Intelligent diagnosis method and system based on knowledge graph and large model

Abstract

The invention discloses an intelligent diagnosis method and system based on a knowledge graph and a large model, and relates to the technical field of large model application, and the method comprises the steps: constructing a medical knowledge graph; a dialogue interface is designed, and the dialogue interface utilizes a large model to guide a patient to describe symptoms and automatically analyzes and extracts chief complaint information; querying related information through the medical knowledge graph, and integrating and sequencing to form a disease condition information base; pre-training the large model, simulating doctor questions and answers based on the illness state information base, guiding the patient to describe symptoms and signs in detail, and updating the illness state information base; inquiring the medical history through a dialogue interface, and dynamically adjusting the problem; the large model comprehensively analyzes chief complaints, signs and medical history, simulates diagnosis thinking, gives preliminary diagnosis results and suggestions, and displays the results and suggestions to patients or doctors. The diagnosis accuracy and efficiency can be improved, and the problems that traditional medical diagnosis depends on doctor experience and is high in subjectivity, low in efficiency and the like are solved.

Description

Technical Field

[0001] The present invention relates to the field of large model application technology, and specifically to an intelligent diagnosis method and system based on knowledge graph and large model. Background Art

[0002] With the explosive growth of medical data and the rapid development of artificial intelligence technology, intelligent medicine has gradually become a research hotspot in the medical field. Traditional medical diagnosis methods rely on the experience and expertise of doctors, and have limitations such as strong subjectivity and low efficiency. In order to overcome these shortcomings, researchers began to explore the application of artificial intelligence technology in the field of medical diagnosis. Among them, knowledge graphs and large model technologies have become important research directions in intelligent medicine due to their powerful knowledge representation and reasoning capabilities.

[0003] Knowledge graph is a structured knowledge representation method that can clearly display the complex relationships between entities. In the medical field, medical knowledge graph can integrate massive amounts of medical data and provide comprehensive knowledge support for medical diagnosis. By constructing a medical knowledge graph, it is easy to query the association information between medical entities such as diseases, symptoms, and drugs, providing a data foundation for intelligent diagnosis.

[0004] Big model technology refers to large-scale neural network models trained using deep learning algorithms. These models have powerful feature extraction and reasoning capabilities and can handle complex nonlinear relationships and data patterns. In the field of medical diagnosis, big models can learn from massive amounts of medical data to discover potential connections between symptoms and diseases, providing decision support for intelligent diagnosis.

[0005] However, a single knowledge graph or large model technology still has limitations in intelligent medical diagnosis. Although the knowledge graph can represent the relationship between medical knowledge, it lacks intelligent reasoning capabilities; and although the large model can perform intelligent reasoning, it often lacks clear medical knowledge support. Therefore, combining the knowledge graph with the large model technology to construct an intelligent diagnosis method based on the knowledge graph and the large model has important research significance and application value. Summary of the invention

[0006] In order to solve the problems of traditional medical diagnosis, which relies on doctors' experience, is highly subjective and inefficient, the present invention provides an intelligent diagnosis method and system based on knowledge graph and large model, aiming to improve the accuracy and efficiency of medical diagnosis.

[0007] In the first aspect, the present invention provides an intelligent diagnosis method based on knowledge graph and large model. The technical solution adopted to solve the above technical problems is as follows:

[0008] An intelligent diagnosis method based on knowledge graph and large model includes the following steps:

[0009] S1. Build a medical knowledge graph containing medical entities and the relationships between them, where the medical entities include diseases, symptoms, signs, and treatments; pre-train a large model with powerful natural language processing capabilities that can realize professional question-answering and diagnosis functions in the medical field;

[0010] S2. Design a dialogue interface that uses the multi-round dialogue capability of the big model to guide patients to describe their symptoms, uses the natural language processing technology of the big model to automatically parse and extract the patient's symptom description, and extracts the chief complaint information in combination with the medical knowledge graph;

