Demand analysis generation method and device based on knowledge graph and large model

Abstract

The invention provides a demand analysis generation method and device based on a knowledge graph and a large model. The method comprises the steps that a demand analysis appeal of a user for a target system is obtained, a programming large model is adopted to convert the demand analysis appeal into a graph query language, query processing is conducted on the graph query language in a knowledge graph, an analysis sub-graph is obtained, and the demand analysis appeal represents an analysis appeal for modifying or querying the target system; the demand analysis appeal is retrieved according to a system detailed arrangement knowledge base, system detailed arrangement documents corresponding to the demand analysis appeal are obtained, and the system detailed arrangement knowledge base comprises system detailed arrangement documents of a plurality of target systems; and inputting the system detailed document and the analysis subgraph into the large reasoning model to obtain a demand analysis report corresponding to the demand analysis appeal. The problem that the preset function point cannot cover the demand scene in the demand analysis process when the business demand evaluation is carried out through the preset function point in the prior art is solved.

Description

Technical Field

[0001] This application relates to the field of requirements analysis technology, and more specifically, to a requirements analysis generation method, apparatus, computer-readable storage medium, and electronic device based on knowledge graphs and large models. Background Technology

[0002] As key participants in the market economy, banks face competitive pressure from internet finance. Market demands are constantly changing, requiring them to adapt their services and products accordingly. However, unlike other organizations, banks must prioritize risk control. Enterprise-level demand modeling can enable rapid response to market demands while ensuring stable business processes, manageable risks, and technological stability. Enterprise-level business models, through a five-level modeling system—domain, value stream, activity, task, and step—form model assets. While this allows for standardized abstraction of business elements, dynamically adapting model assets to demand analysis scenarios that are complex, dynamic, and require rapid response remains a challenge.

[0003] Existing technologies based on enterprise-level business models, which assess business requirements through pre-defined functional points, suffer from a lack of flexibility. These pre-defined functional points cannot be maintained in advance, and during requirements analysis, they may fail to cover all required scenarios. Furthermore, the lack of interrelationships between functional points necessitates manual aggregation for assessing the scope of impact during requirements analysis. Summary of the Invention

[0004] The main purpose of this application is to provide a method, apparatus, computer-readable storage medium and electronic device for generating requirements analysis based on knowledge graphs and large models, so as to at least solve the problem that existing technologies, which evaluate business requirements by using preset functional points, may encounter the problem that the preset functional points cannot cover the requirement scenarios during the requirements analysis process.

[0005] To achieve the above objectives, according to one aspect of this application, a method for generating requirements analysis based on knowledge graphs and large models is provided, comprising: obtaining user requirements analysis requests for a target system; converting the requirements analysis requests into a graph query language using a programming large model; querying and processing the graph query language in a knowledge graph to obtain an analysis subgraph, wherein the requirements analysis requests represent analysis requests for modification or querying of the target system; retrieving and processing the requirements analysis requests according to a system detailed design knowledge base to obtain a system detailed design document corresponding to the requirements analysis requests, wherein the system detailed design knowledge base includes multiple system detailed design documents of the target system; and inputting the system detailed design document and the analysis subgraph into a reasoning large model to obtain a requirements analysis report corresponding to the requirements analysis requests.

[0006] Optionally, before retrieving and processing the requirements analysis requests based on the system detailed design knowledge base, the method further includes: acquiring system information related to the target system, wherein the system information includes the design specification and requirements specifications for each stage of the target system; dividing the system information into blocks based on the minimum functional dimension of the target system to obtain multiple blocks of data, and using a vector model to establish a vector index for each block of data to obtain multiple system detailed design documents; and constructing the system detailed design knowledge base based on the multiple system detailed design documents.

[0007] Optionally, the system detailed design document is obtained by retrieving and processing the requirements analysis request based on the system detailed design knowledge base, including: standardizing and rewriting the requirements analysis request based on the language big model under the constraints of prompting engineering to obtain a text rewritten request; converting the text rewritten request into a vector form to obtain a text request vector; and retrieving and processing the text request vector based on the system detailed design knowledge base to obtain the system detailed design document.

[0008] Optionally, the system detailed design document and the analysis subgraph are input into the inference model to obtain a requirements analysis report corresponding to the requirements analysis requirements. This includes: inputting the analysis subgraph in JSON text format into the inference model to obtain a requirements analysis logic chain; and inputting the requirements analysis logic chain and the system detailed design document into the inference model to obtain the requirements analysis report corresponding to the requirements analysis requirements.

[0009] Optionally, after obtaining the requirements analysis report corresponding to the requirements analysis requirements, the method further includes: verifying the analysis report using the analysis subgraph, the system detailed design document, and the requirements analysis logic chain to obtain verification results, and updating the inference big model based on the verification results, wherein the analysis subgraph includes analysis subgraphs in JSON text form and analysis subgraphs in graphical form.

[0010] Optionally, before performing query processing on the graph query language in the knowledge graph, the method further includes: constructing the knowledge graph based on the functional modules of the target system, wherein the types of the vertices of the knowledge graph include at least one of domain type, value stream type, activity type, task type, and step type; the vertex information of the knowledge graph includes at least one of attributes, name, description, scope, stakeholders, products, channels, partners, roles, and decision conditions; and the relationships between the vertices of the knowledge graph include at least one of inclusion relationship and execution sequence relationship.

[0011] Optionally, before converting the requirements analysis requests into a graph query language using a large programming model, the method further includes: optimizing the large programming model using model optimization techniques to obtain an optimized large programming model, wherein the model optimization techniques include knowledge distillation, reinforcement learning, and prompting engineering.

[0012] According to another aspect of this application, a requirement analysis generation device based on knowledge graphs and large models is provided, comprising: a query processing unit, configured to acquire user requirement analysis requests for a target system, convert the requirement analysis requests into a graph query language, and perform query processing on the graph query language in a knowledge graph to obtain an analysis subgraph, wherein the requirement analysis requests represent requests for modification or querying of the target system; a retrieval processing unit, configured to retrieve and process the requirement analysis requests according to a system detailed design knowledge base to obtain a system detailed design document corresponding to the requirement analysis requests, wherein the system detailed design knowledge base includes multiple system detailed design documents of the target system; and an input unit, configured to input the system detailed design document and the analysis subgraph into a large inference model to obtain a requirement analysis report corresponding to the requirement analysis requests.

