GEO optimization method and system based on medical knowledge graph and user intention driving

Abstract

The invention discloses a GEO optimization method and system based on a medical knowledge map and user intention driving, and relates to the technical field of medical data acquisition, and the method comprises the following steps: S1, user input perception: receiving and analyzing a medical health request of a user, and extracting a symptom entity and a preliminary intention in the medical health request according to the medical health request; and S2, deep analysis: carrying out interactive reasoning on the symptom entity and the preliminary intention with a medical knowledge graph. Deep semantic analysis and intention expansion are carried out on user query through a medical knowledge graph, accurate alignment between requirements and geographic medical resources on the knowledge level is achieved, fuzzy and surface appeals of users are converted into accurate and structured medical requirements, the medical requirements are matched with deep capabilities of the resources, and the requirements of the users are matched with the deep capabilities of the resources. The clinical relevance and the user satisfaction degree of the recommendation result are greatly improved, and through a multi-objective optimization framework, the nearest distance is considered, and a personalized, dynamic and optimal medical action path is generated.

Description

Technical Field

[0001] This invention relates to the field of medical data acquisition, specifically to a GEO optimization method and system based on medical knowledge graphs and user intent. Background Technology

[0002] With the aging population and increased public health awareness, the demand for healthcare services is growing and becoming increasingly complex. When faced with disease symptoms or health problems, users expect efficient and accurate guidance throughout the entire process, from disease awareness and diagnostic advice to actual medical treatment. Traditional medical information services and resource matching models primarily rely on general search engines, static medical information portals, or independent appointment booking platforms. These have significant limitations in meeting users' complex, dynamic, and personalized healthcare needs. For example:

[0003] 1. Users primarily search for information through general search engines or vertical medical platforms. These methods mainly rely on keyword matching and lack a deep understanding of the user's true intent and semantic context. When a user enters "headache, fever," the system may return a massive amount of mixed information ranging from the common cold to meningitis. Users cannot determine the severity of the symptoms or what to do next. Although some platforms have introduced simple symptom self-check tools, the medical logic behind them is often based on static rules or shallow databases, lacking a large-scale, structured, and reasonable medical knowledge graph as support. This makes it impossible to perform multiple inferences, resulting in insufficient accuracy and depth of suggestions.

[0004] 2. Existing medical resource query services primarily focus on the retrieval and display of static information. They severely lack awareness of the real-time dynamic resource status of medical institutions and are unable to intelligently match this resource information with the user's specific medical intentions and plan pathways. For example, for a patient suspected of having acute chest pain, the optimal path is not simply to recommend the nearest tertiary hospital, but rather to comprehensively consider dynamic factors such as the real-time reception capacity of the hospital's chest pain center, the equipment status of the catheterization lab, and the estimated arrival time, generating an emergency action sequence that includes "calling 120, being sent to a hospital with interventional capabilities, and directly entering the green channel." Existing systems are deficient in this type of multi-objective, multi-constraint, and dynamic collaborative decision-making.

[0005] 3. Currently, most recommendation systems lack proactive screening of potential risks in user queries, and also lack real-time compliance verification and filtering mechanisms for recommended resources, which may lead users to unsafe or mismatched services, causing medical risks. Summary of the Invention

[0006] The purpose of this invention is to provide a GEO optimization method and system based on medical knowledge graphs and user intent-driven approaches, thereby solving the problems mentioned in the background art.

[0007] To solve the above-mentioned technical problems, the present invention is achieved through the following technical solution:

[0008] This invention provides a GEO optimization method and system based on medical knowledge graphs and user intent-driven approaches.

[0009] In a first aspect, the present invention provides a technical solution: a GEO optimization method based on medical knowledge graph and user intent-driven approach, characterized by comprising the following steps:

[0010] Step 1, User Input Perception: Receive and parse the user's medical and health requests, and extract the symptom entities and preliminary intents from the medical and health requests;

[0011] Step 2, In-depth analysis: Interact and reason with the symptom entities and initial intents with the medical knowledge graph to generate a structured user intent enhancement graph;

[0012] Step 3, Knowledge Graph Alignment: Obtain real-time medical resource status information from the geographic information system, construct a dynamic medical resource graph, and semantically match the user intent enhancement graph with the dynamic medical resource graph to obtain a candidate resource set;

[0013] Step 4: Personalized Path Generation: Based on the priority of medical needs in the enhanced user intent graph, select the optimal path with the corresponding weight from the candidate resource set;

[0014] Step 5: Presentation and Feedback: Visualize the optimal path to the user and receive user feedback data to optimize subsequent path generation.

