Business expansion project processing method and device

Abstract

The invention relates to the technical field of power system informatization, artificial intelligence and big data, in particular to a business expansion project processing method and device. The method comprises the steps that knowledge and entities are extracted from multi-source heterogeneous data, the incidence relation between the entities is established based on the knowledge and the entities, a knowledge graph of business expansion is constructed, then natural language questions input by a user are received, the natural language questions are analyzed, question intentions and key elements are obtained, and the key elements comprise key entities or key attributes; and based on the problem intention, the key elements and the context information of the business expansion project, executing multi-layer matching retrieval in the knowledge graph to obtain a result set. The problem that in the prior art, accurate understanding and retrieval of related problems of business expansion items cannot be achieved can be solved.

Description

Technical Field

[0001] This invention relates to the fields of power system informatization, artificial intelligence and big data technology, and in particular to a method and apparatus for processing business expansion application projects. Background Technology

[0002] Business expansion application is a core link for power supply companies in connecting user electricity applications with the actual situation of the power grid and handling related business of power supply and electricity expansion. It runs through the entire process of project planning, design, and implementation, and is crucial to ensuring power supply reliability and meeting user electricity needs. With the development of the power industry, the number of business expansion application projects is increasing and the scenarios are becoming more complex, which places higher demands on the accuracy and efficiency of business processing.

[0003] Currently, business expansion and installation related services rely on multiple sources of information, such as policy and regulation databases, technical standard databases, and historical project archives. Problem-solving is mostly achieved through traditional search engines or keyword matching in knowledge bases, and some rely on expert experience guidance, lacking a systematic intelligent processing mechanism.

[0004] Existing technologies have significant shortcomings: there are numerous and rapidly updating policies and standards, making manual searching inefficient, prone to omissions, or using outdated versions; historical experience and expert knowledge are difficult to share and reuse effectively; keyword matching search results are broad and poorly relevant, failing to accurately respond to specific project contexts. Summary of the Invention

[0005] This invention provides a method and apparatus for processing business expansion and installation projects, in order to solve the problem that the prior art cannot achieve accurate understanding and retrieval of issues related to business expansion and installation projects.

[0006] In a first aspect, embodiments of the present invention provide a method for processing business expansion application projects, including: Knowledge and entities are extracted from multi-source heterogeneous data, and the relationships between entities are established based on the knowledge and entities to construct a knowledge graph for business expansion and installation. Receive a natural language question input by the user, parse the natural language question, and obtain the question intent and key elements, including key entities or key attributes; Based on the stated question intent, the stated key elements, and the contextual information of the business expansion and installation project, a multi-level matching retrieval is performed in the knowledge graph to obtain a result set.

[0007] In one possible implementation, after performing multi-level matching retrieval on the knowledge graph based on the problem intent, the key elements, and the context information of the business expansion and installation project, and obtaining the result set, the method further includes: Based on the result set and actual data of the business expansion and installation projects, targeted personalized suggestions are generated and output. Obtain user feedback on the personalized suggestions to optimize subsequent retrieval and personalized suggestion generation processes.

[0008] In one possible implementation, based on the problem intent, the key elements, and the context information of the business expansion and installation project, a multi-level matching retrieval is performed in the knowledge graph to obtain a result set, including: Based on the stated question intent, the stated key elements, and the contextual information of the business expansion application project, the semantic retrieval layer, the project attribute filtering layer, the real-time context matching and ranking layer, and the cross-case reasoning and enhancement layer are sequentially performed in the knowledge graph to obtain a result set.

[0009] In one possible implementation, semantic retrieval is performed on the knowledge graph, including: Based on the question intent and the key entity, a preliminary search is performed in the knowledge graph to obtain a set of knowledge related to the key entity, which is output as a coarse-grained related result set.

[0010] In one possible implementation, the retrieval of the project attribute filtering layer in the knowledge graph includes: In the coarse-grained correlation result set, based on the voltage level of the project in the project profile parameters of the business expansion application project context information, policies and design schemes that are not applicable to the voltage level are filtered out to obtain the first correlation result set; Based on the installation capacity in the profile parameters of the business expansion application project, the policy clauses in the first relevant result set are matched, and the matched clauses are marked to obtain the second relevant result set; In the second relevant result set, based on the electricity usage characteristics in the profile parameters of the business expansion application project, the design cases are sorted according to their relevance to "important loads" to obtain the result set after project attribute filtering.

[0011] In one possible implementation, real-time context matching and ranking layer retrieval is performed on the knowledge graph, including: Determine the project address and power point location of all successfully implemented historical cases in the result set after the project attribute filtering, calculate the spatial distance between the address of the business expansion application project and the project address and power point location, increase the recommendation priority of the corresponding solution with the closest spatial distance, and obtain the first optimized result set; The recommendation priority of historical solutions in all historical cases in the first optimization result set that have a current surrounding power point real-time load rate exceeding a preset threshold is reduced, and a prompt message is generated to obtain the second optimization result set. The context of all parallel historical cases in the second optimization result set is retrieved to detect the access interval resource usage. If a predetermined access interval exists, the interval is automatically excluded and a resource usage prompt is generated to obtain the third optimization result set. By combining the user's expected time with the material inventory status and equipment procurement cycle, the historical solutions in the third optimization result set are ranked according to their time feasibility, resulting in a fine-grained result set.

[0012] In one possible implementation, cross-case reasoning and enhancement layer retrieval of the knowledge graph include: By performing commonality mining and difference alerts on historical cases in the fine-grained result set, the final result set is obtained.

[0013] In one possible implementation, based on the result set and actual data of the business expansion and installation project, targeted personalized suggestions are generated and output, including: Get the predefined business rule engine; Information related to the business expansion and installation project is extracted from the result set. The extracted information is matched with the actual data of the business expansion and installation project. Based on the matching results and the business rule engine, targeted personalized suggestions are generated. The extracted information includes: policy standard clauses, historical successful cases, technical adaptation solutions and resource allocation information.

[0014] In one possible implementation, the business rule engine adopts IF-THEN rules. The IF condition is set based on the policy threshold, case adaptation conditions and resource constraint parameters in the result set. The reasoning in the IF-THEN rule is based on the relationship between entities in the knowledge graph, combined with the historical relationship data in the result set and the characteristics of the business expansion application project for logical deduction. The THEN conclusion corresponds to specific solution requirements, recommended options or precautions.