[0013] According to another aspect of this application, a computer-readable storage medium is provided, the computer-readable storage medium including a stored program, wherein, when the program is executed, it controls the device where the computer-readable storage medium is located to execute any of the aforementioned requirements analysis and generation methods based on knowledge graphs and large models.

[0014] According to another aspect of this application, an electronic device is provided, comprising: one or more processors, a memory, and one or more programs, wherein the one or more programs are stored in the memory and configured to be executed by the one or more processors, the one or more programs including methods for performing any of the described knowledge graph and large model-based requirements analysis generation methods.

[0015] By applying the technical solution of this application, user requirements for the target system are obtained. A large-scale programming model is used to convert these requirements into a graph query language, which is then queried and processed within a knowledge graph to obtain an analysis subgraph. Each requirement represents a modification or query request for the target system. The requirements are then retrieved and processed from a system detailed design knowledge base, yielding corresponding system detailed design documents. This knowledge base includes detailed design documents for multiple target systems. The system detailed design documents and the analysis subgraph are input into a large-scale reasoning model to obtain a requirements analysis report corresponding to the requirements. This requirement analysis and evaluation based on knowledge graphs and large-scale models provides efficient and accurate query and reasoning capabilities, displaying intuitive, interpretable, and causal reasoning-based analysis results. For target systems with complex business logic and frequently changing requirements, this approach reduces decision-making risks associated with complex problems, thereby improving the quality and efficiency of requirements analysis. It also solves the problem in existing technologies where pre-defined functional points cannot cover all requirement scenarios during the requirements analysis process. Attached Figure Description

[0016] The accompanying drawings, which form part of this application, are used to provide a further understanding of this application. The illustrative embodiments and descriptions of this application are used to explain this application and do not constitute an undue limitation of this application. In the drawings:

[0017] Figure 1 A hardware structure block diagram of a mobile terminal that performs a knowledge graph and large model-based requirements analysis generation method is shown in an embodiment of this application.

[0018] Figure 2 A flowchart illustrating a requirement analysis generation method based on knowledge graphs and large models according to an embodiment of this application is shown.

[0019] Figure 3 A flowchart illustrating a specific knowledge graph and large model-based requirement analysis generation method provided according to an embodiment of this application is shown.

[0020] Figure 4 A structural block diagram of a requirement analysis generation device based on knowledge graphs and large models, according to an embodiment of this application, is shown. Detailed Implementation

[0021] It should be noted that, unless otherwise specified, the embodiments and features described in this application can be combined with each other. This application will now be described in detail with reference to the accompanying drawings and embodiments.

[0022] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present application, and not all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative effort should fall within the scope of protection of the present application.

[0023] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate for the embodiments of this application described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0024] As described in the background section, existing technologies assess business requirements by pre-setting functional points. However, during the requirements analysis process, these pre-set functional points may not cover all requirements scenarios. To address this issue, embodiments of this application provide a requirements analysis generation method, apparatus, computer-readable storage medium, and electronic device based on knowledge graphs and large models.

[0025] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention.

[0026] The methods and embodiments provided in this application can be executed on a mobile terminal, computer terminal, or similar computing device. Taking running on a mobile terminal as an example, Figure 1 This is a hardware structure block diagram of a mobile terminal based on a knowledge graph and large model-based requirement analysis and generation method, according to an embodiment of the present invention. Figure 1 As shown, a mobile terminal may include one or more ( Figure 1 Only one is shown in the diagram. A processor 102 (which may include, but is not limited to, a microprocessor MCU or a programmable logic device FPGA, etc.) and a memory 104 for storing data are also shown. The mobile terminal may further include a transmission device 106 for communication functions and an input / output device 108. Those skilled in the art will understand that... Figure 1The structure shown is for illustrative purposes only and does not limit the structure of the mobile terminal described above. For example, the mobile terminal may also include components that are more... Figure 1 The more or fewer components shown, or having the same Figure 1 The different configurations shown.

[0027] The memory 104 can be used to store computer programs, such as application software programs and modules, like the computer program corresponding to the knowledge graph and large model-based requirement analysis generation method in this embodiment of the invention. The processor 102 executes various functional applications and data processing by running the computer programs stored in the memory 104, thereby implementing the above-described method. The memory 104 may include high-speed random access memory and non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some instances, the memory 104 may further include memory remotely located relative to the processor 102, and these remote memories can be connected to the mobile terminal via a network. Examples of the aforementioned networks include, but are not limited to, the Internet, corporate intranets, local area networks, mobile communication networks, and combinations thereof. The transmission device 106 is used to receive or send data via a network. Specific examples of the aforementioned networks may include wireless networks provided by the mobile terminal's communication provider. In one example, the transmission device 106 includes a Network Interface Controller (NIC), which can be connected to other network devices via a base station to communicate with the Internet. In one example, the transmission device 106 may be a radio frequency (RF) module, which is used to communicate with the Internet wirelessly.

[0028] This embodiment provides a knowledge graph and large model-based requirement analysis generation method that runs on a mobile terminal, computer terminal or similar computing device. It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions. Also, although the logical order is shown in the flowchart, in some cases the steps shown or described can be executed in a different order than that shown here.

[0029] Figure 2 This is a flowchart of a requirement analysis and generation method based on knowledge graphs and large models according to an embodiment of this application. For example... Figure 2 As shown, the method includes the following steps:

[0030] Step S201: Obtain the user's requirements analysis requests for the target system, use a large programming model to convert the above requirements analysis requests into a graph query language, and perform query processing on the above graph query language in the knowledge graph to obtain an analysis subgraph, wherein the above requirements analysis requests represent the analysis requests for modification or query of the above target system.