[0011] S3. Use the association query capability of the medical knowledge graph to find relevant information about the chief complaint in the medical knowledge graph; integrate and sort the relevant information found to form a comprehensive and orderly patient condition information database;

[0012] S4. Based on the patient's condition information database, the large model is used to simulate the doctor's professional question-and-answer method to guide the patient to further describe his or her symptoms and changes in the course of the disease in detail, and the acquired information is updated to the patient's condition information database; then, the large model is used to simulate the doctor's professional question-and-answer method to inquire about the patient's current physical signs;

[0013] S5. The dialogue interface asks the patient about his medical history information through a preset medical history question template. During the inquiry process, the multi-round dialogue capability of the large model is used to dynamically adjust subsequent questions based on the patient's current answer.

[0014] S6. The large model uses its powerful learning and reasoning capabilities to comprehensively analyze and judge the patient's chief complaint information, physical sign information, and medical history information, and based on medical knowledge and diagnostic rules, simulates the doctor's diagnostic thinking process and gives preliminary diagnostic results and suggestions;

[0015] S7. The large model intuitively displays its diagnostic results and suggestions to patients or doctors.

[0016] Optionally, executing step S1, the process of constructing a medical knowledge graph is as follows:

[0017] Collect various medical data resources, including clinical guidelines, medical literature, drug instructions and disease databases;

[0018] Screen and organize the collected medical data resources, remove duplicate, erroneous or irrelevant information, and form medical text data;

[0019] Use natural language processing technology (NLP) to perform pre-processing operations on medical text data, such as word segmentation, part-of-speech tagging, named entity recognition, and relationship extraction, to identify medical entities in medical texts and the relationships between medical entities, thereby converting unstructured medical text data into structured data;

[0020] Based on the identified medical entities and the relationships between them, a medical knowledge graph is constructed with medical entities as nodes and the relationships between them as edges.

[0021] Optionally, the dialogue interface designed in step S2 guides the patient to answer questions and describe his or her own symptoms through text or voice input according to a preset symptom question template; in the process of guiding the patient to answer the questions, the multi-round dialogue capability of the big model is utilized to dynamically adjust subsequent questions based on the patient's current answers, and the natural language processing technology of the big model is used to automatically parse and extract the patient's answers, and the chief complaint information is extracted by extracting key symptom description information and combining it with the medical knowledge graph.

[0022] Optionally, when pre-training a large model, the medical datasets used include general natural language text, professional medical literature and clinical data, and real conversation logs.

[0023] Optionally, the large model can visually display the diagnosis results and suggestions to the patient or doctor through diagrams or text descriptions;

[0024] The diagnostic results and recommendations of the big model include the name of the disease, the corresponding probability, and related explanations and suggestions, as well as personalized treatment recommendations and guidance based on the patient's specific circumstances.

[0025] In the second aspect, the present invention provides an intelligent diagnosis system based on knowledge graph and large model. The technical solutions adopted to solve the above technical problems are as follows:

[0026] An intelligent diagnosis system based on knowledge graph and large model, comprising:

[0027] A graph construction module, used to construct a medical knowledge graph containing medical entities and relationships between medical entities, wherein the medical entities include diseases, symptoms, signs and treatment methods;

[0028] The pre-training module is used to pre-train a large model with powerful natural language processing capabilities and capable of realizing professional question-answering and diagnosis functions in the medical field. The large model further includes:

[0029] The conversation interface is used to guide patients to describe their symptoms using the multi-round conversation capabilities of the big model, automatically parse and extract the patient's symptom description using the natural language processing technology of the big model, and extract the chief complaint information in combination with the medical knowledge graph;

[0030] The query integration module is used to find relevant information about the chief complaint in the medical knowledge graph, integrate and sort the relevant information found, and form a comprehensive and orderly patient condition information database;

[0031] The question-and-answer simulation module is used to simulate the professional question-and-answer method of doctors based on the patient's condition information database, guide patients to further describe their symptoms and changes in the course of their disease in detail, and update the acquired information to the patient's condition information database. It is also used to simulate the professional question-and-answer method of doctors to inquire about the patient's current physical signs;

[0032] The diagnostic reasoning module is used to use the powerful learning and reasoning capabilities of the big model to comprehensively analyze and judge the patient's chief complaint information, physical sign information, and medical history information. Based on medical knowledge and diagnostic rules, it simulates the doctor's diagnostic thinking process and gives preliminary diagnostic results and suggestions.