[0015] Furthermore, in step one, the healthcare request includes natural language queries, structured options, and contextual information.

[0016] Furthermore, in step two, the medical knowledge graph includes symptom entities, disease entities, drug entities, department entities, and examination item entities, and the user intent enhancement graph includes inferred suspected disease sets, recommended treatment action sequences, and medical need priorities.

[0017] Furthermore, in step three, the dynamic medical resource map includes medical institution status attributes, real-time service status, and spatial location information.

[0018] Furthermore, in step four, the optimal path is a sequence containing multiple ordered medical action nodes, the types of which include going to a specific medical institution, receiving specific medical services, and obtaining specific medicines.

[0019] Secondly, according to the first aspect above, the present invention also provides a technical solution: a GEO optimization system based on medical knowledge graph and user intent driven, comprising:

[0020] Central Processing Unit (CPU): Used for data processing during system operation;

[0021] User intent parsing module: used to receive and parse the user's original query request, and use medical knowledge graph to complete, disambiguate and standardize the user's unstructured medical needs;

[0022] Medical knowledge query module: Utilizes professional medical knowledge graphs to perform in-depth reasoning and expansion of user intent, generating resource demand information;

[0023] Geographic Information Module: Based on the resource demand information, it filters, weighs, and sorts the most matching medical resource entities from a massive amount of geographic information.

[0024] Risk perception and security barrier module: used for risk screening at each stage, providing early warning of potential medical risks, and blocking and downgrading candidate geographic resources that do not meet safety and compliance conditions to ensure the accuracy, safety and compliance of the recommendation results;

[0025] GEO Resource Awareness and Matching Module: Used to construct a dynamic medical resource map and complete semantic matching between user intent and medical resources;

[0026] Multi-objective optimization path planning module: Based on the urgency of the user's intent, it transforms matched resources into executable optimal action plans;

[0027] Interaction and Learning Module: Used for human-computer interaction, providing recommendation results, and collecting feedback data for system iteration and optimization;

[0028] The risk perception and security barrier module includes a real-time intent risk scanning module, a resource compliance authentication and filtering module, an interactive risk warning and awareness prompt module, an emergency path security intervention module, and an audit and traceability log module. These modules are all bidirectionally connected. The GEO resource perception and matching module includes a multi-source data access module, a resource knowledge graph module, a multi-dimensional feature similarity calculation module, and an elastic resource discovery module. These modules are also bidirectionally connected.

[0029] The output of the risk perception and security barrier module is connected to the input of the central processing unit via a bidirectional signal.

[0030] Furthermore, the user intent parsing module includes a natural language understanding module, a context fusion module, a medical knowledge graph query module, and a structured option module. These modules are all bidirectionally connected. The interaction and learning module includes a personalized recommendation presentation module, a conversational adjustment interface module, a closed-loop data collection module, and an online learning optimization module. These modules are also bidirectionally connected.

[0031] Furthermore, the bidirectional signal output of the user intent parsing module is connected to the input of the central processing unit, the bidirectional signal output of the multi-objective optimized path planning module is connected to the input of the central processing unit, and the bidirectional signal output of the geographic information module is connected to the input of the central processing unit.

[0032] Furthermore, the output signal of the user intent parsing module is connected to the input of the medical knowledge query module; the bidirectional output signal of the medical knowledge query module is connected to the input of the GEO resource perception and matching module; the bidirectional output signal of the GEO resource perception and matching module is connected to the input of the geographic information module; the bidirectional output signal of the geographic information module is connected to the input of the multi-objective optimization path planning module; the output signal of the user intent parsing module is connected to the input of the GEO resource perception and matching module; the bidirectional output signal of the multi-objective optimization path planning module is connected to the input of the risk perception and safety barrier module; the bidirectional output signal of the multi-objective optimization path planning module is connected to the input of the interaction and learning module; and the output signal of the interaction and learning module is connected to the input of the user intent parsing module.