[0031] Large-scale programming models refer to AI models that have undergone extensive training and are capable of understanding and generating program code. These models can handle complex code structures and programming language features, are typically based on deep learning frameworks such as the Transformer architecture, and possess a large number of parameters and powerful natural language processing capabilities. Through large-scale programming models, requirements analysts only need to describe their needs using natural language, and the model can understand and convert them into graph query language, thereby automatically extracting the required information from the knowledge graph. This greatly simplifies the operations of professionals and reduces the time spent on manual queries and writing query code.

[0032] Specifically, the large programming model, through deep learning technology, can accurately understand the user's requirements and translate them into a graph query language. This language is specifically designed for efficient searching on knowledge graphs. As a high-level form of semantic networks, knowledge graphs contain rich information about all key components of the target system and their relationships. Through precise matching using the graph query language, the subgraphs related to the requirements, i.e., the analysis subgraphs, can be quickly located.

[0033] Step S202: Retrieve and process the above-mentioned requirements analysis requests based on the system detailed design knowledge base to obtain the system detailed design documents corresponding to the above-mentioned requirements analysis requests. The system detailed design knowledge base includes system detailed design documents of multiple target systems.

[0034] Specifically, the system detailed design knowledge base retrieval further provides detailed system design documents, which are carefully organized and cover all aspects of the target system, including but not limited to architecture design, module interfaces, and data flow.

[0035] Step S203: Input the above system detailed design document and the above analysis sub-graph into the reasoning big model to obtain the requirement analysis report corresponding to the above requirement analysis requirements.

[0036] Among them, large-scale reasoning models are AI models capable of handling logical reasoning, causal relationship analysis, and complex decision-making problems. They are typically based on pre-trained language models, fine-tuned or customized to adapt to specific domains or tasks, and possess strong comprehension and reasoning abilities.

[0037] Specifically, the comprehensive analysis of sub-graphs and system design documents using the large-scale reasoning model generates a thorough and in-depth requirements analysis report. This report not only covers the specific scope of impact of the requirements but also predicts the consequences of changes, thereby assisting decision-makers in making more accurate judgments. Through the synergistic effect of this series of steps, this solution greatly simplifies the requirements analysis process, improves the accuracy of the analysis, and reduces decision-making risks.

[0038] This embodiment, by applying steps S201, S202, and S203, utilizes knowledge graph and large-scale model technology for requirement analysis and evaluation. This provides efficient and accurate query and reasoning capabilities for requirement analysis, displaying intuitive, interpretable analysis results with causal reasoning. For target systems with complex business logic and frequently changing requirements, requirement analysis and evaluation using knowledge graph and large-scale model technology can reduce decision-making risks for complex problems, thereby improving the quality and efficiency of requirement analysis. It solves the problem in existing technologies where pre-defined functional points cannot cover all requirement scenarios during requirement analysis.

[0039] In the specific implementation process, before retrieving and processing the above-mentioned requirements analysis requests based on the system detailed design knowledge base, the above method also includes: obtaining system information related to the above-mentioned target system, wherein the above-mentioned system information includes the design specification of the above-mentioned target system and the requirement specification specifications for each stage; dividing the above-mentioned system information into blocks according to the minimum functional dimension of the above-mentioned target system to obtain multiple blocks of data, and using a vector model to establish a vector index for each of the above-mentioned blocks of data to obtain multiple above-mentioned system detailed design documents; and constructing the above-mentioned system detailed design knowledge base based on the multiple above-mentioned system detailed design documents.

[0040] This method constructs and optimizes a detailed system design knowledge base by collecting and organizing various design documents of the target system, such as design specifications and requirements specifications. Acquiring system information is fundamental, encompassing the entire system design, including but not limited to functional descriptions, architecture design, module interfaces, and data flows. Next, a minimum functional dimension-based block processing approach is adopted, breaking down large documents into several smaller blocks around specific functions, which facilitates subsequent vector modeling. The establishment of the vector model is a crucial step, utilizing natural language processing technology to convert textual information into numerical vectors, facilitating computer understanding and retrieval. In this way, each block of data is assigned a vector representation, forming the core component of the detailed system design knowledge base. The construction of the detailed system design knowledge base not only integrates scattered system information but also accelerates the retrieval process through vector indexing. This enables rapid location of relevant design documents when facing complex system requirements analysis, significantly improving the efficiency and accuracy of requirements analysis, reducing the time cost of manual review, and providing strong data support for requirements analysts.

[0041] Specifically, the system detailed design knowledge base is used to retrieve and process the above-mentioned requirements analysis requests to obtain the system detailed design documents corresponding to the above-mentioned requirements analysis requests. This includes: standardizing and rewriting the above-mentioned requirements analysis requests based on the language big model under the constraints of prompting engineering to obtain text rewritten requests; converting the above-mentioned text rewritten requests into vector form to obtain text request vectors; and retrieving and processing the above-mentioned text request vectors based on the system detailed design knowledge base to obtain the above-mentioned system detailed design documents.

[0042] This method solves the matching problem between requirements analysis requests and system specification documents by introducing a large language model and vector retrieval technology. Large language models, such as BERT or GPT series, can standardize and rewrite users' original, unstructured requirements analysis requests under the guidance of prompting engineering, transforming them into text rewritten requests in a unified format. This step eliminates comprehension barriers caused by differences in language expression. Subsequently, the text rewritten requests are converted into vector form, i.e., text request vectors, which is a prerequisite for vector retrieval. Vector retrieval technology utilizes a pre-built system specification document vector index, calculating the similarity between the text request vector and the document vectors in the index to quickly locate the most relevant system specification documents. This process not only improves retrieval speed but also ensures the accuracy of retrieval results because the similarity calculation considers semantic matching rather than simple keyword matching. Through this combination of technologies, this solution overcomes the ambiguity and inefficiency problems of traditional document retrieval, providing requirements analysts with timely and accurate information retrieval services, significantly improving the quality and efficiency of requirements analysis work.