[0033] The result display module is used to intuitively display the diagnosis results and suggestions to patients or doctors.

[0034] Optionally, the graph construction modules involved specifically include:

[0035] Data collection unit, used to collect various medical data resources, including clinical guidelines, medical literature, drug instructions and disease databases;

[0036] A screening and sorting unit is used to screen and sort the collected medical data resources, remove duplicate, erroneous or irrelevant information, and form medical text data;

[0037] A preprocessing unit is used to perform preprocessing operations such as word segmentation, part-of-speech tagging, named entity recognition and relationship extraction on medical text data using natural language processing technology NLP, so as to identify medical entities in the medical text and the relationship between medical entities, thereby converting unstructured medical text data into structured data;

[0038] The graph construction unit is used to construct a medical knowledge graph with medical entities as nodes and relationships between medical entities as edges based on the identified medical entities and the relationships between medical entities.

[0039] Optionally, the dialogue interface involved utilizes the multi-round dialogue capability of the big model to guide the patient to describe his or her symptoms, utilizes the natural language processing technology of the big model to automatically parse and extract the patient's symptom description, and extracts the chief complaint information in combination with the medical knowledge graph. In this process: the dialogue interface guides the patient to answer questions and describe his or her own symptoms through text or voice input based on a preset symptom question template; in the process of guiding the patient to answer questions, utilizes the multi-round dialogue capability of the big model to dynamically adjust subsequent questions based on the patient's current answers, utilizes the natural language processing technology of the big model to automatically parse and extract the patient's answers, and extracts the key symptom description information in combination with the medical knowledge graph to extract the chief complaint information.

[0040] Optionally, when the pre-training module is used to pre-train a large model, the medical dataset used includes general natural language text, professional medical literature and clinical data, and real conversation logs.

[0041] Optionally, the result display module involved intuitively displays the diagnosis results and suggestions to the patient or doctor through charts or text descriptions;

[0042] The diagnosis results and recommendations include the name of the disease, the corresponding probability, and related explanations and recommendations, as well as personalized treatment recommendations and guidance based on the patient's specific circumstances.

[0043] The intelligent diagnosis method and system based on knowledge graph and large model of the present invention have the following beneficial effects compared with the prior art:

[0044] 1. The present invention realizes the systematization and structuring of medical knowledge by constructing a medical knowledge graph, providing a solid knowledge foundation for intelligent diagnosis. At the same time, it introduces large model technology and uses its powerful learning and reasoning capabilities to conduct in-depth mining and analysis of complex medical data, improve the accuracy and efficiency of diagnosis, and solve the problems of traditional medical diagnosis that rely on doctor's experience, strong subjectivity, and low efficiency;

[0045] 2. The present invention not only improves the accuracy of diagnosis, but also realizes the optimal allocation of medical resources, provides patients with more convenient and accurate medical services, reduces the workload of doctors, promotes the optimal allocation of medical resources, and promotes the rapid development of the field of intelligent medical care. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] Attached Figure 1 is a flow chart of a method according to Embodiment 1 of the present invention;

[0047] Attached Figure 2 This is a module connection block diagram of the second embodiment of the present invention. DETAILED DESCRIPTION

[0048] In order to make the technical solution, the technical problem solved and the technical effect of the present invention more clearly understood, the technical solution of the present invention is clearly and completely described below in conjunction with specific embodiments.