[0033] A computer-readable storage medium storing a processor-executable program, which, when executed by a processor, is used to perform the method as described in any of the preceding claims.

[0034] This invention provides a GEO optimization method and system based on medical knowledge graphs and user intent-driven approaches. It offers the following advantages:

[0035] (1) The GEO optimization method and system based on medical knowledge graph and user intent drive performs deep semantic analysis and intent expansion on user queries through medical knowledge graph, realizes accurate alignment between demand and geographic medical resources at the knowledge level, transforms the vague and superficial demands of users into accurate and structured medical needs, and matches them with the deep capabilities of resources, which greatly improves the clinical relevance of recommendation results and user satisfaction.

[0036] (2) The GEO optimization method and system based on medical knowledge graph and user intent drive, through a multi-objective optimization framework, not only considers the nearest distance, but also comprehensively considers the urgency of medical intent, real-time resource load, and continuity of treatment path, to generate personalized and dynamically optimal medical action path, upgrading static geographical location recommendation to dynamic, personalized, and multi-step intelligent medical action plan, and mining insights from massive user intent to achieve a leap from individual service to collective intelligence.

[0037] (3) The GEO optimization method and system based on medical knowledge graph and user intent, through the use of risk perception and security barrier modules, deeply integrates medical safety and compliance requirements into the geographic optimization process, reduces medical risks caused by information asymmetry for users, and at the same time helps regulatory compliance and builds a trustworthy medical search environment.

[0038] (4) The GEO optimization method and system based on medical knowledge graph and user intent drive guides patients to be rationally diverted through global optimization and load balancing, alleviates the instantaneous congestion of high-quality medical resources, revitalizes the overall medical resources in the region, shortens the rescue response time for critically ill patients, reduces the comprehensive management cost of patients with chronic diseases, and reduces the risk of ineffective travel and misdiagnosis through precise matching. Attached Figure Description

[0039] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the description of the embodiments 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.

[0040] Figure 1 The flowchart shows the GEO optimization method and system based on medical knowledge graph and user intent driven by the present invention.

[0041] Figure 2 This is the overall system diagram of the GEO optimization method and system based on medical knowledge graph and user intent driven by the present invention;

[0042] Figure 3 This is a schematic diagram of the user intent parsing module of the GEO optimization method and system based on medical knowledge graph and user intent driven by the present invention.

[0043] Figure 4 This is a schematic diagram of the risk perception and security barrier module of the GEO optimization method and system based on medical knowledge graph and user intent driven by the present invention.

[0044] Figure 5This is a schematic diagram of the GEO resource perception and matching module of the GEO optimization method and system based on medical knowledge graph and user intent driven by the present invention.

[0045] Figure 6 This is a schematic diagram of the interaction and learning module of the GEO optimization method and system based on medical knowledge graph and user intent driven by the present invention.

[0046] In the diagram: 1. Central Processing Unit; 2. User Intent Parsing Module; 3. Medical Knowledge Query Module; 4. Geographic Information Module; 5. Risk Perception and Security Barrier Module; 6. GEO Resource Perception and Matching Module; 7. Multi-Objective Optimization Path Planning Module; 8. Interaction and Learning Module. Detailed Implementation

[0047] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments.

[0048] Examples of the embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and intended to explain the invention, and should not be construed as limiting the invention.

[0049] The present invention will be further described below with reference to the accompanying drawings and embodiments:

[0050] Please see Figure 1-6 As shown, a GEO optimization method and system based on medical knowledge graph and user intent-driven approach are presented.

[0051] Firstly, please refer to Figure 1 As shown, this invention provides an implementation scheme, a GEO optimization method based on medical knowledge graph and user intent driven, characterized by the following steps:

[0052] Step 1, User Input Awareness: Receive and parse the user's medical and health requests, extract the symptom entities and preliminary intents from the medical and health requests, which include natural language queries, structured options, and contextual information.