[0043] More specifically, the above-mentioned system detailed design document and the above-mentioned analysis sub-graph are input into the above-mentioned inference model to obtain a requirements analysis report corresponding to the above-mentioned requirements analysis requirements, including: inputting the analysis sub-graph in JSON text format into the above-mentioned inference model to obtain a requirements analysis logic chain; inputting the above-mentioned requirements analysis logic chain and the above-mentioned system detailed design document into the above-mentioned inference model to obtain the above-mentioned requirements analysis report corresponding to the above-mentioned requirements analysis requirements.

[0044] This method, through the application of a large-scale inference model, achieves a deep understanding of requirements analysis needs and intelligent parsing of system design documents, solving the problem of complex logical reasoning in requirements analysis report generation. The large-scale inference model, such as a large-scale pre-trained model based on the Transformer architecture, can handle complex relationships and logic. When it receives an analysis subgraph in JSON text format, it automatically constructs a requirements analysis logic chain—a logical structure that clearly shows the causal relationship between requirements analysis needs and system components. Subsequently, the requirements analysis logic chain is used as input, and the system design document is input into the large-scale inference model. Based on the content of the design document, the model supplements and improves the details in the logic chain, ultimately generating a comprehensive and accurate requirements analysis report. This process not only covers the specific scope of impact of the requirements but also predicts possible consequences of changes, providing decision-makers with intuitive and interpretable analysis results. Through the intelligent processing of the large-scale inference model, this solution overcomes the manual dependence in traditional requirements analysis report generation, improves report quality and generation efficiency, reduces decision-making risks, and ensures the smooth implementation of requirements changes.

[0045] Furthermore, after obtaining the requirements analysis report corresponding to the above requirements analysis requirements, the above method also includes: using the above analysis subgraph, the above system detailed design document and the above requirements analysis logic chain to verify the above analysis report, obtain the verification result, and update the above reasoning big model according to the above verification result, wherein the above analysis subgraph includes analysis subgraph in JSON text form and analysis subgraph in graphical form.

[0046] This method, by introducing a verification mechanism and model update strategy, ensures the accuracy of the requirements analysis report and the continuous optimization of the inference model, resolving long-standing issues with model generalization ability and adaptability. After generating the requirements analysis report, the system automatically verifies it from multiple perspectives using analysis subgraphs, detailed system design documents, and the requirements analysis logic chain, checking whether the report content conforms to the actual system design and whether there are any omissions or errors. The verification results provide feedback for model iteration; if the report has deviations, the system adjusts the parameters of the inference model to make it more accurate in future requirements analyses. This closed-loop mechanism not only improves the reliability of the report but also promotes the learning and growth of the inference model, enabling it to better adapt to constantly changing business needs, reduce errors in future analyses, and ensure continuous improvement and efficient operation of requirements analysis work.

[0047] Furthermore, before performing the graph query language query processing in the knowledge graph, the method further includes: constructing the knowledge graph based on the functional modules of the target system, wherein the types of the vertices of the knowledge graph include at least one of domain type, value stream type, activity type, task type, and step type; the vertex information of the knowledge graph includes at least one of attributes, name, description, scope, stakeholders, products, channels, partners, roles, and decision conditions; and the relationships between the vertices of the knowledge graph include at least one of inclusion relationship and execution sequence relationship.

[0048] Domain Type: Represents a macro-level classification of the business, such as credit card business, savings business, etc. Information at the top of the domain type list includes domain name, domain description, domain scope, stakeholders involved in the domain, products, channels, partners, etc.

[0049] Value stream type: This represents a specific process that provides value to customers, such as a credit card application process or a savings account opening process. Information at the apex of a value stream type includes the process name, process description, activities, tasks and steps involved, process scope, stakeholders, products, channels, and partners.

[0050] Activity Type: Represents a set of one or more tasks, typically corresponding to a stage in a business process. Information at the Activity Type vertex includes the activity name, description, scope, involved tasks, stakeholders, products, channels, and partners.

[0051] Task Type: Represents the specific operation or activity that needs to be performed in the business process, such as inputting customer information or conducting credit assessment. Information at the top of the task type list includes task name, task description, task scope, execution steps, stakeholders, product, channel, partners, and decision-making conditions.

[0052] Step type: Represents the smallest unit of operation when completing a task, such as the sub-task of entering customer information like "name" or "address". Information at the step type vertex includes step name, step description, step scope, stakeholders, product, channel, partner, and decision conditions.

[0053] This method, by constructing a detailed and structured knowledge graph, provides a solid foundation for subsequent graph query language queries, solving the problems of information fragmentation and poor relevance in requirements analysis. The construction of the knowledge graph begins with an in-depth analysis of the target system's functional modules, decomposing the system into a series of nodes of domain type, value stream type, activity type, task type, and step type. These nodes contain rich attribute information, such as name, description, scope, stakeholders, and the relationships between them, such as inclusion relationships and execution sequence relationships. Through this refined classification and association, the knowledge graph can clearly display the internal logical structure of the system and the external business environment, enabling precise matching of requirements analysis needs. This technical feature not only enhances the depth and breadth of requirements analysis but also, through the visualization of the knowledge graph, helps requirements analysts better understand and grasp the overall system, providing intuitive support for subsequent analysis work and significantly improving the efficiency and quality of requirements analysis.

[0054] Specifically, before converting the above requirements analysis into a graph query language using a large programming model, the above method also includes: optimizing the above programming model using model optimization techniques to obtain the optimized programming model, wherein the above model optimization techniques include knowledge distillation, reinforcement learning, and prompting engineering.