[0015] Secondly, embodiments of the present invention provide a processing device for business expansion application projects, comprising: The knowledge graph construction module is used to extract knowledge and entities from multi-source heterogeneous data, and establish the relationship between entities based on the knowledge and entities to construct the knowledge graph for business expansion and installation. The parsing module is used to receive natural language questions input by the user, parse the natural language questions, and obtain the question intent and key elements, including key entities or key attributes. The retrieval module is used to perform multi-level matching retrieval in the knowledge graph based on the question intent, the key elements, and the context information of the business expansion application project, and obtain a result set.

[0016] Thirdly, embodiments of the present invention provide a terminal, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of the processing method for business expansion and installation projects as described in the first aspect or any possible implementation thereof.

[0017] Fourthly, embodiments of the present invention provide a computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps of the processing method for business expansion and installation projects as described in the first aspect or any possible implementation thereof.

[0018] This invention provides a method and apparatus for processing business expansion and installation projects. It extracts knowledge and entities from multi-source heterogeneous data, establishes relationships between entities based on this knowledge and entities, and constructs a knowledge graph for business expansion and installation. Then, it receives natural language questions input by users, parses the questions to obtain the question intent and key elements, including key entities or key attributes. Based on the question intent, key elements, and contextual information of the business expansion and installation project, it performs multi-level matching retrieval in the knowledge graph to obtain a result set. This invention constructs a knowledge graph for business expansion and installation projects using multi-source heterogeneous data, transforming fragmented knowledge scattered in archives and expert minds into structured, reusable digital assets. This avoids knowledge loss due to personnel turnover and breaks down knowledge barriers, allowing for rapid access to historical success stories, technical standards, and solutions, significantly improving knowledge reuse rates. Through the rapid retrieval capabilities and multi-level matching filtering mechanism of the knowledge graph, project personnel can quickly obtain accurate result sets without manually sifting through massive amounts of policy documents, searching historical archives, or repeatedly consulting experts, significantly reducing the time cost of information retrieval and solution demonstration. Attached Figure Description

[0019] To more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0020] Figure 1 This is a flowchart illustrating the implementation of the processing method for business expansion and installation projects provided in this embodiment of the invention. Figure 2 This is a schematic diagram of the structure of the processing device for business expansion and installation projects provided in an embodiment of the present invention; Figure 3 This is a schematic diagram of the terminal provided in an embodiment of the present invention. Detailed Implementation

[0021] In the following description, specific details such as particular system architectures and techniques are set forth for illustrative purposes and not for limitation, in order to provide a thorough understanding of the embodiments of the invention. However, those skilled in the art will understand that the invention can be implemented in other embodiments without these specific details. In other instances, detailed descriptions of well-known systems, apparatuses, circuits, and methods are omitted so as not to obscure the description of the invention with unnecessary detail.

[0022] To make the objectives, technical solutions, and advantages of the present invention clearer, specific embodiments will be described below in conjunction with the accompanying drawings.

[0023] During the planning, design, and implementation of business expansion and installation projects, project personnel often encounter various complex problems, such as: 1. Difficulty in finding policies and regulations: There are numerous relevant policies, procedures and standards from the national, local and power grid companies, which are frequently updated. Manual searching is inefficient and prone to omissions or the use of outdated versions.

[0024] 2. Insufficient reuse of historical experience: A large amount of successful experience from completed projects, typical design schemes, and solutions to common problems are stored in historical archives and the minds of a few experts, and have not been effectively shared and utilized.

[0025] 3. Inaccurate answers to questions: When encountering specific technical or process issues, traditional search engines or knowledge base keyword matching returns broad and poorly relevant results, failing to provide accurate answers directly addressing the specific context of the current project (such as voltage level, electricity usage, regional characteristics, etc.).

[0026] Therefore, based on the above-mentioned technical problems, embodiments of the present invention provide a method for processing business expansion and installation application projects, achieving accurate understanding and retrieval of issues related to business expansion and installation application projects, and deeply integrating multi-source actual data of the current project to automatically generate highly targeted personalized suggestions. For example... Figure 1 The flowchart illustrating a method for processing business expansion application projects is detailed below: Step 101: Extract knowledge and entities from multi-source heterogeneous data, establish relationships between entities based on knowledge and entities, and construct a knowledge graph for business expansion and installation.

[0027] This step aims to build a knowledge graph covering all business scenarios of business expansion and installation, providing comprehensive and structured knowledge support for subsequent user problem analysis, accurate retrieval, and personalized suggestion generation.

[0028] First, the multi-source heterogeneous data sources required for knowledge graph construction are clearly defined, comprehensively covering various key information involved in the business expansion and installation project process, including policies, technologies, cases, geography, and equipment. Specifically, this includes: the national power policy and regulation database, the power grid company's technical standards database, historical business expansion and installation project archives (including detailed information such as design schemes, completion reports, and problem records), typical design atlases, equipment parameter databases, and geographic information data from the power grid Geographic Information System (GIS). Through data interface integration, file parsing, and database export, the unified collection and format standardization of the above-mentioned multi-source heterogeneous data are completed, ensuring data integrity and accessibility, laying the foundation for subsequent knowledge and entity extraction.

[0029] Based on the normalized multi-source heterogeneous data, natural language processing, data mining, and other technologies are used to extract structured, semi-structured, and unstructured knowledge respectively. At the same time, core entities and corresponding attributes related to business expansion and installation services are accurately identified and extracted. The extracted knowledge types cover policy clauses, technical standard requirements, historical project implementation experience, equipment compatibility specifications, and problem solutions. The entities extracted include, but are not limited to: projects (attributes: project number, voltage level, electricity category, address), users, equipment (such as transformers, switch cabinets, etc.), policy provisions, technical standards, geographical locations (such as substations, lines, etc.), problems, and solutions.

[0030] Through the above extraction operations, fragmented information scattered across different data sources is transformed into standardized knowledge units and entity information, solving the problems of knowledge dispersion and difficulty in reuse in the traditional model.

[0031] After completing knowledge and entity extraction, based on the actual logic and inherent data relationships of the business expansion and installation services, clear relationships are established between the entities, including but not limited to: The "compliance" relationship between projects and policy provisions (e.g., "Project A - [compliance] → Policy B"); The "applicable to" relationship between equipment and voltage level (e.g., "Equipment C - [Applicable to] → Voltage Level D"). The "corresponding solution relationship" between problems and solutions (e.g., "Problem E - [Corresponding Solution] → Solution F"); The "location" relationship between a project and its geographical location (e.g., "Project G - [located in] → Geographical location H").