[0049] Embodiment 1:

[0050] Combined with Figure 1 , this embodiment proposes an intelligent diagnosis method based on knowledge graph and large model, which includes the following steps:

[0051] S1. Construct a medical knowledge graph containing medical entities and the relationships between medical entities, wherein the medical entities include diseases, symptoms, signs and treatment methods.

[0052] The process of building a medical knowledge graph is as follows:

[0053] Collect various medical data resources, including clinical guidelines, medical literature, drug instructions and disease databases;

[0054] Screen and organize the collected medical data resources, remove duplicate, erroneous or irrelevant information, and form medical text data;

[0055] Use natural language processing technology (NLP) to perform preprocessing operations on medical text data such as word segmentation, part-of-speech tagging, named entity recognition, and relationship extraction to identify medical entities (such as diseases, symptoms, drugs, etc.) and relationships between medical entities (such as causal relationships, parallel relationships, etc.) in medical texts, thereby converting unstructured medical text data into structured data;

[0056] Based on the identified medical entities and the relationships between them, a medical knowledge graph is constructed with medical entities as nodes and the relationships between them as edges.

[0057] At the same time, a medical dataset including general natural language text, professional medical literature and clinical data, and real conversation logs is used to pre-train a large model with powerful natural language processing capabilities that can realize professional question-answering and diagnosis functions in the medical field.

[0058] S2. Design a dialogue interface that guides patients to answer questions and describe their symptoms through text or voice input based on preset symptom question templates; in the process of guiding patients to answer questions, use the multi-round dialogue capability of the big model to dynamically adjust subsequent questions based on the patient's current answer to gain a more comprehensive understanding of the patient's symptom manifestations, disease course changes and other information; after multiple rounds of dialogue, use the natural language processing technology of the big model to automatically parse and extract the patient's answers, and extract the chief complaint information by extracting key symptom description information and combining it with the medical knowledge graph.

[0059] S3. Use the association query capability of the medical knowledge graph to find relevant information about the chief complaint in the medical knowledge graph, such as all symptoms and signs closely related to the chief complaint, as well as information on auxiliary tests that may be needed;

[0060] In the search results, detailed symptom descriptions, physical signs, auxiliary examination items and other information related to the chief complaint are further extracted, integrated and sorted to form a comprehensive and orderly patient condition information database, providing rich data support for subsequent intelligent diagnosis.

[0061] S4. Based on the patient's medical information database, a large model is used to simulate the doctor's professional question-and-answer method to guide the patient to further describe their symptoms and changes in the course of the disease in detail, and the acquired information is updated to the patient's medical information database.

[0062] Then, the large model was used to simulate the professional question-and-answer method of doctors to ask the patient's current vital signs, such as the patient's current temperature, blood pressure, and heart rate. Vital sign information is important for assessing the patient's physiological state and the severity of the disease.

[0063] S5. The dialogue interface asks the patient about his or her medical history information through a preset medical history question template. During the inquiry process, the multi-round dialogue capability of the large model is utilized to dynamically adjust subsequent questions based on the patient's current answers.

[0064] Medical history information plays a vital role in medical diagnosis because it often reveals a patient's potential health risks, disease susceptibility, or existing health problems. For example, family genetic history can help doctors determine whether a patient is at risk for certain genetic diseases; past medical history can reveal whether a patient has had similar diseases or received related treatments in the past, thus providing important clues for current diagnosis.

[0065] S6. The large model uses its powerful learning and reasoning capabilities to conduct a comprehensive analysis and judgment of the patient's chief complaint information, physical sign information and medical history information, and based on medical knowledge and diagnostic rules, simulates the doctor's diagnostic thinking process, and gives preliminary diagnostic results and suggestions. The diagnostic results and suggestions include the name of the disease, the corresponding possibility, and related explanations and suggestions. It also includes personalized treatment suggestions and guidance based on the patient's specific situation, such as drug treatment plans, lifestyle adjustments, dietary recommendations, etc.