[0055] This method improves the performance of large programming models by applying model optimization techniques, addressing efficiency and accuracy issues in the initial model state. Model optimization techniques, such as knowledge distillation, reinforcement learning, and prompting engineering, aim to improve the performance of large programming models in the process of converting requirements analysis requests into graph query languages. Knowledge distillation is a model compression technique that transfers knowledge from a large model to a smaller model, ensuring accuracy while improving processing speed, suitable for large-scale requirements analysis scenarios. Reinforcement learning simulates the decision-making process, allowing the model to learn the optimal conversion strategy through multiple trials, enhancing the model's intelligence and adaptability, especially when handling complex requirements analysis requests. Prompting engineering adds specific guiding information to the model input, helping the model understand and respond to user needs more quickly, suitable for fast-response and high-precision requirements analysis scenarios. Through the comprehensive application of these optimization techniques, the performance of large programming models in requirements analysis request conversion tasks is significantly improved, not only accelerating the conversion speed but also enhancing the accuracy and relevance of the query language, laying a solid foundation for subsequent knowledge graph queries and requirements analysis report generation.

[0056] In addition, this embodiment also includes a dynamic knowledge graph update mechanism for real-time updates to the enterprise-level business model knowledge graph. Due to the continuous changes in business requirements and system functions, the knowledge graph needs to be able to quickly adapt to these changes to maintain its accuracy and timeliness. Specifically, when the system receives a new requirements specification or system design document, the large model will automatically identify new entities, new relationships, and changes in entity attributes in the document and update them in the knowledge graph.

[0057] For example, if a bank decides to introduce a new credit card product, the system will automatically identify and create a new "Product" node, while establishing relationships with existing nodes such as "Channels," "Role," and "Decision Conditions," and updating the "Value Stream" and "Activity" nodes to reflect the impact of the new product on business processes. This process requires no manual intervention, greatly reducing the complexity and cost of knowledge graph maintenance, while ensuring the real-time nature of the knowledge graph, enabling it to quickly respond to changes in business needs.

[0058] Through a dynamic knowledge graph update mechanism, this embodiment can reflect changes in business requirements in real time, maintaining the accuracy and timeliness of the knowledge graph, thereby improving the efficiency and accuracy of requirements analysis. Furthermore, this mechanism can reduce the workload of manually maintaining the knowledge graph, lower the error rate, and ensure data quality and the foundation for reasoning during the requirements analysis process.

[0059] This embodiment also includes the automatic identification and visualization of relationships between requirements. By optimizing the inference model, it can automatically identify direct and indirect relationships between different requirements based on entity relationships in the knowledge graph, and display these relationships graphically. For example, when a requirements analyst proposes a requirement to "increase the speed of credit card approval," the inference model will identify the direct relationship between this requirement and tasks such as "customer information verification" and "credit scoring," as well as the indirect relationship with value streams such as "risk management" and "customer service satisfaction." Subsequently, the requirements analysis subgraph will display these relationships in an intuitive graphical form, helping analysts to fully understand the scope of impact and potential risks of the requirements.

[0060] This mechanism helps requirements analysts quickly understand the complex relationships between requirements, enabling them to make more informed decisions. By automatically identifying requirement relationships, it effectively prevents the "domino effect" of requirement changes—that is, a change in one requirement leading to a chain reaction of related requirements—thus avoiding additional costs and risks caused by requirement changes. Simultaneously, visually displaying the relationships between requirements makes the requirements analysis process more transparent and traceable, improving the efficiency and quality of enterprise decision-making.

[0061] To enable those skilled in the art to better understand the technical solution of this application, the implementation process of the requirement analysis and generation method based on knowledge graphs and large models of this application will be described in detail below with reference to specific embodiments.

[0062] This embodiment relates to a specific method for generating requirements analysis based on knowledge graphs and large models. The knowledge graph constructed based on an enterprise-level business model, combined with a large language model, provides business requirements analysts with efficient and accurate analysis. For example... Figure 3 As shown, it specifically includes the following:

[0063] 1. Preparation of business system knowledge base and large model capabilities:

[0064] This embodiment involves a business system knowledge base that mainly comprises two parts. The first part is a knowledge graph built based on an enterprise-level business model. The second part is a knowledge base stored in text form based on detailed system specifications and requirements, supporting vector retrieval. Before application, this embodiment requires the capabilities provided by three major models: enhanced retrieval capabilities, graph query code generation capabilities, and large-scale inference model capabilities. These capabilities include, but are not limited to, acquisition through existing model capabilities, fine-tuning, reinforcement learning, and prompt words.

[0065] 1.1 Constructing an enterprise-level business model knowledge graph:

[0066] Based on the process model in the enterprise-level business model, the existing system functional modules are abstracted and defined according to domain, value stream, activity, task, and step. The knowledge graph vertex types include five categories: domain, value stream, activity, task, and step. Each type of vertex contains necessary attributes, including but not limited to name, description, scope, stakeholders, product, channel, partner, role, and decision conditions. Relationships between vertices include, but are not limited to, containment relationships and execution sequence relationships. Containment relationship attributes include, but are not limited to, the ID of the out-degree vertex, the ID of the in-degree vertex, execution dependencies, and execution type (triggered, timed).

[0067] 1.2 Constructing a detailed system knowledge base:

[0068] Collect system documentation related to requirements analysis, including but not limited to detailed system design specifications and requirements specifications for each stage. Divide this documentation into blocks based on minimum functional dimensions and use a vector model to create vector indexes for the block data.

[0069] 1.3 Optimize graph query code to generate large models:

[0070] We leverage open-source programming models and employ techniques including, but not limited to, knowledge distillation, fine-tuning, reinforcement learning, and prompting engineering to enhance the model's ability to generate code for graph queries in requirement analysis scenarios. Here, we take prompting engineering as an example, pre-providing the large model with prompts based on the structure of the enterprise-level business model knowledge graph, including but not limited to vertex types, vertex attributes, edge types, edge attributes, and requirement analysis strategies.

[0071] 1.4 Optimize the large-scale vector retrieval model:

[0072] We improve the performance of open-source large models by employing techniques including but not limited to knowledge distillation, fine-tuning, reinforcement learning, and prompting engineering.

[0073] 2. Obtain relevant documents for requirements analysis:

[0074] Based on the requirements rewriting and vector retrieval capabilities provided by the large model, documents related to requirements analysis can be obtained from the knowledge base, including but not limited to detailed system design documents and requirements specifications for each stage.

