Diagnosis division and guidance method and system for large medical model, electronic equipment and storage medium

Abstract

The invention provides a separate hospital guidance method and system for a large medical model, electronic equipment and a storage medium. The method comprises the following steps: receiving symptom description information of a natural language of a patient; performing natural language processing on the symptom description information by utilizing a large medical model, and determining symptom entity information and basic information of the patient; inputting the symptom entity information into a pre-constructed medical knowledge graph for matching, and determining candidate diseases and departments thereof corresponding to the patient; generating personalized triage suggestions by using the medical big model according to the candidate diseases and the departments thereof corresponding to the patient and in combination with the basic information of the patient; that is to say, according to the medical big model and the medical knowledge graph, symptom entity identification is performed on the symptom description of the patient, and then the corresponding candidate disease and the department thereof are determined, so that a manual separate hospital guide mode is avoided, the separate hospital guide efficiency and accuracy are improved, the waste of medical resources is reduced, and the fairness and sustainability of medical services are realized.

Description

Technical Field

[0001] The present invention belongs to the field of data processing technology, and more specifically, particularly relates to a diagnosis and guidance method and system, electronic equipment, and storage medium for a large medical model. Background Art

[0002] The medical industry is in a stage of rapid development, and patients' demand for medical treatment is growing with the improvement of living standards and the enhancement of health awareness. However, problems such as the imbalance of medical resource allocation and information asymmetry between doctors and patients have gradually surfaced and become key factors affecting the quality and efficiency of medical services. Especially in a busy medical environment, patients often need to spend a lot of time waiting for medical treatment, and doctors are also under work pressure.

[0003] In the current medical system, the traditional triage method mainly relies on the personal experience and subjective judgment of medical staff. This manual judgment method is not only inefficient, but also prone to errors due to human factors, making it difficult to meet patients' expectations for efficient and accurate medical services. In addition, manual triage may also lead to waste and uneven distribution of medical resources, further exacerbating the tension in medical services.

[0004] Faced with the challenges brought by the rapid development of the medical industry, the traditional triage method has been unable to adapt to the needs of modern medical services. How to improve the efficiency and accuracy of triage, reduce the waste of medical resources, and achieve fairness and sustainability of medical services are technical issues that need to be solved urgently. Summary of the invention

[0005] In view of this, the purpose of the present invention is to provide a medical big model triage method and system, electronic equipment, and storage medium, which are used to perform symptom entity recognition on the patient's symptom description based on the medical big model and the medical knowledge graph, and then determine the corresponding candidate diseases and their departments, so as to reduce the waste of medical resources and achieve fairness and sustainability of medical services.

[0006] The first aspect of the present application discloses a method for diagnosis and guidance of a large medical model, comprising:

[0007] Receive symptom description information in natural language from the patient;

[0008] Using the medical big model to perform natural language processing on the symptom description information to determine the symptom entity information and basic information of the patient;

[0009] Input the symptom entity information into the pre-built medical knowledge graph for matching, and determine the candidate disease and its department corresponding to the patient;

[0010] The medical big model is used to generate personalized triage suggestions based on the candidate diseases and departments corresponding to the patient and combined with the basic information of the patient.

[0011] Optionally, the symptom description information is text information and / or voice information.

[0012] Optionally, before receiving the patient's natural language symptom description information, the method further includes:

[0013] The medical big model is constructed by using deep learning technology and learning samples, so that the medical big model can understand and analyze the patient's symptom description information and obtain the patient's disease entity information and basic information.

[0014] Optionally, before receiving the patient's natural language symptom description information, the method further includes:

[0015] Construct a medical knowledge graph and dynamically update the medical knowledge graph; the medical knowledge graph includes: at least two medical entities among diseases, symptoms, examinations, drugs, and treatment plans, and the relationship between at least two medical entities.

[0016] Optionally, after the medical big model generates personalized triage suggestions based on the candidate diseases and departments corresponding to the patient and in combination with the basic information of the patient, it further includes:

[0017] Collect patients' satisfaction with triage recommendations and actual treatment outcomes;

[0018] The medical big model and the medical knowledge graph are optimized according to the satisfaction and actual medical treatment results.

[0019] Optionally, after receiving the patient's natural language symptom description information, the method further includes:

[0020] The patient's natural language symptom description information is encrypted using an encryption algorithm.

[0021] The second aspect of the present application discloses a large medical model diagnosis and guidance system, including:

[0022] A receiving module, used for receiving symptom description information in natural language of the patient;

[0023] A language processing module, used to perform natural language processing on the symptom description information using a medical big model to determine the symptom entity information and basic information of the patient;

[0024] A matching module is used to input the symptom entity information into a pre-built medical knowledge graph for matching, and determine the candidate disease and its department corresponding to the patient;

[0025] The triage module is used to generate personalized triage suggestions by using the medical big model according to the candidate diseases and departments corresponding to the patient and in combination with the basic information of the patient.

[0026] Optionally, also include:

[0027] The feedback update module is used to collect patients' satisfaction with triage recommendations and actual medical results; and optimize the medical big model and the medical knowledge graph based on the satisfaction and actual medical results.

[0028] The third aspect of the present application discloses an electronic device, comprising:

[0029] one or more processors;

[0030] a storage device having one or more programs stored thereon;

[0031] When one or more programs are executed by one or more processors, the one or more processors implement the method for diagnosis and treatment of a large medical model as described in any one of the first aspects of the present application.

[0032] A fourth aspect of the present application discloses a storage medium having a computer program stored thereon, wherein when the computer program is executed by a processor, the method for triage and guidance of a large medical model as described in any one of the first aspects of the present application is implemented.

[0033] It can be seen from the above technical scheme that the present invention provides a medical big model triage method, including: receiving the patient's natural language symptom description information; using the medical big model to perform natural language processing on the symptom description information to determine the patient's symptom entity information and basic information; inputting the symptom entity information into a pre-constructed medical knowledge graph for matching, and determining the candidate diseases and departments corresponding to the patient; using the medical big model to generate personalized triage suggestions based on the candidate diseases and departments corresponding to the patient, and in combination with the patient's basic information; that is, based on the medical big model and the medical knowledge graph, symptom entity recognition is performed on the patient's symptom description, and then the corresponding candidate diseases and departments are determined, avoiding the use of manual triage methods, improving the efficiency and accuracy of triage, reducing the waste of medical resources, and achieving fairness and sustainability of medical services. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0035] Figure 1 It is a flow chart of a method for diagnosis and treatment of a large medical model provided by an embodiment of the present invention;

[0036] Figure 2 It is a schematic diagram of a large medical model diagnosis and guidance system provided by an embodiment of the present invention;

[0037] Figure 3 It is a schematic diagram of an electronic device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0038] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0039] In the present application, the term "comprise", "include" or any other variant thereof is intended to cover non-exclusive inclusion, so that the process, method, article or equipment including a series of elements includes not only those elements, but also includes other elements not explicitly listed, or also includes elements inherent to such process, method, article or equipment. In the absence of more restrictions, the elements defined by the sentence "comprise one ..." do not exclude the existence of other identical elements in the process, method, article or equipment including the elements. In addition, the terms "first", "second", "third", "fourth", etc. (if present) in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects, and do not have to be used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable in appropriate circumstances, so that the embodiments of the present application described here can be implemented in an order other than those illustrated or described here, for example.