[0032] Based on the extracted knowledge, entities, and established relationships, a knowledge graph for business expansion and installation is constructed using graph construction technology, forming a structured knowledge network. This allows various business expansion and installation-related information to be organically integrated through entity association, providing a core knowledge carrier for subsequent precise retrieval based on semantic understanding and context matching.

[0033] Step 102: Receive the natural language question input by the user, parse the natural language question, and obtain the question intent and key elements, including key entities or key attributes.

[0034] This step is used to accurately capture users' needs for business expansion and installation services, transforming unstructured natural language questions into structured question intent and key information, providing a clear direction for subsequent precise retrieval based on knowledge graphs.

[0035] Optionally, this step allows users to input natural language questions through various channels. Accepted input formats include text input and speech-to-text input. Users can directly express their needs using natural language from everyday conversations, without needing to follow a specific format. Examples include: "I want to apply for a new 10kV power connection for a commercial complex; what documents do I need to prepare?", "When will our community's power supply plan be finalized?", and "Does the user need to pay for the meter installation themselves?". The system can completely receive and store the original natural language question information.

[0036] Based on the received raw natural language question, Natural Language Processing (NLP) is used to parse it. The core of this process includes question intent recognition and key element extraction. The specific process is as follows: Intent Recognition: By optimizing the training of a pre-trained language model in conjunction with business expansion and installation scenarios, the system categorizes user natural language questions by intent, covering the core needs of all business expansion and installation scenarios, including but not limited to: inquiries about new installations / capacity increases, power supply solutions, interpretation of policies and regulations (such as electricity price standards and acceptance rules for incomplete applications), verification of application documents, inquiry about application progress, complaints and feedback, equipment compatibility inquiries, and feasibility assessments of access solutions. Simultaneously, a multi-label classification model supports the recognition of compound intents. For example, when a user simultaneously inquires about "document requirements and processing time," the system can simultaneously recognize both the intents of "document inquiry" and "progress inquiry."

[0037] Key Element Extraction: Named Entity Recognition (NER) technology, combined with domain-specific dictionary rules, is used to extract key elements from the problem. These key elements include key entities and their corresponding key attributes. Key entities include: projects, users, electrical equipment, policy provisions, technical standards, geographical location (such as electricity address and power supply location), and business type (such as new low-voltage non-residential installations and high-voltage capacity expansion). Key attributes include: voltage level (e.g., 10kV, 0.4kV), electricity usage category (e.g., commercial electricity, agricultural electricity), installed capacity, user's expected electricity connection time, electricity address details, project number, etc.

[0038] Furthermore, to address potential ambiguities, terminological ambiguities, or missing information in user inquiries, the system employs multiple mechanisms to ensure parsing accuracy: for questions with vague geographical locations, such as "nearby power supply access," contextual information is supplemented by combining the user's historical electricity application records or current location; for questions with unclear business types, such as "what materials are needed for electricity application," information is completed through option-based guidance (e.g., "Are you inquiring about new installation or change of service?"). The final output is a structured parsing result, including a clear question intent label (e.g., "New installation service information inquiry") and a list of key elements (e.g., "Entity: Low-voltage non-residential new installation; Attributes: Electricity address - XX street XX number, Application capacity - 50kW"). This result will directly serve as the core input condition for the contextual association retrieval in step 103.

[0039] Step 103: Based on the question intent, key elements, and contextual information of the business expansion application project, perform multi-level matching retrieval in the knowledge graph to obtain the result set.

[0040] This step is the core of achieving accurate retrieval of issues related to business expansion and installation projects. Its core logic is based on the problem intent and key elements obtained from step 102, combined with the contextual information of the business expansion and installation projects. In the knowledge graph, a four-layer progressive retrieval process of "semantic retrieval - attribute filtering - real-time matching - cross-case reasoning" is used to gradually narrow down the scope of results, improve the relevance and feasibility of the results, and finally output an accurate set of retrieval results.

[0041] In one embodiment, based on the question intent, key elements, and contextual information of the business expansion application project, a multi-level matching retrieval is performed in the knowledge graph to obtain a result set, which may include: Based on the question intent, key elements, and contextual information of the business expansion application project, the retrieval is performed sequentially in the knowledge graph through semantic retrieval, project attribute filtering, real-time context matching and ranking, and cross-case reasoning and enhancement layers to obtain the result set.

[0042] The core purpose of the semantic retrieval layer is to break through the limitations of traditional keyword matching, realize the broad association retrieval of knowledge based on semantic understanding, and lay the foundation for subsequent accurate screening.

[0043] In one embodiment, performing semantic retrieval in a knowledge graph may include: Based on the question intent and key entities, a preliminary search is performed in the knowledge graph to obtain a set of knowledge related to the key entities, which is output as a coarse-grained relevant result set.

[0044] Optionally, the specific processing steps of the semantic retrieval layer are as follows: Input information: The question intent obtained from step 102 (such as "query access point selection criteria"), key elements (including key entities and attributes, such as "commercial complex", "10kV", "city center"), and the user's original natural language question.

[0045] Implementation Process: The system uses the parsed question intent and key elements as search criteria, performing semantic association retrieval within the constructed knowledge graph of business expansion and installation applications. The retrieval process is not limited to literal matching but is based on the semantic relationships between entities in the knowledge graph, mining all knowledge units related to the question intent and key elements, covering various information such as policy provisions, technical standards, historical project cases, and solutions. For example, when a user's question is "For a 10kV installation project in a commercial complex in the city center, how to choose the optimal power access point?", the system will search the knowledge graph for all policy requirements, historical access schemes, technical specifications, and other content related to "10kV access" and "power access point selection".

[0046] Output: A coarse-grained relevant result set R0 with broad coverage and related to the core needs of the problem is obtained. This coarse-grained relevant result set R0 contains all potentially relevant knowledge information, providing a sufficient data foundation for subsequent targeted screening.

[0047] The project attribute filtering layer aims to use the core profile parameters of the business expansion and installation project to perform targeted filtering on the coarse-grained result set R0, remove inapplicable information, strengthen key matching content, and improve the relevance of the results.

