent is the same as the external database of the industrial chain analysis model.

4. The industrial chain analysis method according to claim 1, characterized in that: The analysis result of the target sub-query task is stored in the form of a data structure of a relationship graph; The step of determining the industrial chain analysis result corresponding to the information to be queried based on the analysis results of all the target sub-query tasks includes: A union operation is performed on the analysis results of all the target sub-query tasks to obtain an industrial chain relationship diagram corresponding to the information to be queried.

5. The industrial chain analysis method according to any one of claims 1 to 4, characterized in that: The external database of the industrial chain analysis model is based on a knowledge graph formed by converting pre-processed industrial chain data into vector form and storing the industrial chain data converted into vector form in a vector library of a graph database.

6. The industrial chain analysis method according to any one of claims 1 to 4, characterized in that: The analysis method further comprises: The industrial chain analysis results corresponding to the information to be queried are displayed visually.

7. An industrial chain analysis system, characterized in that: include: An external database construction module is used to collect industrial chain data, convert the collected industrial chain data into vector format, and store it in the vector library of the graph database; The intelligent retrieval and association analysis module is used to perform intent recognition on the input information to be queried based on the industrial chain analysis model to obtain an intent recognition result; wherein the intent recognition result includes at least the name of a pre-set industrial chain, the industrial chain analysis model is a large language model, and its external database is a vector library built based on a graph database; An agent task decomposition module is used to decompose the intention recognition result based on the intention recognition result into at least one sub-query task, and determine the analysis result of the target sub-query task based on the target sub-query task; wherein each sub-query task corresponds to the name of an industrial chain in the intention recognition result, and the target sub-query task is any sub-query task; A result integration module, used to determine the industrial chain analysis result corresponding to the information to be queried based on the analysis results of all the target sub-query tasks; The visualization and interaction module is used to visualize the industrial chain analysis results corresponding to the information to be queried.

8. An electronic device, characterized in that: The method comprises a memory and a processor, wherein the memory stores a computer program, and the processor implements the method according to any one of claims 1 to 6 when executing the computer program.

9. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the method according to any one of claims 1 to 6 is implemented.

10. A computer program product, characterized in that The method comprises a computer program, which implements the method according to any one of claims 1 to 6 when being executed by a processor.

Images