[0040] The embodiment of the present application provides a method for triage and diagnosis of a large medical model, which is used to solve the problem that the traditional triage and diagnosis method in the prior art has been difficult to adapt to the needs of modern medical services; how to improve the efficiency and accuracy of triage and diagnosis, reduce the waste of medical resources, and achieve the fairness and sustainability of medical services is a technical problem that needs to be solved urgently.

[0041] See also Figure 1 The diagnosis and treatment method of the medical model includes:

[0042] S101. Receive symptom description information in natural language from a patient.

[0043] Symptom description information based on natural language of the user may be obtained through the user interface; of course, a guidance page may be provided in the user interface, for example, the user may accurately describe his or her own symptoms based on the guidance page.

[0044] It should be noted that the symptom description information is text information and / or voice information. That is, the user can describe his or her symptoms in text or in voice.

[0045] Specifically, speech recognition and text processing technologies are used to capture and parse complex symptom descriptions expressed by patients in natural language, and whether they are oral reports or written inputs, they can get instant and detailed responses. These symptom descriptions may include the patient's various discomforts, medical history, a wide range of information from mild discomfort to severe pain, and any relevant health details.

[0046] Symptom description information may include: multiple rounds of interactive dialogues between the patient and the big model, and of course it can also be the first round of interactive dialogue, that is, at this stage, the medical condition information is input into the big medical model.

[0047] S102. Use the medical big model to perform natural language processing on the symptom description information to determine the symptom entity information and basic information of the patient.

[0048] The system's built-in advanced medical big model uses deep learning algorithms to perform natural language analysis on these descriptions. This process not only accurately extracts symptom entities, such as pain location, intensity, duration, etc., but also intelligently identifies and summarizes the patient's basic health information, such as age, gender characteristics, living habits, work environment, and medical history, to ensure the comprehensiveness and accuracy of the information.

[0049] Specifically, the symptom description information is input into the medical big model for processing, and the medical big model outputs the symptom entity information and basic information.

[0050] The medical model first uses a complex neural network structure to perform multi-level semantic analysis on symptom description information. In this process, the model uses word embedding technology to convert each word in the text into a high-dimensional vector to capture the semantic relationship between words. Then, through structures such as convolutional neural networks (CNN) or recurrent neural networks (RNN), the model can identify and extract key information in the text, such as pain location, intensity, duration, etc. These identified key information are called symptom entities.

[0051] At the same time, the medical big model also has the ability of named entity recognition (NER), which can intelligently summarize the patient's age, gender, past medical history and other basic health data. In this process, the model will use technologies such as conditional random fields (CRF) or attention mechanisms to accurately label entities in the text, thereby extracting the patient's basic information.

[0052] Optionally, before receiving the patient's natural language symptom description information, the method further includes:

[0053] The medical big model is constructed by using deep learning technology and learning samples, so that the medical big model can understand and analyze the patient's symptom description information and obtain the patient's disease entity information and basic information.

[0054] Using deep learning technology and rich learning samples, a large medical model was carefully constructed. The core goal of this model is to deeply understand and analyze the patient's symptom description information, so as to accurately extract the patient's disease entity information and basic information.

[0055] During the construction process, advanced deep learning algorithms were selected. These algorithms can automatically learn complex features and patterns in data, thereby achieving efficient processing and analysis of large-scale medical data. Through deep learning, the model can gradually master the knowledge and laws in the medical field, providing strong support for subsequent disease analysis.

[0056] In order to ensure the accuracy and reliability of the model, a large number of learning samples were carefully collected. These samples cover a wide range of medical fields, including symptom descriptions, diagnosis results, treatment plans, etc. of various diseases. By learning from these samples, the model can gradually accumulate rich medical knowledge and experience, providing a solid foundation for parsing patients' symptom descriptions.

[0057] After determining the deep learning algorithm and learning samples, we began to build a large medical model. During the model construction process, we fully considered the particularity and complexity of the medical field and adopted a multi-level neural network structure and complex algorithm logic. Through continuous training and optimization, the model gradually acquired strong semantic understanding and parsing capabilities.

[0058] When the medical big model receives the patient's symptom description information, it can quickly analyze and process the information. Using the powerful capabilities of the deep learning algorithm, the model can accurately identify the disease entity information in the symptom description, such as the pain location, pain nature, and accompanying symptoms. At the same time, the model can also extract the patient's basic information, such as age, gender, and past medical history. This information plays a vital role in the subsequent diagnosis of the disease and the formulation of treatment plans.

[0059] S103. Input the symptom entity information into a pre-constructed medical knowledge graph for matching, and determine the candidate diseases and their departments corresponding to the patient.

[0060] The medical knowledge graph is built on millions of medical documents and massive clinical data. Through graph database technology and efficient graph matching algorithms, it can quickly identify potential diseases that best match the patient's symptoms and the departments they should be treated in. In addition, the medical knowledge graph also has the ability to be dynamically updated, and can absorb the latest medical research results in real time to ensure the timeliness and accuracy of knowledge.

[0061] That is, before step S101, the step of receiving the patient's natural language symptom description information, further includes:

[0062] Construct a medical knowledge graph and dynamically update the medical knowledge graph; the medical knowledge graph includes: at least two medical entities among diseases, symptoms, examinations, drugs, and treatment plans, and the relationship between at least two medical entities.

[0063] The construction of a medical knowledge graph is a complex and sophisticated process that relies on a large amount of medical data, literature, and clinical cases. First, key information needs to be extracted from a large amount of medical resources, including but not limited to medical entities such as diseases, symptoms, examinations, drugs, and treatment plans. Then, natural language processing technology and machine learning algorithms are used to identify and define the relationships between these entities, such as the correspondence between symptoms and diseases, the association between diseases and examination methods, and the matching of drugs and treatment plans. These relationships constitute the core of the knowledge graph, enabling the graph to reflect the inherent logic and connections of medical knowledge.

[0064] During the construction process, the graph also needs to be quality controlled and verified to ensure its accuracy and reliability. This includes verifying the accuracy of the extracted information, evaluating the rationality of the relationship definition, and optimizing the overall structure of the graph.

[0065] Dynamic update of medical knowledge graph: Since knowledge in the medical field is updated rapidly, new research results, treatment methods and technologies are constantly emerging, so the medical knowledge graph needs to have the ability to be dynamically updated. This includes the introduction of new medical entities, the modification and expansion of existing entity relationships, and the adjustment and optimization of the overall structure of the graph.

[0066] In order to achieve dynamic updates, the system needs to regularly collect and analyze the latest medical literature, clinical cases, and research results to extract new medical entities and relationships. At the same time, it is also necessary to regularly review and evaluate existing maps to discover and correct possible errors or omissions. In addition, the system can also use user feedback and clinical data to continuously optimize the map and improve its practicality and accuracy.

[0067] Medical entities and their relationships in the medical knowledge graph: In the medical knowledge graph, there are at least the following two medical entities and their relationships:

[0068] Disease and Symptom: Disease entities represent specific medical conditions or pathological processes, while symptom entities represent abnormal signs or feelings exhibited by patients. The relationship between disease and symptoms is the core part of the atlas, which reflects the symptoms caused by the disease and the role of symptoms in indicating the disease.