[0066] S7. The large model intuitively displays the diagnosis results and suggestions to patients or doctors through charts or text descriptions.

[0067] Doctors can make comprehensive judgments based on the diagnosis results of the big model and the actual situation of the patient, and formulate more accurate and personalized treatment plans. At the same time, the big model can also provide professional advice on disease treatment, prevention and rehabilitation, helping patients better manage their health.

[0068] Embodiment 2:

[0069] Combined with Figure 2 , this embodiment proposes an intelligent diagnosis system based on knowledge graph and large model, which includes:

[0070] A graph construction module, used to construct a medical knowledge graph containing medical entities and relationships between medical entities, wherein the medical entities include diseases, symptoms, signs and treatment methods;

[0071] The pre-training module is used to pre-train a large model with powerful natural language processing capabilities and capable of realizing professional question-answering and diagnosis functions in the medical field using a medical dataset including general natural language texts, professional medical literature and clinical data, and real conversation logs. The large model further includes:

[0072] The conversation interface is used to guide patients to describe their symptoms using the multi-round conversation capabilities of the big model, automatically parse and extract the patient's symptom description using the natural language processing technology of the big model, and extract the chief complaint information in combination with the medical knowledge graph;

[0073] The query integration module is used to find relevant information about the chief complaint in the medical knowledge graph, integrate and sort the relevant information found, and form a comprehensive and orderly patient condition information database;

[0074] The question-and-answer simulation module is used to simulate the professional question-and-answer method of doctors based on the patient's condition information database, guide patients to further describe their symptoms and changes in the course of their disease in detail, and update the acquired information to the patient's condition information database. It is also used to simulate the professional question-and-answer method of doctors to inquire about the patient's current physical signs;

[0075] The diagnostic reasoning module is used to use the powerful learning and reasoning capabilities of the large model to conduct a comprehensive analysis and judgment of the patient's chief complaint information, physical sign information, and medical history information. Based on medical knowledge and diagnostic rules, it simulates the doctor's diagnostic thinking process and gives preliminary diagnostic results and suggestions. The diagnostic results and suggestions include the name of the disease, the corresponding possibility, and related explanations and suggestions. It also includes personalized treatment suggestions and guidance based on the patient's specific situation, such as drug treatment plans, lifestyle adjustments, dietary suggestions, etc.;

[0076] The result display module is used to intuitively display the diagnosis results and suggestions to patients or doctors through charts or text descriptions.

[0077] In this embodiment, the graph construction modules involved specifically include:

[0078] Data collection unit, used to collect various medical data resources, including clinical guidelines, medical literature, drug instructions and disease databases;

[0079] A screening and sorting unit is used to screen and sort the collected medical data resources, remove duplicate, erroneous or irrelevant information, and form medical text data;

[0080] A preprocessing unit is used to perform preprocessing operations such as word segmentation, part-of-speech tagging, named entity recognition and relationship extraction on medical text data using natural language processing technology NLP, so as to identify medical entities in the medical text and the relationship between medical entities, thereby converting unstructured medical text data into structured data;

[0081] The graph construction unit is used to construct a medical knowledge graph with medical entities as nodes and relationships between medical entities as edges based on the identified medical entities and the relationships between medical entities.

[0082] In this embodiment, the dialogue interface involved uses the multi-round dialogue capability of the big model to guide the patient to describe his symptoms, uses the natural language processing technology of the big model to automatically parse and extract the patient's symptom description, and extracts the chief complaint information in combination with the medical knowledge graph. In this process: the dialogue interface guides the patient to answer questions and describe his own symptoms through text or voice input according to the preset symptom question template; in the process of guiding the patient to answer questions, the multi-round dialogue capability of the big model is used to dynamically adjust subsequent questions based on the patient's current answer, and the natural language processing technology of the big model is used to automatically parse and extract the patient's answer, and the chief complaint information is extracted by extracting key symptom description information and combining it with the medical knowledge graph.