[0075] 2.1 Rewriting the request:

[0076] like Figure 3 As shown, requirements analysts express their requirements based on natural language. The requirements are then rewritten using a large model under the constraints of prompting engineering to standardize the expression of the requirements, forming a structured and consistent expression, with the aim of improving the performance of vector retrieval.

[0077] 2.2. Vector retrieval related documents:

[0078] Based on the revised expression of the request, key information of the request is extracted, including but not limited to functional entities, the intent of the request, and conditional restrictions. The above information is vectorized, and various documents and materials related to the request are retrieved and analyzed in the constructed vector knowledge base.

[0079] 3. Obtain sub-graphs related to requirements analysis:

[0080] Based on the graph query code generation capability provided by the large model, the requirement analysis subgraph can be retrieved from the already constructed enterprise-level business model knowledge graph. The subgraph includes JSON text format and graphical format.

[0081] 3.1 Rewriting the request:

[0082] This is equivalent to 2.1, a rewritten requirement, aimed at improving the accuracy of graph query code.

[0083] 3.2 Obtaining the Requirements Analysis Sub-diagram:

[0084] Based on the rewritten requirements, generate graph query code, and then use the graph query code to retrieve the requirement analysis subgraph from the enterprise-level business model knowledge graph. The requirement analysis subgraph includes both JSON text and graphical formats.

[0085] 4. Generate requirements analysis results based on relevant documents and requirements analysis sub-diagrams:

[0086] The requirement analysis subgraphs in JSON text format, obtained from the aforementioned requirement-related documents, are used as input to the large-scale inference model. These subgraphs provide causal inference for the model to analyze requirements. Combined with requirement-related documents obtained through vector retrieval, the large-scale inference model ultimately provides an analysis report based on the requirements for requirement analysts to refer to.

[0087] 4.1 Generate the requirements analysis logic chain:

[0088] The reasoning model uses the requirement analysis sub-graph in JSON text format as a basis to reason about the analysis requirements, forming a requirement analysis logic chain with accurate causal logic.

[0089] 4.2 Generate a requirements analysis report:

[0090] The reasoning model uses the above logical chain as input prompts and combines them with the retrieved system design documents to generate a requirements analysis report.

[0091] 4.3 Assist requirements analysts in evaluation and decision-making:

[0092] Requirements analysts combine the requirements analysis logic chain generated by the reasoning model, analysis reports, requirements analysis sub-graphs, and retrieved system design documents to conduct requirements assessments and assist in decision-making.

[0093] This embodiment uses an enterprise-level business model knowledge graph to semantically and structurally store the process model in the enterprise-level business model. The triple structure based on the attribute relationship graph has a natural correspondence with the subject-verb-object-adverb-complement structure of the Chinese language, making it easier to combine with large language models.

[0094] Furthermore, by having requirements analysts express their analytical needs using natural language, and using graph query code generated from an enterprise-level business model knowledge graph and a large programming model as input to the enterprise-level business model knowledge graph, a requirements analysis subgraph in JSON text format is obtained. This requirements analysis subgraph serves as input to the large inference model, outputting a requirements analysis reasoning logic chain, which then guides the large inference model to generate the final analysis report based on the retrieved detailed system information. Requirements analysts then use the analysis report generated by the large model, the reasoning logic chain, the retrieved detailed system design documents, and the requirements analysis subgraph to conduct requirements assessment and decision-making.

[0095] This embodiment achieves the following specific effects:

[0096] 1. Supports efficient reasoning and complex decision-making:

[0097] It provides efficient query and reasoning capabilities. Based on a graph query language, it can efficiently query complex relationship networks, find critical paths to achieve business requirements, and quickly assess the scope of impact when facing changes in requirements to adapt to rapid business responses. Combined with the reasoning capabilities of knowledge graphs, it can find key issues in complex and ever-changing credit card requirements, derive new knowledge, and enhance the intelligence of requirements management. The enterprise-level business model asset knowledge graph supports multi-dimensional analysis of requirements model assets, enabling comprehensive evaluation and decision-making on business requirements from multiple perspectives. By mining correlations through graph analysis, it can identify potential business requirement risks and opportunities, thereby supporting complex business requirement decisions.

[0098] 2. Provide intuitive evidence for the requirements analysis:

[0099] The large-scale reasoning model provides requirements analysts with analysis reports, along with relevant system documentation, requirements analysis sub-diagrams, and requirements analysis reasoning logic chains. It also provides supporting materials for the requirements analysis reports, facilitating timely review by requirements analysts.

[0100] 3. Eliminate the illusion of large model applications:

[0101] The large-scale model application phase includes rewriting requirements, document retrieval, graph query code generation, generating requirement analysis reasoning logic chains based on requirement analysis subgraphs, generating requirement analysis reports based on reasoning logic chains, requirement analysis subgraph JSON text, and system-related documents. These steps provide accurate data sources and reasoning basis for the requirement analysis phase, effectively eliminating the illusion problem caused by directly applying the large-scale model.

[0102] This application also provides a requirement analysis generation apparatus based on knowledge graphs and large models. It should be noted that this apparatus can be used to execute the requirement analysis generation method based on knowledge graphs and large models provided in this application. This apparatus is used to implement the above embodiments and preferred embodiments; details already described will not be repeated. As used below, the term "module" can refer to a combination of software and / or hardware that performs a predetermined function. Although the apparatus described in the following embodiments is preferably implemented in software, hardware implementation, or a combination of software and hardware, is also possible and contemplated.

[0103] The following describes the requirement analysis and generation device based on knowledge graphs and large models provided in the embodiments of this application.

[0104] Figure 4This is a schematic diagram of a requirement analysis and generation device based on knowledge graphs and large models according to an embodiment of this application. Figure 4 As shown, the device includes:

[0105] The query processing unit 41 is used to obtain the user's requirements analysis requests for the target system, convert the above requirements analysis requests into graph query language, and perform query processing on the above graph query language in the knowledge graph to obtain an analysis subgraph, wherein the above requirements analysis requests represent the requests for modification or query of the above target system.