[0069] Disease and examination / treatment: The disease entity is also associated with entities such as examination method and treatment plan. These relationships help doctors determine the examination items and treatment methods required for a specific disease, thereby developing a personalized diagnosis and treatment plan.

[0070] In addition, the medical knowledge graph may also contain other types of medical entities and their relationships, such as the correspondence between drugs and diseases, the interaction between drugs, etc. The introduction of these entities and relationships further enriches the content of the graph and improves its value in practical applications.

[0071] In summary, the construction and dynamic update of the medical knowledge graph is one of the key technologies for the system to achieve accurate triage and personalized diagnosis and treatment recommendations. By continuously optimizing the content and structure of the graph, the system can more accurately understand the patient's symptoms and needs and provide more effective medical services and support.

[0072] The construction process of the medical big model and knowledge graph is actually a systematic organization and inheritance of medical knowledge. With the continuous addition of new cases and new research results, the system can continuously update the medical knowledge base, maintain its timeliness and accuracy, and provide solid knowledge support for medical services.

[0073] S104. Generate personalized triage suggestions using the medical big model based on the candidate diseases and departments corresponding to the patient and in combination with the basic information of the patient.

[0074] First, the system has identified the candidate diseases and corresponding departments that best match the patient's symptoms through the medical knowledge graph. This step is based on the matching results of the symptom entity information and basic information in the graph.

[0075] Secondly, after obtaining candidate diseases and departments, the system further considers the patient's basic information, such as age, gender, past medical history, allergy history, lifestyle habits, etc. This information is crucial for making personalized triage recommendations because it can affect the incidence of the disease, the severity of symptoms, and the choice of treatment options.

[0076] Next, the system uses a deep learning medical big model to conduct a comprehensive analysis of candidate diseases, departments, and basic patient information. During this process, the model will consider multiple factors, such as the severity of the disease, the professionalism of the department, the patient's health status and preferences, to generate the most suitable triage recommendations for the patient.

[0077] The triage recommendations may include the recommended department to visit, recommended examination items, possible disease diagnosis, an overview of treatment options, and subsequent follow-up recommendations, etc. These recommendations are designed to provide patients with clear and specific guidance on medical treatment and help them find the right treatment path more quickly.

[0078] It is worth noting that the triage recommendations generated by this system are highly personalized. It not only takes into account the patient's specific symptoms and basic information, but also utilizes the deep learning capabilities of the medical big model and the rich knowledge resources of the medical knowledge graph to achieve accurate understanding and personalized analysis of the patient's condition.

[0079] In addition, the intelligence of the system is also reflected in its ability to continuously optimize based on patient feedback and clinical data. Through continuous learning and updating, the system can more accurately understand the needs of patients and provide more effective triage recommendations, thereby improving the overall quality and efficiency of medical services.

[0080] In other words, using the medical big model to generate personalized triage recommendations based on the patient's corresponding candidate diseases and departments, combined with the patient's basic information, is one of the important technologies for this system to achieve precision medicine and personalized services. By continuously optimizing and improving this process, the system can provide patients with more efficient and accurate medical services and support.

[0081] Personalized triage recommendations, combined with the patient's personal characteristics (such as age, gender, medical history, etc.), this method can generate personalized triage recommendations to meet the differentiated needs of different patients and improve the pertinence and satisfaction of medical services.

[0082] By combining deep learning with knowledge graphs, this method can accurately analyze patient symptom descriptions, quickly match the most likely diseases and corresponding departments, effectively reduce misdiagnosis and missed diagnosis rates, shorten patient waiting time, and improve overall triage efficiency. This application can reasonably guide the flow of patients, avoid the waste of medical resources caused by blind registration of patients, enable more reasonable and efficient allocation of medical resources, and improve the quality of medical services.

[0083] In this embodiment, the patient's natural language symptom description information is received; the symptom description information is processed in natural language using the medical big model to determine the patient's symptom entity information and basic information; the symptom entity information is input into a pre-constructed medical knowledge graph for matching to determine the candidate diseases and departments corresponding to the patient; the medical big model is used to generate personalized triage suggestions based on the candidate diseases and departments corresponding to the patient and the basic information of the patient; that is, based on the medical big model and the medical knowledge graph, symptom entity recognition is performed on the patient's symptom description, and then the corresponding candidate diseases and departments are determined, thereby avoiding the use of manual triage methods, improving the efficiency and accuracy of triage, reducing the waste of medical resources, and achieving fairness and sustainability of medical services.

[0084] Optionally, after the medical big model generates personalized triage suggestions based on the candidate diseases and departments corresponding to the patient and in combination with the basic information of the patient, it further includes:

[0085] Collect patients' satisfaction with triage recommendations and actual treatment results; optimize the medical big model and the medical knowledge graph based on the satisfaction and actual treatment results.

[0086] In other words, this application also sets up an evaluation and feedback mechanism for triage recommendations. By collecting patients' satisfaction with triage recommendations and actual medical results, it optimizes the performance of the medical big model and knowledge graph and improves the accuracy and reliability of triage recommendations.

[0087] Specifically, in the process of medical services, patient feedback is extremely valuable. In order to further improve the accuracy and practicality of the medical big model and medical knowledge graph, it is necessary to systematically collect patients' satisfaction with triage recommendations and their actual medical results.

[0088] A special satisfaction survey questionnaire can be designed, which should cover aspects such as patients' acceptance of triage recommendations, the clarity of the recommendations, and whether the recommendations help them understand their health conditions.

[0089] Questionnaires are distributed online (such as hospital apps, websites, etc.) and offline (such as paper questionnaires, face-to-face interviews, etc.) to ensure that extensive and comprehensive patient feedback is collected. The collected satisfaction data is regularly collated and analyzed to identify areas of patient satisfaction and dissatisfaction, providing a basis for subsequent optimization.

[0090] It can also track the patient's actual medical treatment process and record their diagnosis results, treatment plans and treatment effects.

[0091] Connect with the hospital's electronic medical record system to automatically obtain patients' medical records to ensure the accuracy and completeness of the data. Analyze the consistency between actual medical results and triage recommendations, and identify the strengths and weaknesses of the model in prediction and diagnosis.

[0092] After collecting enough satisfaction and actual medical treatment results data, we began to optimize the medical big model and medical knowledge graph.

[0093] According to patient satisfaction data, adjust the model's parameters and algorithms to improve the accuracy and acceptability of triage recommendations. Carry out targeted optimization for the dissatisfied parts of patient feedback, such as improving the clarity of recommendations and adding personalized recommendations. Use reinforcement learning and other technologies in machine learning algorithms to allow the model to continuously learn and improve in actual applications, thereby improving the model's adaptive capabilities.

[0094] Update and improve the medical knowledge and rules in the medical knowledge graph based on actual medical treatment results. Identify possible errors or omissions in the medical knowledge graph, make corrections and supplements, and ensure the accuracy and completeness of the graph. Strengthen the semantic relationships in the medical knowledge graph, improve the semantic understanding ability of the graph, and provide richer knowledge support for the model.