[0053] We perform basic natural language processing on the natural language, extract key symptom entities and basic intent classifications, and integrate contextual information to form an initial user request vector.

[0054] Step 2, In-depth analysis: Interact and reason with the symptom entities and initial intents with the medical knowledge graph to generate a structured user intent enhancement graph. The medical knowledge graph includes symptom entities, disease entities, drug entities, department entities, and examination item entities. The user intent enhancement graph includes the inferred set of suspected diseases, recommended treatment action sequences, and medical need priorities.

[0055] Interactive reasoning involves inferring at least one related disease entity or department entity based on symptom entities and through relational paths in a knowledge graph, thereby expanding the initial intent label.

[0056] Step 3: Knowledge Graph Alignment: Obtain real-time medical resource status information from the geographic information system, construct a dynamic medical resource graph, and perform semantic matching between the user intent augmentation graph and the dynamic medical resource graph to obtain a set of candidate resources. The dynamic medical resource graph includes medical institution status attributes, real-time service status, and spatial location information.

[0057] The status attributes of medical institutions include real-time dynamic information and static attribute information. Real-time dynamic information includes at least the real-time waiting number of each department of the medical institution, the availability status of medical equipment, the real-time inventory of medicines, and the real-time location and status of emergency vehicles. Static attribute information includes at least the level of the medical institution, the department setup, and the expertise of specialists.

[0058] Step 4: Personalized Path Generation: Based on user intent, enhance the priority of medical needs in the graph, select the optimal path with corresponding weight from the candidate resource set. The optimal path is a sequence containing multiple ordered medical action nodes. The types of ordered medical action nodes include going to a specific medical institution, receiving specific medical services, and obtaining specific medicines.

[0059] The selection of the optimal path is based on a multi-objective optimization function, which generates at least one comprehensive optimal path that includes time sequence, spatial movement, and medical action. The objective variables of the multi-objective optimization function include at least user arrival time, economic cost, medical service quality score, and resource load balance. Medical demand priority is used to dynamically adjust the weight coefficients of each objective variable. When the user intent enhancement graph includes emergency medical needs, the weight coefficient of the user arrival time variable is set to the highest priority. The optimal path solution process must satisfy the medical logic constraints and spatiotemporal accessibility constraints between nodes.

[0060] Based on user intent, prioritize the enhancement graph and dynamically set optimization target weights, including:

[0061] (1) Emergency scenario:

[0062] Objective function = min(α*arrival time + β*waiting time inside the hospital)

[0063] Among them, α has an extremely high weight.

[0064] (2) Chronic disease management scenario:

[0065] Objective function = min(γ*transportation cost + δ*economic cost - ε*service quality score)

[0066] Consider the order among multiple destinations.

[0067] By employing operations research algorithms (such as the improved Dijkstra algorithm and multi-agent reinforcement learning), one or more optimal action paths can be generated by searching the dynamic resource graph. These paths include the order of movement, suggested modes of transportation, and estimated time points.

[0068] Step 5, Presentation and Feedback: Visualize the optimal path to the user and receive user feedback data to optimize subsequent path generation.

[0069] The optimized path is clearly presented to the user in the form of a visual map and step-by-step guidance cards, and the reasons for the recommendation are explained. Users are allowed to make fine adjustments based on their own preferences. The system then re-executes the third and fourth steps for rapid re-optimization. Users follow the path, and the system collects anonymized execution feedback data through authorization, forming a closed-loop data flow.