[0048] In one embodiment, retrieving the item attribute filtering layer in the knowledge graph may include: In the coarse-grained relevance result set, based on the voltage level of the project in the project profile parameters of the business expansion application project context information, policies and design schemes that are not applicable to the voltage level are filtered out to obtain the first relevance result set; Based on the installation capacity in the profile parameters of the business expansion application project, the policy clauses in the first relevant result set are matched, and the matched clauses are marked to obtain the second relevant result set; In the second relevant results set, based on the electricity usage characteristics in the profile parameters of the business expansion application project, the design cases are sorted according to their relevance to "important loads" to obtain the result set after project attribute filtering.

[0049] The specific processing flow of the project attribute filtering layer is as follows: Input information: coarse-grained correlation result set R0, and project profile parameters (including core attributes such as project voltage level, installed capacity, and power consumption nature) in the context information of the business expansion application project.

[0050] Implementation process: Voltage level adaptation filtering: Based on the voltage level (e.g., 10kV, 0.4kV) in the project profile, the policy provisions, design schemes, and other contents in the coarse-grained relevant result set R0 are filtered, and information that is not applicable to the voltage level is directly removed. For example, design schemes applicable to 35kV projects are excluded from the search results of 10kV projects, and the first relevant result set is obtained.

[0051] Application capacity clause matching: Based on the application capacity in the project profile, the policy clauses in the first relevant result set are accurately matched and judged. For example, for a project with an application capacity of 5000kVA, clauses such as "a dedicated power supply is required for a capacity exceeding 8000kVA" are matched. The directly applicable clauses are highlighted to facilitate the quick location of the core policy basis and obtain the second relevant result set.

[0052] Prioritization of power consumption characteristics: Based on the power consumption characteristics in the project profile (such as important power users, general commercial users), the design cases in the second relevant result set are prioritized, with reliability design cases related to "important loads" given top priority to ensure more comprehensive support for key power consumption scenarios.

[0053] Output: The result set R1 is filtered through multiple dimensions of the project's core attributes. This result set R1 has been initially adapted to the basic characteristics of the current project and excludes obviously irrelevant information.

[0054] Real-time context matching and ranking layers are key to ensuring the accuracy and feasibility of retrieval. By integrating real-time dynamic information to optimize and rank the result set R1, the recommendation scheme is ensured to fit the current actual scenario and be feasible for implementation.

[0055] In one embodiment, performing real-time context matching and ranking layer retrieval in a knowledge graph may include: After determining the project addresses and power supply locations of all successfully implemented historical cases in the result set after filtering project attributes, calculate the spatial distance between the address of the business expansion application project and the project address and power supply location, increase the recommendation priority of the corresponding solution with the closest spatial distance, and obtain the first optimized result set; The recommendation priority of historical solutions in the first optimization result set that have a current surrounding power point real-time load rate exceeding a preset threshold is reduced, and a prompt message is generated to obtain the second optimization result set. The context of all parallel historical cases in the second optimization result set is retrieved to detect the resource usage of the access interval. If a predetermined access interval exists, the interval is automatically excluded and a resource usage prompt is generated to obtain the third optimization result set. By combining the user's expected time with the status of material inventory and the equipment procurement cycle, the historical solutions in the third optimization result set are ranked according to their time feasibility, resulting in a fine-grained result set.

[0056] The specific processing flow of the real-time context matching and sorting layer is as follows: Input information: The result set R1 after filtering project attributes, and real-time context information (including the project's real-time geographical location, the load status of surrounding power points, the progress of parallel projects, material inventory, equipment procurement cycle, etc.).

[0057] Implementation process: Geographic location matching optimization: Extract the project address and power source location (such as substation, line node) of all successfully implemented historical cases in result set R1. Calculate the spatial distance between the current business expansion application project address and the above addresses using a spatial distance calculation algorithm (such as the Haversine algorithm). Prioritize the access scheme and design ideas corresponding to the historical cases with the closest spatial distance, as the geographical environment and power grid conditions of nearby cases are more relevant. This yields the first optimized result set.

[0058] Grid carrying capacity verification and adjustment: Obtain the real-time load rate of the current surrounding power points corresponding to each historical case in the first optimization result set, set a preset threshold. If the current load rate of a power point corresponding to a historical scheme exceeds the threshold, it indicates that its access feasibility has decreased. The system lowers the recommendation priority of the scheme and generates a prompt message "The associated power point has a high load and needs careful evaluation," thus obtaining the second optimization result set. The preset threshold can be set according to experience or needs. In this embodiment, the value of the preset threshold is not limited. For example, the preset threshold can be 80%.

[0059] Multi-project conflict detection and elimination: Retrieve the context information of parallel projects involved in the second optimization result set, and focus on checking the occupation of key resources such as access intervals. If it is found that an access interval has been reserved by other parallel projects, the system will automatically exclude the interval from the recommended scheme of the current project and generate a resource occupation prompt of "This resource has been occupied by project XXX" to avoid resource conflicts and obtain the third optimization result set.

[0060] Time feasibility ranking: Combining the user's expected project completion time with the material inventory status and procurement cycle of the equipment required for each scheme in the third optimization result set (e.g., long cycle for imported equipment, and direct use of inventory equipment), all schemes are ranked in terms of time feasibility. This not only recommends the technically optimal scheme, but also provides alternative schemes that are more time-suitable (e.g., the technically suboptimal scheme that can use inventory equipment), resulting in a fine-grained result set R2.

[0061] Output: A fine-grained result set R2 optimized by real-time dynamic information in multiple dimensions. This result set takes into account both technical adaptability and practical feasibility.

[0062] The cross-case reasoning and enhancement layer extracts implicit knowledge and risk warnings through case analysis of the fine-grained result set R2, achieving in-depth enhancement of the results and providing richer support for personalized suggestions.

[0063] In one embodiment, cross-case reasoning and enhancement layer retrieval of a knowledge graph may include: By analyzing the commonalities and identifying the differences in historical cases within the fine-grained results set, the final results set is obtained.

[0064] The specific processing flow for cross-case reasoning and the enhancement layer is as follows: Input information: Fine-grained result set R2.

[0065] Implementation process: The system employs case-based reasoning technology to conduct in-depth analysis of multiple historical cases in the fine-grained result set R2, focusing on two core aspects: Commonality mining: Extract common choices and successful experiences from multiple similar projects. For example, "In three 10kV commercial complex installation projects similar to the current business expansion project, Type A switchgear was used because of its low failure rate in localized operation and maintenance," providing a reference for the preferred solution that has been verified in practice for the current project.