[0095] Establish a continuous feedback and optimization mechanism to regularly collect and analyze patient feedback and actual medical treatment results. Based on the results of data analysis, continuously optimize the medical big model and medical knowledge graph to improve the quality and efficiency of medical services. Maintain close cooperation with medical experts to ensure that the optimization direction of the model and knowledge graph is in line with the latest progress and actual needs in the medical field.

[0096] In other words, by collecting patients' satisfaction with triage recommendations and actual medical treatment results, and optimizing the medical big model and medical knowledge graph based on these data, the intelligence level of medical services and patient satisfaction can be continuously improved.

[0097] Optionally, after receiving the patient's natural language symptom description information, the method further includes:

[0098] The patient's natural language symptom description information is encrypted using an encryption algorithm.

[0099] In other words, this application can also provide protection measures for patient privacy data, ensuring that encryption technology, anonymization processing and other means are used during data processing and transmission to protect patient privacy and security.

[0100] Specifically, in addition to using basic encryption algorithms, you can consider using more advanced encryption technologies, such as quantum encryption or homomorphic encryption, to improve the strength and security of data encryption. Encryption algorithms should be updated and upgraded regularly to respond to new security threats and vulnerabilities that may arise.

[0101] When anonymizing patient data, ensure that the processed data cannot be directly associated with a specific patient, thereby protecting the patient's privacy and identity. Strengthen supervision and review of the anonymization process to ensure compliance and accuracy of the process.

[0102] Establish a strict data access control mechanism so that only authorized personnel or systems can access and process patient data. Monitor and audit data access records in real time to promptly detect and handle any unauthorized access.

[0103] During data transmission, secure transmission protocols and encryption technologies are used to ensure the security and integrity of data during transmission. The transmitted data is monitored and detected in real time to promptly detect and handle any potential security threats.

[0104] Develop detailed patient privacy protection policies and procedures to clarify the rights and responsibilities of patients in the process of data collection, processing, storage and transmission. Regularly review and update policies and procedures to adapt to the changing security environment and business needs.

[0105] In other words, by strengthening data encryption, improving anonymization mechanisms, establishing data access control, strengthening data transmission security, establishing patient privacy protection policies and processes, and strengthening employee training and education, the security and privacy of patient privacy data can be fully protected. The implementation of these measures will help improve the quality and efficiency of medical services while protecting the legitimate rights and interests of patients.

[0106] The following is an explanation of the entire process of how the medical big model interacts with patients to generate triage recommendations:

[0107] (1) Simulation of interactive dialogue between doctors and patients.

[0108] The following is a simulated doctor-patient dialogue to illustrate the triage process:

[0109] Patient: "I've been feeling tightness in my chest lately, and sometimes it hurts a little."

[0110] Large Model: "How long has this been going on?"

[0111] Patient: "It's been about two weeks. I didn't pay much attention to it at first, but I feel that the symptoms have worsened in the past few days."

[0112] The Big Model: “Is the pain constant or episodic? Are there specific times or situations when it is more likely to occur?”

[0113] Patient: “It’s mainly paroxysmal, especially when I get up in the morning and lie down at night.”

[0114] Large model: "Besides chest tightness and pain, are there any other discomforts? For example, coughing, shortness of breath or palpitations?"

[0115] Patient: "I have a little cough occasionally, but it's not serious. I don't pay much attention to shortness of breath and palpitations."

[0116] …

[0117] The above conversation may go on for multiple rounds until the large model collects enough information.

[0118] The reception and pre-treatment of the patient's symptom description can be designed during the conversation process. At this stage, the patient inputs the symptom description to the medical big model in the form of voice or text. The system first receives these inputs and uses them as the initial source of disease information. It can also involve: the conversion of natural language to semantic vectors. The medical big model layer uses advanced natural language processing technology to convert the symptoms described by the patient into semantic vectors for subsequent more in-depth analysis and understanding.

[0119] It should be noted that the conversation content is monitored in real time and passed to the medical big model, which makes a rough judgment on the completeness of the conversation information. When the conversation content basically covers all the information required by the big model (patient personal information, chief complaint information, etc.), which is sufficient to support the big model's reasoning, the system will stop asking questions and start executing subsequent actions, and finally give disease analysis and department recommendations.

[0120] (2) Text cleaning.

[0121] After several rounds of conversations, the system cleans the collected text, removes irrelevant symbols, punctuation marks, etc., and performs word segmentation to provide clean text data for subsequent steps.

[0122] (3) Preliminary identification and removal of symptom keywords.

[0123] In the NLP (natural language processing) submodule of the medical model, the above conversation was subjected to in-depth language analysis, including key steps such as part-of-speech tagging and named entity recognition. Through these means, a series of keywords that may represent symptoms were initially screened and identified.

[0124] In order to further improve the accuracy and refinement of the data, OpenRefine, a powerful data processing tool, was used to deduplicate the identified symptoms. The specific steps are as follows:

[0125] First, we carefully annotated different spoken expressions of the same symptom and added corresponding medical terminology tags to them. The purpose of this step is to unify and standardize the symptom descriptions and lay the foundation for subsequent deduplication processing.

[0126] Next, we use these medical term labels to determine which symptoms belong to the same category. Specifically, if two or more symptoms have the same medical term label, we treat them as duplicates and perform deduplication.

[0127] Through such a process, redundant symptom information can be effectively removed, ensuring that the final data set is more accurate, refined, and easy to analyze. This not only improves the processing efficiency of the medical big model, but also provides more reliable data support for its application in actual medical scenarios.

[0128] (4) Semantic understanding, entity recognition and relationship extraction.

[0129] The preprocessed text data is input into the medical big model. With its profound pre-training foundation and excellent semantic understanding ability, the medical big model can deeply explore and analyze the complex medical meanings contained in the text to ensure the accuracy and depth of the information.

[0130] With the help of the medical big model, a detailed analysis of the text content was achieved. The medical big model can automatically identify and mark the key entities in the text. These entities cover multiple aspects such as symptoms, body parts, and the severity of symptoms. At the same time, the model will assign corresponding type labels to each identified entity, and finally generate a series of symptom entities for subsequent processing and analysis.

[0131] After successfully identifying the key entities in the text, the relationship extraction capability of the medical big model is further used to deeply analyze the complex relationships between these entities. For example, the model can accurately identify the modification relationship between "headache" and "long duration". This in-depth understanding of the relationship between entities provides extremely valuable and rich information support for the subsequent construction of knowledge graphs and the realization of accurate matching.

[0132] (5) Entity mapping and disease diagnosis.

[0133] First, the symptom entities analyzed by the medical big model are mapped to the vast network of the medical knowledge graph. This process is like locating specific coordinate points in the vast ocean of medical knowledge, that is, the corresponding symptom entity nodes. These nodes are not only the crystallization of medical wisdom, but also the cornerstone of subsequent analysis and diagnosis.

[0134] Secondly, once the symptom entity node is successfully located, the engine layer immediately starts the path search mechanism. In the complex network of the medical knowledge graph, the algorithm flexibly adopts depth-first search or breadth-first search strategies based on the established relationships and rules between entities, aiming to find the shortest or optimal path from the symptom entity to the disease entity. This search process not only tests the efficiency and wisdom of the algorithm, but also reflects the deep integration of medical logic and data science.