[0070] Secondly, please refer to Figure 2-6 As shown, according to the implementation scheme of the first aspect above, the present invention also provides an implementation scheme, a GEO optimization system based on medical knowledge graph and user intent driven, comprising:

[0071] Central Processing Unit 1: Used for data processing during system operation;

[0072] User Intent Parsing Module 2: Used to receive and parse the user's original query request, and use medical knowledge graph to complete, disambiguate and standardize the user's unstructured medical needs;

[0073] Medical knowledge query module 3: Utilizes a professional medical knowledge graph to perform in-depth reasoning and expansion of user intent, generating resource demand information;

[0074] Geographic Information Module 4: Based on resource demand information, filter, weigh, and rank the most matching medical resource entities from a massive amount of geographic information;

[0075] Risk Perception and Security Barrier Module 5: Used for risk screening at each stage, providing early warnings of potential medical risks, and blocking and downgrading candidate geographic resources that do not meet safety and compliance conditions to ensure the accuracy, safety and compliance of the recommendation results;

[0076] GEO Resource Awareness and Matching Module 6: Used to construct a dynamic medical resource map and complete the semantic matching between user intent and medical resources;

[0077] Multi-objective optimization path planning module 7: Based on the urgency of the user's intent, the matched resources are transformed into an executable optimal action plan;

[0078] Interaction and Learning Module 8: Used for human-computer interaction, providing recommendation results, and collecting feedback data for system iteration and optimization;

[0079] The Risk Perception and Security Barrier Module 5 includes a real-time intent risk scanning module, a resource compliance authentication and filtering module, an interactive risk warning and awareness prompt module, an emergency path security intervention module, and an audit and traceability log module. These modules are all bidirectionally connected. The GEO Resource Perception and Matching Module 6 includes a multi-source data access module, a resource knowledge graph module, a multi-dimensional feature similarity calculation module, and an elastic resource discovery module. These modules are also bidirectionally connected.

[0080] The bidirectional signal output of the risk perception and safety barrier module 5 is connected to the input of the central processing unit 1.

[0081] In Module 5 of Risk Perception and Security Barrier:

[0082] Real-time intent risk scanning module: Utilizes the risk rule base in the knowledge graph (such as drug contraindications and high-risk symptom associations) to conduct real-time risk assessment of user queries. For example, it can identify the purchase intent of a "pregnant" user who queries "a certain teratogenic drug" and trigger an advanced alert.

[0083] Resource compliance certification and filtering module: When the decision engine generates a list of candidate resources, this module enforces qualification verification and automatically filters out unlicensed, unauthorized, or penalized institutions to ensure the underlying compliance of the recommended list.

[0084] Interactive risk warning and informed consent module: For low to medium risk situations where users need to be aware, structured prompts are generated. For example, when a user searches for medical aesthetic procedures, the "qualifications of the attending physician for medical aesthetic procedures" of the institution are automatically highlighted, or when prescription drugs are recommended, a warning message "Please purchase with a doctor's prescription" is forcibly displayed.

[0085] Emergency Pathway Safety Intervention Module: Specifically designed to handle acute and critical illness intentions (such as "chest pain" or "stroke"), it integrates disease-treatment capability mapping in the knowledge graph (such as chest pain center certification) to prioritize guiding patients to hospitals with corresponding treatment capabilities and may block clinics that are nearby but lack treatment conditions, thus avoiding the risk of "misdirection" from a medical ethics perspective.

[0086] Audit and traceability log module: Completely records the risk rules, filtering actions, warning information and final recommendation reasons triggered by each query, forming an immutable audit log, serving system optimization, and providing a transparent and traceable chain of evidence for potential compliance reviews.

[0087] In GEO Resource Awareness Matching Module 6:

[0088] The multi-source data access module includes: hospital HIS system interface (real-time appointment availability, waiting time), drug inventory monitoring system, ambulance GPS tracking system, public health data platform (epidemic and epidemiological data), and IoT device data stream (status of smart medical devices).

[0089] The resource knowledge graph module includes a static attribute graph, a dynamic state graph, and a spatiotemporal accessibility graph. The static attribute graph includes institutional qualifications, departmental capabilities, and expert profiles; the dynamic state graph includes real-time load, equipment availability, and service capabilities; and the spatiotemporal accessibility graph includes traffic conditions and a distance-time conversion model.

[0090] Multidimensional feature similarity calculation module: Matching score = α * medical relevance + β * spatiotemporal accessibility + γ * service quality + δ * economy.

[0091] Elastic resource discovery module: generates alternative solutions when the primary resource is unavailable, and discovers cross-organizational collaborative paths (such as examination A + treatment B).