[0106] The retrieval and processing unit 42 is used to retrieve and process the above-mentioned requirements analysis requests based on the system detailed design knowledge base, and obtain the system detailed design documents corresponding to the above-mentioned requirements analysis requests. The system detailed design knowledge base includes system detailed design documents of multiple target systems.

[0107] Input unit 43 is used to input the above system detailed design document and the above analysis subgraph into the reasoning big model to obtain a requirement analysis report corresponding to the above requirement analysis requirements.

[0108] In this embodiment, the query processing unit is used to obtain the user's requirements analysis requests for the target system, convert the requirements analysis requests into graph query language, and perform query processing on the graph query language in the knowledge graph to obtain an analysis subgraph. The requirements analysis requests represent requests for modification or querying of the target system. The retrieval processing unit is used to retrieve and process the requirements analysis requests according to the system detailed design knowledge base to obtain the system detailed design documents corresponding to the requirements analysis requests. The system detailed design knowledge base includes system detailed design documents for multiple target systems. The input unit is used to input the system detailed design documents and the analysis subgraph into the reasoning big model to obtain a requirements analysis report corresponding to the requirements analysis requests. By using knowledge graph and big model technology for requirements analysis and evaluation, efficient and accurate query and reasoning capabilities can be provided for requirements analysis, displaying intuitive, interpretable, and causal reasoning-based analysis results. For target systems with complex business logic and frequently changing requirements, requirements analysis and evaluation using knowledge graph and big model technology can reduce the decision-making risk of complex problems and improve the quality and efficiency of requirements analysis. This solves the problem that existing technologies, which evaluate business requirements by pre-setting functional points, may encounter situations where the pre-set functional points cannot cover the requirements scenarios during the requirements analysis process.

[0109] As an optional solution, the device further includes an acquisition unit, a block processing unit, and a first construction unit; the acquisition unit is used to acquire system information related to the target system before retrieving and processing the above-mentioned requirements analysis requests based on the system detailed design knowledge base, wherein the above-mentioned system information includes the design specification of the target system and the requirements specification specifications for each stage; the block processing unit is used to block the above-mentioned system information according to the minimum functional dimension of the target system to obtain multiple block data, and to use a vector model to establish a vector index for each of the above-mentioned block data to obtain multiple above-mentioned system detailed design documents; the first construction unit is used to construct the above-mentioned system detailed design knowledge base based on the multiple above-mentioned system detailed design documents.

[0110] An optional scheme is that the retrieval processing unit includes a standardized rewriting processing module and a retrieval processing module; the standardized rewriting processing module is used to perform standardized rewriting processing on the above-mentioned requirements analysis requests based on the language big model under the constraints of prompting engineering to obtain text rewriting requests; the retrieval processing module is used to convert the above-mentioned text rewriting requests into vector form to obtain text request vectors, and to retrieve and process the above-mentioned text request vectors according to the system detailed design knowledge base to obtain the above-mentioned system detailed design documents.

[0111] An optional scheme is provided, wherein the input unit includes a first input module and a second input module; the first input module is used to input the analysis subgraph in JSON text form into the above-mentioned reasoning model to obtain the requirement analysis logic chain; the second input module is used to input the above-mentioned requirement analysis logic chain and the above-mentioned system detailed design document into the above-mentioned reasoning model to obtain the above-mentioned requirement analysis report corresponding to the above-mentioned requirement analysis requirements.

[0112] In an optional embodiment, the apparatus further includes a verification unit, which, after obtaining a requirements analysis report corresponding to the aforementioned requirements analysis requirements, verifies the aforementioned analysis report using the aforementioned analysis subgraph, the aforementioned system detailed design document, and the aforementioned requirements analysis logic chain to obtain a verification result, and updates the aforementioned inference big model based on the aforementioned verification result, wherein the aforementioned analysis subgraph includes analysis subgraphs in JSON text form and analysis subgraphs in graphical form.

[0113] In an optional embodiment, the apparatus further includes a second construction unit for constructing the knowledge graph based on the functional modules of the target system before performing query processing on the graph query language in the knowledge graph. The types of vertices in the knowledge graph include at least one of domain type, value stream type, activity type, task type, and step type. The vertex information of the knowledge graph includes at least one of attributes, name, description, scope, stakeholders, products, channels, partners, roles, and decision conditions. The relationships between the vertices of the knowledge graph include at least one of inclusion relationship and execution sequence relationship.

[0114] In an alternative embodiment, the apparatus further includes an optimization processing unit, used to optimize the programming model using model optimization techniques before converting the aforementioned requirements analysis requests into a graph query language using a large programming model, to obtain the optimized programming model, wherein the aforementioned model optimization techniques include knowledge distillation, reinforcement learning, and prompting engineering.

[0115] The aforementioned requirement analysis and generation device based on knowledge graphs and large models includes a processor and a memory. The query processing unit, retrieval processing unit, input unit, etc., are all stored as program units in the memory. The processor executes these program units stored in the memory to achieve the corresponding functions. All of the above modules reside in the same processor; alternatively, the modules may be located in different processors in any combination.

[0116] The processor contains a kernel, which retrieves the corresponding program units from memory. One or more kernels can be configured. By adjusting kernel parameters, the problem arises where existing technologies rely on preset function points for business requirement assessment, and these preset function points may not cover all requirement scenarios during the requirements analysis process.

[0117] The memory may include non-permanent memory in computer-readable media, such as random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM, and the memory includes at least one memory chip.

[0118] This invention provides a computer-readable storage medium including a stored program, wherein, when the program is running, it controls the device where the computer-readable storage medium is located to execute the aforementioned requirement analysis and generation method based on knowledge graphs and large models.

[0119] This invention provides a processor for running a program, wherein the program executes the aforementioned requirement analysis and generation method based on knowledge graphs and large models.

[0120] This invention provides an electronic device, which includes a processor, a memory, and a program stored in the memory and executable on the processor. When the processor executes the program, it implements at least the steps of the above-described requirement analysis and generation method based on knowledge graphs and large models.