[0135] Then, after successfully finding multiple possible paths connecting symptoms and diseases, the system enters the similarity calculation stage. The core task of this stage is to accurately evaluate the similarity or match between each potential disease and the symptom description. During the evaluation process, the system will comprehensively consider multiple key factors, including but not limited to the comprehensiveness of symptoms (i.e., whether the symptoms cover all the patient's discomfort), specificity (i.e., whether the symptoms point to a specific disease), and the prevalence of the disease (i.e., the incidence of the disease in a specific population). Through these comprehensive considerations, the system can screen out the disease diagnosis that best suits the patient's actual situation, provide strong data support for doctors, and also provide patients with timely and accurate medical guidance.

[0136] (6) Generate triage suggestions and department recommendations.

[0137] In the intelligent medical system, the results of similarity calculation are not only an important basis for disease diagnosis, but also the cornerstone for generating triage recommendations. The system will intelligently generate a series of triage recommendations based on the similarity ranking of symptoms and potential diseases. These recommendations are designed to provide patients with clear medical guidance, helping them quickly find the most suitable diagnosis and treatment direction, thereby improving medical efficiency and satisfaction.

[0138] At the same time, the system will pass the identified disease entities into the graph database for in-depth query. The graph database stores rich multidimensional relational data with its unique graph structure, which covers diseases, departments, doctors and the complex relationships between them. Based on these relational data, the system can intelligently match the optimal path between the disease and the most relevant department. This matching process not only takes into account the professional field of the department, but also comprehensively considers multiple dimensions such as the department's diagnosis and treatment capabilities, the professional level of the doctor, and the actual needs of the patient. In the end, the system will recommend one or more most suitable departments for patients to ensure that they can receive the most professional and considerate medical services.

[0139] It should be noted that compared with using large models to recommend departments, knowledge graphs to recommend departments have multiple advantages such as explainability, credibility, traceability and maintainability.

[0140] Among them, explainability: the knowledge graph makes the relationship between diseases and departments clearly visible through clear structured representation, so that doctors can intuitively understand the basis for recommending departments. In contrast, the decision-making process of large models is often more complicated, and its internal mechanism is difficult to fully explain, which to some extent affects doctors' trust in the recommendation results.

[0141] Credibility: The construction of knowledge graphs is usually based on authoritative medical literature, clinical guidelines and expert experience, and has been rigorously reviewed and verified to ensure the accuracy and reliability of the information. Although large models can process massive amounts of data, their results may be affected by multiple factors such as data quality and model bias, so their credibility may be relatively low.

[0142] Traceability: Every piece of information in the knowledge graph can be traced back to its original source, which helps doctors further verify and validate the information when needed. However, when processing data, large models often merge information from multiple sources, making it difficult to directly trace back to the specific data source.

[0143] Maintainability: The knowledge graph has a flexible and reliable update mechanism, which can be adjusted and updated according to the development of medical knowledge or differences between hospital departments to improve the accuracy of recommendations.

[0144] It should be noted that the above data processing process may also include:

[0145] 1) Data interface conversion and natural language generation.

[0146] In the operation of the intelligent medical system, the data interface of the medical big model is called. This step is like building a bridge to connect the information flow between the knowledge graph and the medical big model. The entity information output by the knowledge graph is carefully sorted, passed to this interface, and converted into a format that can be easily recognized by the big model, such as structured XML data.

[0147] Once the data is successfully received, the medical big model immediately starts its internal data fusion and preprocessing mechanism. This process is like deep mining and refining of information, aiming to extract the most critical and valuable information from the complex raw data. With its powerful semantic understanding and processing capabilities, the big model can accurately grasp the internal logic and association of the data, laying a solid foundation for the subsequent natural language text generation.

[0148] Finally, based on data fusion and preprocessing, the medical big model uses its excellent natural language generation technology to convert the extracted key information into natural language text that conforms to grammatical norms and semantic coherence. These texts not only convey medical knowledge clearly and accurately, but also provide convenient and efficient medical information support for patients and doctors in a humane way.

[0149] 2) Data analysis and packaging.

[0150] When the medical model has finished outputting its disease and department information, it immediately enters the data decoding and analysis stage. This process is like decoding the code of medical information, converting the raw data output by the medical model into understandable and usable information.

[0151] In the decoding phase, specific algorithms or tools are used to decode the data output by the medical big model to restore key content such as disease names and department information. This process ensures the accuracy and integrity of the data and provides a solid foundation for subsequent data processing and utilization.

[0152] The analysis phase is to further refine and interpret the decoded data. According to the structure and content of the data, it is divided into different fields or attributes, such as disease name, department number, recommendation level, etc., to facilitate subsequent data processing and analysis.

[0153] After decoding and parsing, the data will be packaged and formatted. This step aims to organize the parsed data according to specific format requirements to ensure that it meets the format standards returned by the interface. Packaging and formatting can not only improve the readability and ease of use of data, but also ensure the security and stability of data during transmission.

[0154] Ultimately, the data that has been decoded, parsed, packaged and formatted will be accurately returned to the requester, providing strong information support for subsequent processes such as medical decision-making and patient guidance.

[0155] 3) Data replenishment and continuous optimization.

[0156] The medical information generated by this round of inquiries will be fed back into the medical big model and knowledge graph in a timely manner to ensure that the timeliness and accuracy of the system are continuously optimized.

[0157] In the continuous operation and optimization of the intelligent medical system, the medical information generated by this round of inquiries is regarded as a valuable resource. In order to ensure the continuous improvement of the timeliness and accuracy of the system, a key measure has been taken: data backlog.

[0158] In short, the process of data replenishment is to timely replenish the medical information generated in this round of inquiries into the medical big model and knowledge graph after careful screening and processing. This step not only enriches the data source of the system, but also provides strong support for the continuous optimization of the big model and knowledge graph.

[0159] For the medical big model, the newly added data can help it to continuously learn and update, thereby improving its understanding and processing capabilities of complex medical problems. These new data points are like injecting new vitality into the model, enabling it to make more accurate and reliable predictions and judgments when facing unknown or unpredictable medical situations.

[0160] For the knowledge graph, the addition of new data means the continuous expansion and improvement of its network structure. These newly added entities, relationships and attributes not only enrich the content of the graph, but also provide more possibilities for the intelligent application of the graph. Through continuous data replenishment, the knowledge graph can maintain its timeliness and accuracy, and provide more accurate and comprehensive information support for key processes such as medical decision-making and patient guidance.

[0161] In summary, data replenishment is a crucial link in the intelligent medical system. It not only ensures the continuous optimization of the timeliness and accuracy of the system, but also provides a more solid foundation for the intelligent application of medical big models and knowledge graphs.

[0162] 4) Update the user interface.

[0163] In the user interaction process of the intelligent medical system, when the user receives the response from the server, a series of sophisticated data processing and interface update operations are immediately carried out.

[0164] First, the user end will deeply analyze the data in the response body. This process is like extracting the pearls that users care about most from the server's ocean of information, that is, key information such as diseases and departments. Through specific parsing algorithms or tools, the user end can accurately identify and extract this information, making full preparations for subsequent interface updates.