[0092] Furthermore, the user intent parsing module 2 includes a natural language understanding module, a context fusion module, a medical knowledge graph query module, and a structured option module. The natural language understanding module, context fusion module, medical knowledge graph query module, and structured option module are all bidirectionally connected. The interaction and learning module 8 includes a personalized recommendation presentation module, a conversational adjustment interface module, a closed-loop data collection module, and an online learning optimization module. The personalized recommendation presentation module, conversational adjustment interface module, closed-loop data collection module, and online learning optimization module are all bidirectionally connected.

[0093] Furthermore, the bidirectional signal at the output of the user intent parsing module 2 is connected to the input of the central processing unit 1, the bidirectional signal at the output of the multi-objective optimized path planning module 7 is connected to the input of the central processing unit 1, and the bidirectional signal at the output of the geographic information module 4 is connected to the input of the central processing unit 1.

[0094] Furthermore, the output signal of the user intent parsing module 2 is connected to the input of the medical knowledge query module 3, the bidirectional output signal of the medical knowledge query module 3 is connected to the input of the GEO resource perception and matching module 6, the bidirectional output signal of the GEO resource perception and matching module 6 is connected to the input of the geographic information module 4, the bidirectional output signal of the geographic information module 4 is connected to the input of the multi-objective optimization path planning module 7, the output signal of the user intent parsing module 2 is connected to the input of the GEO resource perception and matching module 6, the bidirectional output signal of the multi-objective optimization path planning module 7 is connected to the input of the risk perception and safety barrier module 5, the bidirectional output signal of the multi-objective optimization path planning module 7 is connected to the input of the interaction and learning module 8, and the output signal of the interaction and learning module 8 is connected to the input of the user intent parsing module 2.

[0095] At the same time, the present invention also proposes a computer-readable storage medium storing a processor-executable program, characterized in that the processor-executable program, when executed by a processor, is used to perform the method as described above.

[0096] The above are merely preferred embodiments of the present invention. It should be noted that those skilled in the art can make various modifications and improvements without departing from the inventive concept of the present invention, and these all fall within the protection scope of the present invention.

Claims

1. A GEO optimization method based on medical knowledge graph and user intent-driven approach, characterized in that, Includes the following steps: S1. User Input Perception: Receive and parse the user's medical and health requests, and extract the symptom entities and preliminary intentions from the medical and health requests; S2. Deep Analysis: Interact and reason with the symptom entities and initial intents with the medical knowledge graph to generate a structured user intent enhancement graph; S3. Knowledge Graph Alignment: Obtain real-time medical resource status information from the geographic information system, construct a dynamic medical resource graph, and semantically match the user intent enhancement graph with the dynamic medical resource graph to obtain a candidate resource set; S4. Personalized Path Generation: Based on the priority of medical needs in the enhanced user intent graph, select the optimal path with the corresponding weight from the candidate resource set; S5. Presentation and Feedback: Visualize the optimal path to the user and receive user feedback data to optimize subsequent path generation.

2. The GEO optimization method based on medical knowledge graph and user intent driven according to claim 1, characterized in that, In step S1, the medical and health request includes natural language queries, structured options, and contextual information.

3. The GEO optimization method based on medical knowledge graph and user intent driven according to claim 1, characterized in that, In step S2, the medical knowledge graph includes symptom entities, disease entities, drug entities, department entities, and examination item entities, and the user intent enhancement graph includes inferred suspected disease sets, recommended treatment action sequences, and medical need priorities.

4. The GEO optimization method based on medical knowledge graph and user intent driven according to claim 1, characterized in that, In step S3, the dynamic medical resource map includes medical institution status attributes, real-time service status, and spatial location information.

5. The GEO optimization method based on medical knowledge graph and user intent driven according to claim 1, characterized in that, In step S4, the optimal path is a sequence containing multiple ordered medical action nodes, the types of which include going to a specific medical institution, receiving specific medical services, and obtaining specific medicines.