[0083] In summary, the intelligent diagnosis method and system based on knowledge graph and large model of the present invention can improve the accuracy and efficiency of diagnosis, and solve the problems of traditional medical diagnosis relying on doctor's experience, strong subjectivity and low efficiency.

[0084] The above specific examples are used to explain the principles and implementation methods of the present invention in detail. These examples are only used to help understand the core technical content of the present invention. Based on the above specific embodiments of the present invention, any improvements and modifications made by technicians in this technical field without departing from the principles of the present invention should fall within the scope of patent protection of the present invention.

Claims

1. An intelligent diagnosis method based on knowledge graph and large model, characterized in that: The steps include: S1. Build a medical knowledge graph containing medical entities and the relationships between them, where the medical entities include diseases, symptoms, signs, and treatments; pre-train a large model with powerful natural language processing capabilities that can realize professional question-answering and diagnosis functions in the medical field; S2. Design a dialogue interface that uses the multi-round dialogue capability of the big model to guide patients to describe their symptoms, uses the natural language processing technology of the big model to automatically parse and extract the patient's symptom description, and extracts the chief complaint information in combination with the medical knowledge graph; S3. Use the association query capability of the medical knowledge graph to find relevant information about the chief complaint in the medical knowledge graph; Integrate and sort the relevant information found to form a comprehensive and orderly patient condition information database; S4. Based on the patient's condition information database, the large model is used to simulate the doctor's professional question-and-answer method to guide the patient to further describe his or her symptoms and changes in the course of the disease in detail, and the acquired information is updated to the patient's condition information database; then, the large model is used to simulate the doctor's professional question-and-answer method to inquire about the patient's current physical signs; S5. The dialogue interface asks the patient's medical history information through a preset medical history question template; During the inquiry process, the multi-round dialogue capability of the large model is used to dynamically adjust subsequent questions based on the patient's current answer; S6. The large model uses its powerful learning and reasoning capabilities to comprehensively analyze and judge the patient's chief complaint information, physical sign information, and medical history information, and based on medical knowledge and diagnostic rules, simulates the doctor's diagnostic thinking process and gives preliminary diagnostic results and suggestions; S7. The large model intuitively displays its diagnostic results and suggestions to patients or doctors.

2. According to claim 1, the intelligent diagnosis method based on knowledge graph and large model is characterized in that: Execute step S1 to build a medical knowledge graph as follows: Collect various medical data resources, including clinical guidelines, medical literature, drug instructions and disease databases; Screen and organize the collected medical data resources, remove duplicate, erroneous or irrelevant information, and form medical text data; Use natural language processing technology (NLP) to perform pre-processing operations on medical text data, such as word segmentation, part-of-speech tagging, named entity recognition, and relationship extraction, to identify medical entities in medical texts and the relationships between medical entities, thereby converting unstructured medical text data into structured data; Based on the identified medical entities and the relationships between them, a medical knowledge graph is constructed with medical entities as nodes and the relationships between them as edges.

3. According to claim 1, the intelligent diagnosis method based on knowledge graph and large model is characterized in that: The dialogue interface designed in step S2 guides the patient to answer questions and describe his or her own symptoms through text or voice input according to a preset symptom question template; in the process of guiding the patient to answer the questions, the multi-round dialogue capability of the large model is utilized to dynamically adjust the subsequent questions based on the patient's current answer, and the natural language processing technology of the large model is used to automatically parse and extract the patient's answer, and the main complaint information is extracted by extracting key symptom description information and combining it with the medical knowledge graph.

4. According to claim 1, the intelligent diagnosis method based on knowledge graph and large model is characterized in that: When pre-training large models, the medical datasets used include general natural language text, professional medical literature and clinical data, and real conversation logs.