[0165] Next, the extracted key information is converted into the specific format required by the interface elements. This process ensures that the information can be presented in a way that is most suitable for users to read and understand. For example, the name of the disease may be converted into an easily recognizable text label, and the department information may be converted into a button or link that the user can click. Through this format conversion, users can more easily obtain and use this information.

[0166] Finally, the user end will dynamically update the interface elements based on the converted information. This process is like putting a new set of clothes on the user interface to make it more in line with the user's current needs. Interface elements such as text boxes, buttons, links, etc. will be adjusted or updated accordingly based on the latest information to ensure that users can obtain the most accurate and comprehensive medical information in real time.

[0167] That is to say, after receiving the response from the server, the client will go through a series of delicate operations such as data analysis, format conversion and interface update to ensure that the user can easily obtain and use the latest medical information. This process not only improves the user experience, but also provides strong support for the continuous optimization and development of the intelligent medical system.

[0168] 5) Generate report.

[0169] Organizing triage recommendations into an easy-to-understand report format and presenting them clearly to patients and medical staff is an important step in improving the quality of medical services and patient satisfaction.

[0170] Specifically, first of all, the triage recommendations will be carefully sorted out and integrated. This process is like stringing scattered pearls into a beautiful necklace, ensuring that each recommendation is closely related and logically clear. According to multiple dimensions such as the severity of the disease, the professionalism of the department, and the personal needs of the patient, the triage recommendations are sorted and optimized to make them more in line with clinical reality and patient expectations.

[0171] Next, these recommendations will be presented in a concise, easy-to-understand report format. The report may use charts, lists, text descriptions, and other methods to ensure that users with different reading habits and comprehension abilities can easily obtain the required information. For example, for patients, intuitive charts can be used to show the severity of the disease and the recommended order of departments; for medical staff, more detailed symptom descriptions, disease diagnosis basis, and inter-departmental collaboration recommendations can be provided.

[0172] Finally, the report will be presented to patients and medical staff through a user-friendly interface. Whether through a mobile phone APP, a computer web page or a self-service inquiry machine in the hospital, users can easily obtain the report. At the same time, auxiliary functions such as printing and downloading will also be provided to meet the different needs of users.

[0173] In other words, organizing triage recommendations into easy-to-understand reports and clearly presenting them to patients and medical staff is an indispensable part of the smart medical system. This process not only improves the transparency and efficiency of medical services, but also builds a bridge for effective communication between patients and medical staff.

[0174] 6) Model performance evaluation and optimization.

[0175] Regularly evaluate model performance, adjust model parameters and structure, and optimize algorithm efficiency to improve the accuracy and reliability of triage recommendations.

[0176] Model performance is critical to the accuracy and reliability of triage recommendations, so a key measure was taken: regularly evaluating and optimizing the model.

[0177] This optimization process begins with a comprehensive evaluation of the model's performance. We pay close attention to the model's performance in actual applications, including its processing speed, accuracy, stability, and other aspects. By collecting and analyzing a large amount of user feedback and actual application data, we can accurately identify the shortcomings of the model.

[0178] Next, the model will be adjusted according to the evaluation results. This may include modifying the model's parameter settings to better adapt to the characteristics of different diseases and patients, or optimizing the model's structure to improve its ability to handle complex medical problems. At the same time, new algorithms and technologies will be continuously explored and introduced to improve the overall efficiency and performance of the model.

[0179] During the optimization process, improving the accuracy and reliability of triage recommendations is always the core goal. By continuously adjusting model parameters, optimizing algorithm structure, and introducing new technologies, the gap between model prediction results and actual clinical conditions can be gradually narrowed, providing patients and medical staff with more accurate and reliable triage recommendations.

[0180] In other words, regular evaluation and optimization of the smart medical model is an important means to improve the quality of triage recommendations, and continuous commitment to this work can ensure that the smart medical system can always provide patients with high-quality medical services.

[0181] The above process describes in detail the entire process of the medical big model interacting with the patient to generate triage suggestions. From receiving symptom descriptions to finally generating reports and continuously optimizing model performance, each step is closely linked and together constitutes a complete medical triage system.

[0182] In this embodiment, the problems of low efficiency, high misdiagnosis rate, and poor patient experience in the current medical triage process are solved. Through automation and intelligent methods, the patient's description of the condition is quickly understood and accurately analyzed, and personalized triage suggestions are provided to the patient, guiding him to find the appropriate department and doctor quickly and accurately; by providing clear and accurate triage suggestions, the patient is helped to have a deep understanding of his condition and the steps of medical treatment, and the anxiety and misunderstanding caused by information asymmetry are effectively alleviated. This measure not only promotes smooth communication between doctors and patients, but also significantly enhances the trust between each other, laying a solid foundation for building a harmonious doctor-patient relationship. At the same time, the high degree of transparency of the system operation makes the medical service process more open and fair, and further enhances the credibility and trustworthiness of the medical industry. Through these efforts, it is intended to create a more secure and efficient medical experience for patients and promote the continuous optimization and upgrading of medical services. In addition, the implementation of the present invention will effectively improve the utilization rate of medical resources, reduce the workload of medical staff, and promote the development of medical services in a more intelligent and humanized direction.

[0183] It is worth noting that the intelligent triage method integrates deep learning and natural language processing technology, aiming to make medical consultation intelligent and efficient. This method builds a comprehensive medical big model, integrates medical literature, clinical data, diagnostic guidelines and expert experience, and forms a structured medical knowledge system. At the same time, the knowledge graph technology is used to finely represent medical entities and their relationships, and to build a dynamic medical knowledge network. When seeking medical treatment, the system analyzes the patient's description, predicts the disease and analyzes the symptoms through the medical big model, and then uses the knowledge graph to plan the path, match the department, doctor and examination, and realize the personalized medical process. This method reduces blind registration and improves the efficiency of resource allocation. At the same time, it has the ability to continuously learn, automatically update the knowledge base, and maintain timeliness and accuracy. Overall, it provides innovative intelligent triage solutions for medical services to improve patient experience and medical service quality.

[0184] Another embodiment of the present application provides a large medical model diagnosis and guidance system.

[0185] See also Figure 2 , the medical model diagnosis and treatment system includes:

[0186] The receiving module 101 is used to receive the patient's natural language symptom description information.

[0187] The language processing module 102 is used to perform natural language processing on the symptom description information using a medical big model to determine the symptom entity information and basic information of the patient.

[0188] The matching module 103 is used to input the symptom entity information into the pre-built medical knowledge graph for matching, and determine the candidate disease and its department corresponding to the patient.

[0189] The triage module 104 is used to generate personalized triage suggestions by using the medical big model according to the candidate diseases and departments corresponding to the patient and in combination with the basic information of the patient.

[0190] Optionally, also include:

[0191] The feedback update module is used to collect patients' satisfaction with triage recommendations and actual medical results; and optimize the medical big model and the medical knowledge graph based on the satisfaction and actual medical results.

[0192] For details of the specific working process and principle of each of the above modules, please refer to the medical large model diagnosis method provided in the above embodiment, which will not be described here one by one. It depends on the actual situation and is within the protection scope of this application.