[0121] The devices mentioned in this article can be servers, PCs, tablets, mobile phones, etc.

[0122] This application also provides a computer program product that, when executed on a data processing device, is suitable for executing a program that initializes a requirement analysis generation method based on at least the above-described knowledge graph and large model.

[0123] It is obvious to those skilled in the art that the modules or steps of the present invention described above can be implemented using general-purpose computing devices. They can be centralized on a single computing device or distributed across a network of multiple computing devices. They can be implemented using computer-executable program code, and thus can be stored in a storage device for execution by a computing device. In some cases, the steps shown or described can be performed in a different order than those described herein, or they can be fabricated as separate integrated circuit modules, or multiple modules or steps can be fabricated as a single integrated circuit module. Thus, the present invention is not limited to any particular combination of hardware and software.

[0124] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0125] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0126] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0127] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0128] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0129] Memory may include non-persistent memory in computer-readable media, such as random access memory (RAM) and / or non-volatile memory, like read-only memory (ROM) or flash RAM. Memory is an example of computer-readable media.

[0130] Computer-readable media includes both permanent and non-permanent, removable and non-removable media that can store information using any method or technology. Information can be computer-readable instructions, data structures, modules of programs, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other non-transferable medium that can be used to store information accessible by a computing device. As defined herein, computer-readable media does not include transient computer-readable media, such as modulated data signals and carrier waves.

[0131] The technical features of the above embodiments can be combined in any way. For the sake of brevity, 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.

[0132] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element.

[0133] The above description is merely a preferred embodiment of this application and is not intended to limit this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the protection scope of this application.

Claims

1. A method for generating requirements analysis based on knowledge graphs and large models, characterized in that, include: The user's requirements for the target system are obtained, and the requirements are converted into a graph query language using a large programming model. The graph query language is then queried and processed in a knowledge graph to obtain an analysis subgraph. The requirements represent the analysis requests for modification or query of the target system. The system detailed design knowledge base is used to retrieve and process the requirements analysis requests to obtain system detailed design documents corresponding to the requirements analysis requests. The system detailed design knowledge base includes system detailed design documents of multiple target systems. The system detailed design document and the analysis subgraph are input into the reasoning model to obtain a requirements analysis report corresponding to the requirements analysis requirements.

2. The method according to claim 1, characterized in that, Before retrieving and processing the requirements analysis requests based on the system's detailed knowledge base, the method further includes: Obtain system information related to the target system, wherein the system information includes the design specification and requirements specification for each stage of the target system; The system data information is divided into blocks based on the minimum functional dimension of the target system to obtain multiple blocks of data. A vector model is then used to establish a vector index for each block of data to obtain multiple detailed system documents. The system detailed design knowledge base is constructed based on multiple system detailed design documents.

3. The method according to claim 1, characterized in that, Based on the system detailed design knowledge base, the requirements analysis requests are retrieved and processed to obtain the system detailed design documents corresponding to the requirements analysis requests, including: Based on the language big model, the requirements analysis requests are standardized and rewritten under the constraints of prompting engineering to obtain text rewriting requests. The text rewriting request is converted into a vector form to obtain a text request vector. The text request vector is then retrieved and processed according to the system detailed design knowledge base to obtain the system detailed design document.

4. The method according to claim 1, characterized in that, The system detailed design document and the analysis subgraph are input into the inference model to obtain a requirements analysis report corresponding to the requirements analysis needs, including: Input the analysis subgraph in JSON text format into the inference model to obtain the requirement analysis logic chain; The requirement analysis logic chain and the system detailed design document are input into the reasoning model to obtain the requirement analysis report corresponding to the requirement analysis requirements.

5. The method according to claim 4, characterized in that, After obtaining the requirements analysis report corresponding to the stated requirements analysis needs, the method further includes: The analysis report is verified using the analysis subgraph, the system detailed design document, and the requirement analysis logic chain to obtain verification results. The reasoning model is then updated based on the verification results. The analysis subgraph includes analysis subgraphs in JSON text format and analysis subgraphs in graphical format.

6. The method according to claim 1, characterized in that, Before performing query processing on the graph query language within the knowledge graph, the method further includes: The knowledge graph is constructed based on the functional modules of the target system. The types of vertices in the knowledge graph include at least one of domain type, value stream type, activity type, task type, and step type. The vertex information of the knowledge graph includes at least one of attributes, name, description, scope, stakeholders, products, channels, partners, roles, and decision conditions. The relationships between the vertices of the knowledge graph include at least one of inclusion relationship and execution sequence relationship.

7. The method according to claim 1, characterized in that, Before using a large programming model to convert the requirements analysis needs into a graph query language, the method also includes: The programming model is optimized using model optimization techniques to obtain the optimized programming model. The model optimization techniques include knowledge distillation, reinforcement learning, and prompting engineering.

8. A requirement analysis and generation device based on knowledge graphs and large models, characterized in that, include: The query processing unit is used to obtain the user's requirements analysis requests for the target system, convert the requirements analysis requests into a graph query language, and perform query processing on the graph query language in the knowledge graph to obtain an analysis subgraph, wherein the requirements analysis requests represent the requests for modification or query of the target system. The retrieval and processing unit is used to retrieve and process the requirements analysis request based on the system detailed design knowledge base to obtain the system detailed design document corresponding to the requirements analysis request, wherein the system detailed design knowledge base includes multiple system detailed design documents of the target system. The input unit is used to input the system detailed design document and the analysis subgraph into the reasoning model to obtain a requirement analysis report corresponding to the requirement analysis requirements.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium includes a stored program, wherein, when the program is executed, it controls the device containing the computer-readable storage medium to perform the requirement analysis and generation method based on any one of claims 1 to 7.

10. An electronic device, characterized in that, include: One or more processors, a memory, and one or more programs, wherein the one or more programs are stored in the memory and configured to be executed by the one or more processors, the one or more programs including methods for performing the knowledge graph and large model-based requirements analysis generation method according to any one of claims 1 to 7.

Images