5. According to claim 1, the intelligent diagnosis method based on knowledge graph and large model is characterized in that: The large model can visually display the diagnosis results and suggestions to patients or doctors through diagrams or text descriptions; The diagnostic results and recommendations of the big model include the name of the disease, the corresponding probability, and related explanations and suggestions, as well as personalized treatment recommendations and guidance based on the patient's specific circumstances.

6. An intelligent diagnosis system based on knowledge graph and large model, characterized in that: It includes: A graph construction module, used to construct a medical knowledge graph containing medical entities and relationships between medical entities, wherein the medical entities include diseases, symptoms, signs and treatment methods; The pre-training module is used to pre-train a large model with powerful natural language processing capabilities and capable of realizing professional question-answering and diagnosis functions in the medical field. The large model further includes: The conversation interface is used to guide patients to describe their symptoms using the multi-round conversation capabilities of the big model, automatically parse and extract the patient's symptom description using the natural language processing technology of the big model, and extract the chief complaint information in combination with the medical knowledge graph; The query integration module is used to find relevant information about the chief complaint in the medical knowledge graph, integrate and sort the relevant information found, and form a comprehensive and orderly patient condition information database; The question-and-answer simulation module is used to simulate the professional question-and-answer method of doctors based on the patient's condition information database, guide patients to further describe their symptoms and changes in the course of their disease in detail, and update the acquired information to the patient's condition information database. It is also used to simulate the professional question-and-answer method of doctors to inquire about the patient's current physical signs; The diagnostic reasoning module is used to use the powerful learning and reasoning capabilities of the big model to comprehensively analyze and judge the patient's chief complaint information, physical sign information, and medical history information. Based on medical knowledge and diagnostic rules, it simulates the doctor's diagnostic thinking process and gives preliminary diagnostic results and suggestions. The result display module is used to intuitively display the diagnosis results and suggestions to patients or doctors.

7. The intelligent diagnosis system based on knowledge graph and large model according to claim 6 is characterized in that: The graph construction module specifically includes: Data collection unit, used to collect various medical data resources, including clinical guidelines, medical literature, drug instructions and disease databases; A screening and sorting unit is used to screen and sort the collected medical data resources, remove duplicate, erroneous or irrelevant information, and form medical text data; A preprocessing unit is used to perform preprocessing operations such as word segmentation, part-of-speech tagging, named entity recognition and relationship extraction on medical text data using natural language processing technology NLP, so as to identify medical entities in the medical text and the relationship between medical entities, thereby converting unstructured medical text data into structured data; The graph construction unit is used to construct a medical knowledge graph with medical entities as nodes and relationships between medical entities as edges based on the identified medical entities and the relationships between medical entities.

8. The intelligent diagnosis system based on knowledge graph and large model according to claim 6 is characterized in that: The dialogue interface uses the multi-round dialogue capability of the big model to guide patients to describe their symptoms, uses the natural language processing technology of the big model to automatically parse and extract the patient's symptom description, and extracts the chief complaint information in combination with the medical knowledge graph. In this process: the dialogue interface guides patients to answer questions and describe their own symptoms through text or voice input according to the preset symptom question template; In the process of guiding patients to answer questions, the multi-round dialogue capability of the big model is utilized to dynamically adjust follow-up questions based on the patient's current answers. The natural language processing technology of the big model is used to automatically parse and extract the patient's answers, and the main complaint information is extracted by extracting key symptom description information and combining it with the medical knowledge graph.

9. The intelligent diagnosis system based on knowledge graph and large model according to claim 6 is characterized in that: When the pre-training module pre-trains the large model, the medical data set used includes general natural language text, professional medical literature and clinical data, and real conversation logs.

10. The intelligent diagnosis system based on knowledge graph and large model according to claim 6, characterized in that: The result display module intuitively displays the diagnosis results and suggestions to the patient or doctor through charts or text descriptions; The diagnosis results and recommendations include the name of the disease, the corresponding probability, and related explanations and recommendations, as well as personalized treatment recommendations and guidance based on the patient's specific circumstances.

Images