[0193] It should be noted that the system can also be divided according to the architecture, specifically into the medical big model layer, the medical big model layer, the knowledge graph layer and the intelligent diagnosis and guidance engine layer.

[0194] 1. Medical large model layer.

[0195] This layer is the core of the intelligent medical system. It uses deep learning technology, especially natural language processing technology, to deeply understand and analyze the patient's symptom description. Through pre-training with a large amount of medical literature, clinical records and real patient guidance corpus, the medical big model layer can accurately identify the symptom entities in the patient's description and understand the relationship and meaning between these symptoms. The work of this layer provides a solid foundation for subsequent accurate triage.

[0196] When deepening our understanding of the system's natural language processing (NLP) capabilities, we pay special attention to the process of receiving and processing patient symptom descriptions, using deep learning algorithms to understand text, and accurately identifying and parsing symptom entities. This level of functionality is the key foundation for the entire system to achieve precision medicine and personalized services.

[0197] First, the system receives the patient's symptom description in natural language through the user interface. These descriptions may contain the patient's subjective feelings, abnormal physical signs, disease history, etc. The system uses advanced speech recognition and text processing technology to convert the patient's oral description or written input into text data that can be processed by the computer.

[0198] After receiving the patient's symptom description, the system uses deep learning algorithms to understand the text. During this process, the system's built-in deep learning model analyzes the vocabulary, grammar, and semantic structure in the text to capture the key information in the patient's description. The selection and training of deep learning algorithms are crucial to achieve accurate recognition of symptom entities.

[0199] In order to accurately identify and parse symptom entities, the system uses advanced named entity recognition (NER) technology. NER technology can identify specific entities in text, such as disease names, symptom descriptions, body parts, etc., and match them with corresponding nodes in the knowledge graph. Through this technology, the system can accurately extract symptom entities in patient descriptions, providing key information for subsequent medical analysis and triage recommendations.

[0200] In order to achieve the above functions, the system has undergone a lot of pre-training work. The pre-training data includes a large amount of medical literature, clinical records, and real patient guidance corpus. These data provide the system with a rich medical knowledge background and semantic understanding capabilities. Through pre-training, the system can learn professional terminology in the medical field, the correspondence between common symptoms and diseases, and the logical relationship between symptoms.

[0201] In addition, the system has strong semantic understanding capabilities. This means that the system can not only recognize words and phrases in the text, but also understand the semantic relationships and contextual meanings between them. This capability enables the system to more accurately understand the patient's symptom description and provide more accurate triage recommendations.

[0202] In summary, the natural language processing layer of this system provides a solid foundation for subsequent medical analysis and triage recommendations by receiving and processing patients' natural language symptom descriptions, using deep learning algorithms for text understanding, and accurately identifying and parsing symptom entities. With rich pre-training data and powerful semantic understanding capabilities, the system can achieve accurate understanding and personalized analysis of patients' conditions, thereby improving the overall quality and efficiency of medical services.

[0203] 2. Knowledge graph layer.

[0204] Based on the acquired medical knowledge, the knowledge system is constructed using medical ontology, and the entities and relationships in the annotated knowledge are vectorized. Based on this representation, a comprehensive and dynamic medical knowledge graph is constructed, which contains a variety of medical entities such as diseases, symptoms, examinations, drugs, treatment plans, and their complex relationships (causal relationships, similarity relationships, inclusion relationships, etc.). This graph not only covers a wide range of medical knowledge, but also maintains the timeliness and accuracy of information through a real-time update mechanism.

[0205] In particular, considering the differences between hospital departments, in order to localize department information, the intelligent diagnosis and treatment engine layer supports the standardization of data from different hospitals, including unified specifications for department names, service content, professional terms, etc., and establishes mapping relationships to transform the unique data formats of each hospital into a unified format that can be processed by the knowledge graph. This design is particularly important. It is one of the advantages that promotes the better implementation of the intelligent diagnosis and treatment system, and is of great significance for promoting information sharing and collaboration and supporting personalized medical services.

[0206] Based on the collected medical data, medical ontology is used to construct a systematic knowledge system. In this system, various medical entities (such as diseases, symptoms, examination items, drugs, and treatment plans) and their interrelationships (including complex relationships such as causality, similarity, and inclusion) are carefully annotated and further converted into vectorized expressions. Based on these vectorized representations, a comprehensive medical knowledge graph with dynamic update capabilities is successfully constructed.

[0207] The atlas covers a wide range of medical knowledge. At the same time, in order to ensure the timeliness and accuracy of the information, a real-time update mechanism has been specially designed. This mechanism ensures that the information in the atlas is always synchronized with the latest medical advances.

[0208] It is worth mentioning that in view of the differences in hospital department settings, the intelligent diagnosis and treatment engine layer provides a powerful data standardization processing function. This function can standardize the data of different hospitals, including the standardization of department names, service contents, and professional terms, and establish corresponding mapping relationships. Through these processes, the unique data formats of each hospital can be smoothly converted into a unified format that can be recognized by the knowledge graph.

[0209] This design not only improves the practicality and adaptability of the system, but also plays a key role in promoting the better application of the intelligent diagnosis and treatment system in actual medical environments. It greatly promotes the sharing and collaboration of medical information and provides patients with more personalized and accurate medical services. This innovative design is undoubtedly one of the important advantages of the implementation of the intelligent diagnosis and treatment system.

[0210] 3. Intelligent diagnosis and guidance engine layer.

[0211] The intelligent diagnosis engine layer plays a key role in connecting the medical big model with the knowledge graph. Its main task is to accurately map the symptom entities analyzed by the medical big model to the vast field of the knowledge graph. With the help of a series of advanced algorithms such as graph query and path reasoning, this layer can efficiently search for the disease that best matches the patient's symptoms and recommend the most appropriate department accordingly.

[0212] In addition, the core engine layer of intelligent triage guidance also has the ability to provide personalized triage suggestions. It will comprehensively consider various aspects of the patient's information, such as age, gender, and past medical history, to ensure that the suggestions given are both scientific and considerate. This personalized service not only improves the accuracy of medical services, but also makes patients feel unprecedented care and respect.

[0213] In summary, the core engine layer of intelligent triage has become an indispensable part of the intelligent medical system with its powerful connection and reasoning capabilities, bringing patients a more efficient and accurate triage experience.

[0214] This layered design enables the intelligent medical system to better cope with complex medical scenarios and provide patients with more accurate and personalized medical services. At the same time, this design also facilitates system maintenance and upgrades, providing broad space for the future development of intelligent medical care.

[0215] In this embodiment, the receiving module 101 receives the patient's natural language symptom description information; the language processing module 102 uses the medical big model to perform natural language processing on the symptom description information to determine the patient's symptom entity information and basic information; the matching module 103 inputs the symptom entity information into the pre-constructed medical knowledge graph for matching, and determines the candidate diseases and departments corresponding to the patient; the triage module 104 uses the medical big model to generate personalized triage suggestions based on the candidate diseases and departments corresponding to the patient, and in combination with the patient's basic information; that is, based on the medical big model and the medical knowledge graph, the patient's symptom description is subjected to symptom entity recognition, and then the corresponding candidate diseases and departments are determined, thereby avoiding the use of manual triage methods, improving the efficiency and accuracy of triage, reducing the waste of medical resources, and achieving fairness and sustainability of medical services.