[0066] Difference alert: Identify key differences and potential risks between cases, such as "Project C is similar to the current business expansion and installation project in terms of load and location, but it was reworked during acceptance because it did not meet standard X. Please focus on checking standard clause Y." Proactively remind users of compliance points and risk points that they need to pay attention to, and help avoid similar problems.

[0067] Output: The final result set R includes in-depth insights such as common experiences and risk warnings. This result set not only contains precisely matched information, but also has derivative value and risk prediction capabilities.

[0068] The existing technology lacks forward-looking suggestions: existing systems are mostly based on passive queries and cannot proactively combine actual project data (such as load forecasting, geographical location, and surrounding power grid conditions) to provide users with personalized and forward-looking optimization suggestions, such as optimal access point recommendations, potential risk warnings, and energy-saving solution suggestions.

[0069] Therefore, in this embodiment of the invention, after obtaining the final result set R by completing the multi-level matching retrieval in step 103, the subsequent transformation from accurate retrieval results to actionable decision support is achieved through two core steps: "personalized suggestion generation and output" and "user feedback learning and optimization," thereby driving the continuous iteration of the system.

[0070] In one embodiment, after performing multi-level matching retrieval on the knowledge graph based on the question intent, key elements, and contextual information of the business expansion application project to obtain the result set, the process may further include: Based on the result set and actual data from business expansion and installation projects, targeted personalized suggestions are generated and output. Obtain user feedback on personalized suggestions to optimize subsequent retrieval and personalized suggestion generation processes.

[0071] The core of this step is to generate personalized suggestions that are not merely "document listings" and have strong practical application, based on the high-quality result set obtained from the retrieval and combined with the actual data of the business expansion and installation projects, through the dual support of a predefined business rule engine and graph reasoning.

[0072] In one embodiment, based on the result set and actual data of the business expansion and installation project, targeted personalized suggestions are generated and output, which may include: Get the predefined business rule engine; Information related to business expansion and installation projects is extracted from the results set. The extracted information is then matched with the actual data of the business expansion and installation projects. Based on the matching results and business rules, the engine generates targeted and personalized suggestions. The extracted information includes: policy standards and clauses, historical successful cases, technical adaptation solutions, and resource allocation information.

[0073] The business rule engine is a solidified carrier of domain expert experience and business expansion logic, and its core adopts the IF-THEN rule format. The IF condition is based on the policy threshold, case adaptation conditions and resource constraint parameters set in the result set. The reasoning in the IF-THEN rule is based on the relationship between entities in the knowledge graph, and logical deduction is performed by combining historical relationship data in the result set and the characteristics of business expansion application projects.

[0074] Optionally, the IF condition is closely based on the key information in the final result set R, specifically including: quantitative thresholds in policy standards (such as "application capacity exceeds 8000kVA" and "power supply point load rate is less than 80%), adaptation conditions of historical cases (such as "large commercial complex" and "city center area project"), and resource constraint parameters (such as "access interval is idle" and "inventory equipment is available"); THEN conclusion corresponds to specific scheme requirements, recommended options, or precautions, such as "adopt dual power supply", "prefer to select a certain substation for access", and "verify a certain standard clause".

[0075] Meanwhile, the rule engine's reasoning logic is not isolated but deeply integrated with the relationships between entities in the industry-wide knowledge graph. For example, if it is known that "important electricity users" are a subclass of "special electricity users," and that "special electricity users" are subject to the "dual power supply policy," then graph reasoning can deduce that "important electricity users" are also subject to the same policy. This refines the reasoning chain of the IF-THEN rule, ensuring the comprehensiveness and accuracy of rule application. Furthermore, the rule engine supports the integration of machine learning models (such as text generation models) to improve the fluency and accuracy of suggested text.

[0076] Extract core information directly related to the current business expansion and installation project from the final result set R, covering four main categories: Policy and Standard Clauses: The original text of applicable policies, quantitative requirements, and compliance points after multi-level retrieval and matching (such as the clauses on dedicated power supply in the "Implementation Rules for Business Expansion Application of XX Power Supply Company"). Historical success stories: Information on completed projects similar to the current project in terms of voltage level, power consumption, load size, and geographical location, including project number, design scheme (such as electrical main wiring diagram), equipment selection, implementation experience, etc. Technical adaptation solutions: access solutions, equipment configuration suggestions, construction process requirements, etc. that are suitable for the characteristics of the current project (such as transformer model selection and access point recommendations); Resource configuration information: Dynamic resource data such as the load status of surrounding power points, the availability of access intervals, the inventory status of materials, and the equipment procurement cycle.

[0077] The extracted information will be categorized and organized according to "policy-case-technology-resources" to form a structured information support library, avoiding fragmented information from interfering with the generation of suggestions.

[0078] The extracted structured information is precisely matched with the actual data of current business expansion and installation projects to ensure that the recommendations are fully aligned with the specific project context. This actual project data includes, but is not limited to: project installation capacity, voltage level, electricity usage type (e.g., major electricity users / ordinary commercial users), electricity address, load forecast data, user's expected completion time, and budget constraints.

[0079] During the matching process, the system will focus on verifying the suitability of the information: for example, comparing the load data of historical cases with the load of the current project, and selecting cases with load deviation within a preset range (such as ±20%) as core references; comparing the voltage level requirements of policy clauses with the voltage level of the current project to ensure that only fully compatible policies are referenced; and matching resource allocation information with the expected time of the project to determine the time feasibility of the solution.

[0080] Based on the above information matching results, and combining the IF-THEN rules of the business rule engine with knowledge graph reasoning, targeted personalized suggestions are generated. The suggestions have three main characteristics: Compliance: Strictly adhere to matching policy standards and provisions to ensure that recommendations comply with relevant national, local, and power grid company requirements; Practicality: Includes specific solution options, equipment models, data support, and operating procedures, rather than general descriptions; Personalized: Closely integrated with actual project data and dynamic resource conditions, each project's recommendations are unique.