6. A GEO optimization system based on medical knowledge graph and user intent-driven approach, characterized in that: The method applied to the GEO optimization method based on medical knowledge graph and user intent driven by any one of claims 1-5 includes: Central Processing Unit (1): Used for data processing during system operation; User intent parsing module (2): used to receive and parse the user's original query request, and use medical knowledge graph to complete, disambiguate and standardize the user's unstructured medical needs; Medical knowledge query module (3): Utilizes professional medical knowledge graphs to perform in-depth reasoning and expansion of user intent, and generates resource demand information; Geographic Information Module (4): Based on the resource demand information, the most matching medical resource entities are selected, weighed and ranked from a large amount of geographic information; Risk perception and security barrier module (5): used for risk screening at each stage, to provide early warning of potential medical risks, and to intercept and downgrade candidate geographic resources that do not meet safety and compliance conditions, so as to ensure the accuracy, safety and compliance of the recommendation results; GEO Resource Awareness and Matching Module (6): Used to construct a dynamic medical resource map and complete the semantic matching between user intent and medical resources; Multi-objective optimization path planning module (7): Based on the urgency of the user's intent, the matched resources are transformed into an executable optimal action plan; Interaction and Learning Module (8): Used for human-computer interaction, providing recommendation results, and collecting feedback data for system iteration and optimization; The risk perception and security barrier module (5) includes a real-time intent risk scanning module, a resource compliance authentication and filtering module, an interactive risk warning and awareness prompt module, an emergency path security intervention module, and an audit and traceability log module. The real-time intent risk scanning module, the resource compliance authentication and filtering module, the interactive risk warning and awareness prompt module, the emergency path security intervention module, and the audit and traceability log module are all bidirectionally connected. The GEO resource perception and matching module (6) includes a multi-source data access module, a resource knowledge graph module, a multi-dimensional feature similarity calculation module, and an elastic resource discovery module. The multi-source data access module, the resource knowledge graph module, the multi-dimensional feature similarity calculation module, and the elastic resource discovery module are all bidirectionally connected. The output of the risk perception and security barrier module (5) is bidirectionally connected to the input of the central processing unit (1).

7. The GEO optimization system based on medical knowledge graph and user intent as described in claim 6, characterized in that: The user intent parsing module (2) includes a natural language understanding module, a context fusion module, a medical knowledge graph query module, and a structured option module. The natural language understanding module, context fusion module, medical knowledge graph query module, and structured option module are all bidirectionally connected. The interaction and learning module (8) includes a personalized recommendation presentation module, a conversational adjustment interface module, a closed-loop data collection module, and an online learning optimization module. The personalized recommendation presentation module, conversational adjustment interface module, closed-loop data collection module, and online learning optimization module are all bidirectionally connected.

8. The GEO optimization system based on medical knowledge graph and user intent as described in claim 6, characterized in that: The output of the user intent parsing module (2) is connected to the input of the central processing unit (1) via a bidirectional signal. The output of the multi-objective optimized path planning module (7) is connected to the input of the central processing unit (1) via a bidirectional signal. The output of the geographic information module (4) is connected to the input of the central processing unit (1) via a bidirectional signal.

9. The GEO optimization system based on medical knowledge graph and user intent as described in claim 6, characterized in that: The output signal of the user intent parsing module (2) is connected to the input of the medical knowledge query module (3). The bidirectional output signal of the medical knowledge query module (3) is connected to the input of the GEO resource perception matching module (6). The bidirectional output signal of the GEO resource perception matching module (6) is connected to the input of the geographic information module (4). The bidirectional output signal of the geographic information module (4) is connected to the input of the multi-objective optimization path planning module (7). The output signal of the user intent parsing module (2) is connected to the input of the GEO resource perception matching module (6). The bidirectional output signal of the multi-objective optimization path planning module (7) is connected to the input of the risk perception and safety barrier module (5). The bidirectional output signal of the multi-objective optimization path planning module (7) is connected to the input of the interaction and learning module (8). The output signal of the interaction and learning module (8) is connected to the input of the user intent parsing module (2).

10. A computer-readable storage medium storing a processor-executable program, characterized in that, The processor-executable program, when executed by the processor, is used to perform the method as described in any one of claims 1-5.

Images