[0216] Another embodiment of the present application provides a storage medium having a computer program stored thereon, wherein when the computer program is executed by a processor, the method for triage and guidance of a large medical model as described in any one of the above embodiments is implemented.

[0217] In the context of the present disclosure, a machine-readable medium may be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, device, or equipment. A machine-readable medium may be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium may include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, device, or equipment, or any suitable combination of the foregoing. A more specific example of a machine-readable storage medium may include an electrical connection based on one or more lines, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0218] It should be noted that the computer-readable medium disclosed above may be a computer-readable signal medium or a computer-readable storage medium or any combination of the above two. The computer-readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or device, or any combination of the above. More specific examples of computer-readable storage media may include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In the present disclosure, a computer-readable storage medium may be any tangible medium containing or storing a program that may be used by or in combination with an instruction execution system, device or device. In the present disclosure, a computer-readable signal medium may include a data signal propagated in a baseband or as part of a carrier wave, in which a computer-readable program code is carried. This propagated data signal may take a variety of forms, including but not limited to an electromagnetic signal, an optical signal, or any suitable combination of the above. The computer readable signal medium may also be any computer readable medium other than a computer readable storage medium, which may send, propagate or transmit a program for use by or in conjunction with an instruction execution system, apparatus or device. The program code contained on the computer readable medium may be transmitted using any suitable medium, including but not limited to: wires, optical cables, RF (radio frequency), etc., or any suitable combination of the above.

[0219] The computer-readable medium may be included in the electronic device, or may exist independently without being incorporated into the electronic device.

[0220] Another embodiment of the present invention provides an electronic device, such as Figure 3 As shown, including:

[0221] One or more processors 201.

[0222] The storage device 202 stores one or more programs.

[0223] When one or more programs are executed by one or more processors 201, the one or more processors 201 implement the method for diagnosis and treatment of a large medical model as described in any one of the above embodiments.

[0224] In particular, according to an embodiment of the present disclosure, the process described above with reference to the flowchart can be implemented as a computer software program. For example, an embodiment of the present disclosure includes a computer program product, which includes a computer program carried on a non-transitory computer-readable medium, and the computer program includes a program code for executing the method shown in the flowchart.

[0225] Although the subject matter has been described in language specific to structural features and / or methodological logical actions, it should be understood that the subject matter defined in the appended claims is not necessarily limited to the specific features or actions described above. On the contrary, the specific features and actions described above are merely example forms of implementing the claims.

[0226] Although several specific implementation details are included in the above discussion, these should not be interpreted as limiting the scope of the present disclosure. Certain features described in the context of a separate embodiment can also be implemented in a single embodiment in combination. On the contrary, the various features described in the context of a single embodiment can also be implemented in multiple embodiments individually or in any suitable sub-combination.

[0227] The above description is only a preferred embodiment of the present disclosure and an explanation of the technical principles used. Those skilled in the art should understand that the scope of disclosure involved in the present disclosure is not limited to the technical solutions formed by a specific combination of the above technical features, but should also cover other technical solutions formed by any combination of the above technical features or their equivalent features without departing from the above disclosed concept. For example, the above features are replaced with the technical features with similar functions disclosed in the present disclosure (but not limited to) by each other to form a technical solution.

[0228] The features recorded in the various embodiments in this specification can be replaced or combined with each other, and the same and similar parts between the various embodiments can refer to each other, and each embodiment focuses on the differences from other embodiments. In particular, for the system or system embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and the relevant parts can refer to the partial description of the method embodiment. The system and system embodiments described above are only schematic, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules can be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Ordinary technicians in this field can understand and implement it without paying creative work.

[0229] Professionals may further appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of the two. In order to clearly illustrate the interchangeability of hardware and software, the composition and steps of each example have been generally described in the above description according to function. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professionals and technicians may use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the present invention.

[0230] The above description of the disclosed embodiments enables one skilled in the art to implement or use the present invention. Various modifications to these embodiments will be apparent to one skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but rather to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for diagnosis and treatment of a large medical model, characterized in that: include: Receive symptom description information in natural language from the patient; Using the medical big model to perform natural language processing on the symptom description information to determine the symptom entity information and basic information of the patient; Input the symptom entity information into the pre-built medical knowledge graph for matching, and determine the candidate disease and its department corresponding to the patient; The medical big model is used to generate personalized triage suggestions based on the candidate diseases and departments corresponding to the patient and combined with the basic information of the patient.

2. The method for diagnosis and treatment of a large medical model according to claim 1, characterized in that: The symptom description information is text information and / or voice information.

3. The method for diagnosis and treatment of a large medical model according to claim 1, characterized in that: Before receiving the patient's natural language symptom description information, the method further includes: The medical big model is constructed by using deep learning technology and learning samples, so that the medical big model can understand and analyze the patient's symptom description information and obtain the patient's disease entity information and basic information.

4. The method for diagnosis and treatment of a large medical model according to claim 1, characterized in that: Before receiving the patient's natural language symptom description information, the method further includes: Construct a medical knowledge graph and dynamically update the medical knowledge graph; the medical knowledge graph includes: at least two medical entities among diseases, symptoms, examinations, drugs, and treatment plans, and the relationship between at least two medical entities.

5. The method for diagnosis and treatment of a large medical model according to claim 1, characterized in that: After the medical big model generates personalized triage suggestions based on the candidate diseases and departments corresponding to the patient and in combination with the basic information of the patient, it also includes: Collect patients' satisfaction with triage recommendations and actual treatment outcomes; The medical big model and the medical knowledge graph are optimized according to the satisfaction and actual medical treatment results.

6. The method for diagnosis and treatment of a large medical model according to claim 1, characterized in that: After receiving the patient's natural language symptom description information, the method further includes: The patient's natural language symptom description information is encrypted using an encryption algorithm.

7. A medical model diagnosis and treatment system, characterized in that: include: A receiving module, used for receiving symptom description information in natural language of the patient; A language processing module, used to perform natural language processing on the symptom description information using a medical big model to determine the symptom entity information and basic information of the patient; A matching module is used to input the symptom entity information into a pre-built medical knowledge graph for matching, and determine the candidate disease and its department corresponding to the patient; The triage module is used to generate personalized triage suggestions by using the medical big model according to the candidate diseases and departments corresponding to the patient and in combination with the basic information of the patient.

8. The medical model diagnosis and guidance system according to claim 7, characterized in that: Also includes: Feedback update module, used to collect patients’ satisfaction with triage recommendations and actual treatment results; The medical big model and the medical knowledge graph are optimized according to the satisfaction and actual medical treatment results.

9. An electronic device, characterized in that: include: one or more processors; a storage device having one or more programs stored thereon; When one or more programs are executed by one or more processors, the one or more processors implement the method for diagnosis triage of a large medical model as described in any one of claims 1 to 6.

10. A storage medium, characterized in that: A computer program is stored thereon, wherein when the computer program is executed by a processor, the method for diagnosis and guidance of a large medical model as described in any one of claims 1-6 is implemented.

Images