[0081] For example, when a user asks, "My project is expected to have a load of 5000kVA, what precautions should I take?", the system generates the following suggestion: "Based on your load (5000kVA), according to the 'Implementation Rules for Business Expansion Application of XX Power Supply Company,' you need to use a dedicated power supply line. Recommended solutions: 1) Prioritize connection from the 10kV busbar of the nearby 110kV Chaoyang Substation. The current load rate of this substation is 65%, meeting the connection requirements (specific bay needs to be verified). 2) It is recommended to select two 2500kVA transformers as the main transformers to meet the N-1 operation requirements. 3) You can refer to the electrical main wiring diagram of project number Y2023-045 (also a large commercial complex) as a typical design." Ultimately, the recommendations will be presented to users in a clear and structured format, including modules such as "Policy Basis - Core Recommendations - Data Support - Reference Cases - Risk Warnings," enabling users to quickly grasp key information. At the same time, it will provide links to the original case texts and policy documents for users to access in-depth information.

[0082] Furthermore, to enable the system to evolve itself and improve the accuracy of subsequent retrieval and suggestion generation, the system records user feedback on personalized suggestions and uses it for model optimization. The specific process is as follows: The system supports users in submitting feedback in various ways, including: suggestion adoption status (e.g., "fully adopted," "partially adopted," "not adopted"), specific evaluations (e.g., "suggestion is accurate," "access point recommendation is unreasonable," "missing certain information"), and supplementary requirements (e.g., "need more equipment selection options," "want to understand cost estimates"). Feedback information will be linked to the current project's problem intent, key elements, search results, and generated suggestions, and stored in the system's feedback database, forming a complete business closed-loop data.

[0083] The feedback-based model optimization system periodically analyzes the data in the feedback database to optimize subsequent retrieval algorithms and suggestion generation models accordingly. Search model optimization: If a user reports that "a recommended case is not applicable", the system will analyze the differences between the case and the current project and adjust the weights in the multi-level matching search (such as reducing the weight of geographic location matching and increasing the weight of load adaptability); if a user frequently adopts suggestions corresponding to a certain type of policy clause, the system will increase the ranking priority of that type of clause in the search results.

[0084] Optimize the suggestion generation model: If users report that "the suggestion lacks cost information", the system will update the business rule engine and add a rule that "if the project has cost budget constraints, then the suggestion should include a cost estimate"; if the suggestions output by the text generation model have ambiguous expressions, the model will be fine-tuned based on positive suggestions from users to improve the fluency and accuracy of the expression.

[0085] Through continuous feedback and learning, the system can adapt to changes in business expansion (such as policy updates, technology upgrades, and changes in user needs), achieving a virtuous cycle of "search-suggestion-feedback-optimization" and gradually improving the performance of core functions and user experience.

[0086] This invention provides a method for processing business expansion and installation projects. It extracts knowledge and entities from multi-source heterogeneous data, establishes relationships between entities based on this knowledge and entities, and constructs a knowledge graph for business expansion and installation. The method receives natural language questions input by users, parses the questions to obtain the question intent and key elements, including key entities or key attributes. Based on the question intent, key elements, and contextual information of the business expansion and installation project, a multi-level matching retrieval is performed in the knowledge graph to obtain a result set. This invention constructs a knowledge graph for business expansion and installation projects using multi-source heterogeneous data, transforming fragmented knowledge scattered in archives and expert minds into structured, reusable digital assets. This avoids knowledge loss due to personnel turnover and breaks down knowledge barriers, allowing for rapid access to historical success stories, technical standards, and solutions, significantly improving knowledge reuse rates. Through the rapid retrieval capabilities and multi-level matching filtering mechanism of the knowledge graph, project personnel can quickly obtain accurate result sets without manually sifting through massive amounts of policy documents, searching historical archives, or repeatedly consulting experts, significantly reducing the time cost of information retrieval and solution demonstration. The embodiments of the present invention can not only provide accurate result sets, but also proactively prompt users with risks that they may have overlooked, better solutions or potential cost-saving opportunities based on data analysis, transforming passive querying into proactive empowerment.

[0087] In this embodiment of the invention, targeted personalized suggestions are generated and output based on the result set and actual data of the business expansion and installation project. User feedback on the personalized suggestions is obtained to optimize subsequent retrieval and personalized suggestion generation processes. This ensures that the suggestions provided to customers are no longer a "one-size-fits-all" answer, but rather unique and customized suggestions closely integrated with specific data such as the current project's load, location, and equipment, providing strong practical guidance.

[0088] It should be understood that the sequence number of each step in the above embodiments does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0089] The following are device embodiments of the present invention. For details not described in detail, please refer to the corresponding method embodiments described above.

[0090] Figure 2 The diagram shows a schematic representation of a processing device for business expansion and installation projects according to an embodiment of the present invention. For ease of explanation, only the parts relevant to the embodiment of the present invention are shown, and are described in detail below: like Figure 2 As shown, the processing device for business expansion application projects includes: a knowledge graph construction module 21, a parsing module 22, and a retrieval module 23.

[0091] The knowledge graph construction module 21 is used to extract knowledge and entities from multi-source heterogeneous data, and to establish the relationship between entities based on knowledge and entities to construct the knowledge graph for business expansion application. Parsing module 22 is used to receive natural language questions input by the user, parse the natural language questions, and obtain the question intent and key elements, including key entities or key attributes; The retrieval module 23 is used to perform multi-level matching retrieval in the knowledge graph based on the question intent, key elements and business expansion application project context information, and obtain the result set.

[0092] In one possible implementation, after performing multi-level matching retrieval on the knowledge graph based on the question intent, key elements, and contextual information of the business expansion application project, the retrieval module 23 is further used to: Based on the result set and actual data from business expansion and installation projects, targeted personalized suggestions are generated and output. Obtain user feedback on personalized suggestions to optimize subsequent retrieval and personalized suggestion generation processes.

[0093] In one possible implementation, the retrieval module 23 performs multi-level matching retrieval in the knowledge graph based on the question intent, key elements, and contextual information of the business expansion application project. When obtaining the result set, it is used for: Based on the question intent, key elements, and contextual information of the business expansion application project, the retrieval is performed sequentially in the knowledge graph through semantic retrieval, project attribute filtering, real-time context matching and ranking, and cross-case reasoning and enhancement layers to obtain the result set.

[0094] In one possible implementation, when the retrieval module 23 performs semantic retrieval in the knowledge graph, it is used for: Based on the question intent and key entities, a preliminary search is performed in the knowledge graph to obtain a set of knowledge related to the key entities, which is output as a coarse-grained relevant result set.

[0095] In one possible implementation, when the retrieval module 23 performs a retrieval of the item attribute filtering layer in the knowledge graph, it is used for: In the coarse-grained relevance result set, based on the voltage level of the project in the project profile parameters of the business expansion application project context information, policies and design schemes that are not applicable to the voltage level are filtered out to obtain the first relevance result set; Based on the installation capacity in the profile parameters of the business expansion application project, the policy clauses in the first relevant result set are matched, and the matched clauses are marked to obtain the second relevant result set; In the second relevant results set, based on the electricity usage characteristics in the profile parameters of the business expansion application project, the design cases are sorted according to their relevance to "important loads" to obtain the result set after project attribute filtering.

[0096] In one possible implementation, when the retrieval module 23 performs real-time context matching and ranking layer retrieval in the knowledge graph, it is used for: After determining the project addresses and power supply locations of all successfully implemented historical cases in the result set after filtering project attributes, calculate the spatial distance between the address of the business expansion application project and the project address and power supply location, increase the recommendation priority of the corresponding solution with the closest spatial distance, and obtain the first optimized result set; The recommendation priority of historical solutions in the first optimization result set that have a current surrounding power point real-time load rate exceeding a preset threshold is reduced, and a prompt message is generated to obtain the second optimization result set. The context of all parallel historical cases in the second optimization result set is retrieved to detect the resource usage of the access interval. If a predetermined access interval exists, the interval is automatically excluded and a resource usage prompt is generated to obtain the third optimization result set. By combining the user's expected time with the status of material inventory and the equipment procurement cycle, the historical solutions in the third optimization result set are ranked according to their time feasibility, resulting in a fine-grained result set.

[0097] In one possible implementation, when the retrieval module 23 performs cross-case reasoning and enhancement layer retrieval on the knowledge graph, it is used for: By analyzing the commonalities and identifying the differences in historical cases within the fine-grained results set, the final results set is obtained.

[0098] In one possible implementation, when the retrieval module 23 generates and outputs targeted personalized suggestions based on the result set and actual data of the business expansion and installation project, it is used for: Get the predefined business rule engine; Information related to business expansion and installation projects is extracted from the results set. The extracted information is then matched with the actual data of the business expansion and installation projects. Based on the matching results and business rules, the engine generates targeted and personalized suggestions. The extracted information includes: policy standards and clauses, historical successful cases, technical adaptation solutions, and resource allocation information.

[0099] In one possible implementation, the business rule engine adopts IF-THEN rules. The IF condition is based on the policy threshold, case adaptation conditions and resource constraint parameters set in the result set. The reasoning in the IF-THEN rule is based on the relationship between entities in the knowledge graph. It combines the historical relationship data in the result set with the characteristics of the business expansion application project to make logical deductions. The THEN conclusion corresponds to the specific solution requirements, recommended options or precautions.

[0100] The above embodiments provide a processing device for business expansion and installation projects. A knowledge graph construction module extracts knowledge and entities from multi-source heterogeneous data and establishes relationships between entities based on this knowledge and entities, constructing a knowledge graph for business expansion and installation. A parsing module receives natural language questions input by the user, parses the questions to obtain the question intent and key elements, including key entities or key attributes. A retrieval module performs multi-level matching retrieval in the knowledge graph based on the question intent, key elements, and contextual information of the business expansion and installation project, obtaining a result set. In this embodiment, the knowledge graph for business expansion and installation projects is constructed from multi-source heterogeneous data, transforming fragmented knowledge scattered in archives and expert minds into structured, reusable digital assets. This avoids knowledge loss due to personnel turnover and breaks down knowledge barriers, allowing for rapid access to historical success stories, technical standards, and solutions, significantly improving knowledge reuse rates. Through the rapid retrieval capabilities and multi-level matching filtering mechanism of the knowledge graph, project personnel can quickly obtain accurate result sets without manually sifting through massive amounts of policy documents, searching historical archives, or repeatedly consulting experts, significantly reducing the time cost of information retrieval and solution demonstration. The embodiments of the present invention can not only provide accurate result sets, but also proactively prompt users with risks that they may have overlooked, better solutions or potential cost-saving opportunities based on data analysis, transforming passive querying into proactive empowerment.

[0101] In this embodiment of the invention, targeted personalized suggestions are generated and output based on the result set and actual data of the business expansion and installation project. User feedback on the personalized suggestions is obtained to optimize subsequent retrieval and personalized suggestion generation processes. This ensures that the suggestions provided to customers are no longer a "one-size-fits-all" answer, but rather unique and customized suggestions closely integrated with specific data such as the current project's load, location, and equipment, providing strong practical guidance.

[0102] Figure 3 This is a schematic diagram of a terminal provided in an embodiment of the present invention. Figure 3 As shown, the terminal 3 in this embodiment includes: a processor 30, a memory 31, and a computer program 32 stored in the memory 31 and executable on the processor 30. When the processor 30 executes the computer program 32, it implements the steps in the processing method embodiments for the various business expansion application projects described above, for example... Figure 1 Steps 101 to 103 are shown. Alternatively, when processor 30 executes computer program 32, it implements the functions of each module / unit in the above-described device embodiments, for example... Figure 2 The functions of each module / unit are shown.

[0103] For example, computer program 32 can be divided into one or more modules / units, one or more of which are stored in memory 31 and executed by processor 30 to complete the present invention. One or more modules / units can be a series of computer program instruction segments capable of performing a specific function, which describe the execution process of computer program 32 in terminal 3. For example, computer program 32 can be divided into... Figure 2 The modules / units shown are shown.

[0104] Terminal 3 may include, but is not limited to, processor 30 and memory 31. Those skilled in the art will understand that... Figure 3 This is merely an example of terminal 3 and does not constitute a limitation on terminal 3. It may include more or fewer components than shown, or combine certain components, or different components. For example, the terminal may also include input / output devices, network access devices, buses, etc.

[0105] The processor 30 may be a Central Processing Unit (CPU), or 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. A general-purpose processor may be a microprocessor or any conventional processor.

[0106] The memory 31 can be an internal storage unit of the terminal 3, such as a hard disk or RAM of the terminal 3. The memory 31 can also be an external storage device of the terminal 3, such as a plug-in hard disk, Smart MediaCard (SMC), Secure Digital (SD) card, or Flash Card equipped on the terminal 3. Furthermore, the memory 31 can include both internal and external storage units of the terminal 3. The memory 31 is used to store computer programs and other programs and data required by the terminal. The memory 31 can also be used to temporarily store data that has been output or will be output.

[0107] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the above-described division of functional units and modules is merely an example. In practical applications, the above functions can be assigned to different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above. The functional units and modules in the embodiments can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit. Furthermore, the specific names of the functional units and modules are only for easy differentiation and are not intended to limit the scope of protection of this application. The specific working process of the units and modules in the above system can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0108] In the above embodiments, the descriptions of each embodiment have different focuses. For parts that are not described in detail or recorded in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0109] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this invention.

[0110] In the embodiments provided by this invention, it should be understood that the disclosed devices / terminals and methods can be implemented in other ways. For example, the device / terminal embodiments described above are merely illustrative. For instance, the division of modules or units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between devices or units may be electrical, mechanical, or other forms.

[0111] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0112] Furthermore, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0113] If integrated modules / units are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, all or part of the processes in the above embodiments of the present invention can also be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the above embodiments of the processing methods for various business expansion and installation projects. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. The computer-readable medium can include: any entity or device capable of carrying computer program code, recording media, USB flash drives, portable hard drives, magnetic disks, optical disks, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signals, telecommunication signals, and software distribution media, etc.

[0114] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be included within the protection scope of the present invention.

Claims

1. A method for processing business expansion application projects, characterized in that, include: Knowledge and entities are extracted from multi-source heterogeneous data, and the relationships between entities are established based on the knowledge and entities to construct a knowledge graph for business expansion and installation. Receive a natural language question input by the user, parse the natural language question, and obtain the question intent and key elements, including key entities or key attributes; Based on the stated question intent, the stated key elements, and the contextual information of the business expansion and installation project, a multi-level matching retrieval is performed in the knowledge graph to obtain a result set.

2. The method for processing business expansion and installation projects according to claim 1, characterized in that, After performing multi-level matching retrieval on the knowledge graph based on the stated question intent, the key elements, and the context information of the business expansion and installation project, and obtaining the result set, the process further includes: Based on the result set and actual data of the business expansion and installation projects, targeted personalized suggestions are generated and output. Obtain user feedback on the personalized suggestions to optimize subsequent retrieval and personalized suggestion generation processes.

3. The method for processing business expansion and installation projects according to claim 1 or 2, characterized in that, Based on the stated question intent, the stated key elements, and the contextual information of the business expansion and installation project, a multi-level matching retrieval is performed in the knowledge graph to obtain a result set, including: Based on the stated question intent, the stated key elements, and the contextual information of the business expansion application project, the semantic retrieval layer, the project attribute filtering layer, the real-time context matching and ranking layer, and the cross-case reasoning and enhancement layer are sequentially performed in the knowledge graph to obtain a result set.

4. The method for processing business expansion and installation projects according to claim 3, characterized in that, The semantic retrieval layer of the knowledge graph includes: Based on the question intent and the key entity, a preliminary search is performed in the knowledge graph to obtain a set of knowledge related to the key entity, which is output as a coarse-grained related result set.

5. The method for processing business expansion and installation projects according to claim 4, characterized in that, The retrieval of the project attribute filtering layer in the knowledge graph includes: In the coarse-grained correlation result set, based on the voltage level of the project in the project profile parameters of the business expansion application project context information, policies and design schemes that are not applicable to the voltage level are filtered out to obtain the first correlation result set; Based on the installation capacity in the profile parameters of the business expansion application project, the policy clauses in the first relevant result set are matched, and the matched clauses are marked to obtain the second relevant result set; In the second relevant result set, based on the electricity usage characteristics in the profile parameters of the business expansion application project, the design cases are sorted according to their relevance to "important loads" to obtain the result set after project attribute filtering.

6. The method for processing business expansion and installation projects according to claim 5, characterized in that, The retrieval of the knowledge graph in real-time context matching and ranking layer includes: Determine the project address and power point location of all successfully implemented historical cases in the result set after the project attribute filtering, calculate the spatial distance between the address of the business expansion application project and the project address and power point location, increase the recommendation priority of the corresponding solution with the closest spatial distance, and obtain the first optimized result set; The recommendation priority of historical solutions in all historical cases in the first optimization result set that have a current surrounding power point real-time load rate exceeding a preset threshold is reduced, and a prompt message is generated to obtain the second optimization result set. The context of all parallel historical cases in the second optimization result set is retrieved to detect the access interval resource usage. If a predetermined access interval exists, the interval is automatically excluded and a resource usage prompt is generated to obtain the third optimization result set. By combining the user's expected time with the material inventory status and equipment procurement cycle, the historical solutions in the third optimization result set are ranked according to their time feasibility, resulting in a fine-grained result set.

7. The method for processing business expansion and installation projects according to claim 6, characterized in that, The knowledge graph is subjected to cross-case reasoning and enhancement layer retrieval, including: By performing commonality mining and difference alerts on historical cases in the fine-grained result set, the final result set is obtained.

8. The method for processing business expansion and installation projects according to claim 2, characterized in that, Based on the aforementioned result set and actual data from business expansion and installation projects, targeted personalized suggestions are generated and output, including: Get the predefined business rule engine; Information related to the business expansion and installation project is extracted from the result set. The extracted information is matched with the actual data of the business expansion and installation project. Based on the matching results and the business rule engine, targeted personalized suggestions are generated. The extracted information includes: policy standard clauses, historical successful cases, technical adaptation solutions and resource allocation information.

9. The method for processing business expansion and installation projects according to claim 8, characterized in that, The business rule engine adopts IF-THEN rules. The IF condition is set based on the policy threshold, case adaptation conditions and resource constraint parameters in the result set. The reasoning in the IF-THEN rule is based on the relationship between entities in the knowledge graph. It combines the historical relationship data in the result set with the characteristics of the business expansion application project to make logical deductions. The THEN conclusion corresponds to specific solution requirements, recommended options or precautions.

10. A processing device for business expansion and installation projects, characterized in that, include: The knowledge graph construction module is used to extract knowledge and entities from multi-source heterogeneous data, and establish the relationship between entities based on the knowledge and entities to construct the knowledge graph for business expansion and installation. The parsing module is used to receive natural language questions input by the user, parse the natural language questions, and obtain the question intent and key elements, including key entities or key attributes. The retrieval module is used to perform multi-level matching retrieval in the knowledge graph based on the question intent, the key elements, and the context information of the business expansion application project, and obtain a result set.

Images