Hospital guidance method and equipment based on large model, medium and product

Abstract

The invention provides a hospital guide method and device based on a large model, a medium and a product, and relates to the technical field of medical hospital guide. The hospital guide method comprises the following steps: inputting description information of a doctor-seeing object into a first large model to obtain doctor-seeing information; obtaining doctor seeing pointing information corresponding to the doctor seeing information, wherein the doctor seeing pointing information comprises at least one of a knowledge base retrieval result, a knowledge graph retrieval result and a similar case of the doctor seeing information; and inputting the doctor-seeing department and the doctor-seeing pointing information into the second large model, and obtaining a target doctor-seeing department of the doctor-seeing object. The embodiment of the invention can quickly and accurately help the doctor-seeing object to know the department needing to be registered, so that the doctor-seeing time of the doctor-seeing object is effectively reduced, the doctor-seeing object is prevented from being hung on a wrong department, the treatment cost is reduced, and the life health of the doctor-seeing object is protected.

Description

Technical Field

[0001] The present application relates to the field of medical guidance technology. Specifically, the present application relates to a guidance method, equipment, medium and product based on a large model. Background Art

[0002] With the development and progress of society, medical technology, which is closely related to people's life and health, has also developed rapidly. In order to simplify the medical treatment process, many medical institutions have opened online registration systems, and patients can receive medical treatment through the online registration system.

[0003] However, the existing registration system requires users to know which department they should register for before they can register correctly. However, patients are not clear about which department they should register for. They need to check a lot of information based on their symptoms or ask medical staff offline to determine the department they need to visit, which wastes the patients' medical time. In serious cases, patients register for the wrong department, delaying the time for treatment, increasing treatment costs, and even endangering the patients' lives and health, causing irreparable losses to the patients. Summary of the invention

[0004] The present application embodiment provides a large model-based guidance method, device, medium and product, which can solve the problem that it is difficult for existing patients to quickly and accurately know the department they need to register before registering, which wastes medical time, increases treatment costs, and even endangers the life and health of the patients. To achieve this purpose, the present application embodiment provides the following solutions.

[0005] According to one aspect of an embodiment of the present application, a diagnosis guidance method based on a large model is provided, the method comprising:

[0006] Inputting description information of the patient into the first large model to obtain medical information, wherein the medical information includes at least one of symptom information and medical needs;

[0007] Obtaining medical consultation information corresponding to the medical consultation information, wherein the medical consultation information includes at least one of a knowledge base search result, a knowledge graph search result, and similar cases of the medical consultation information;

[0008] The treatment department and the treatment direction information are input into the second large model to obtain the target treatment department of the treatment object. The first large model and the second large model are the same or different.

[0009] In a possible implementation, the medical consultation information includes a knowledge base search result of the medical consultation information, and obtaining medical consultation direction information corresponding to the medical consultation information includes:

[0010] Obtaining the word segmentation result of the medical information;

[0011] Based on the word segmentation result, a knowledge text in the knowledge base that matches the medical information is obtained, and the knowledge text is determined as a retrieval result of the knowledge base, wherein the knowledge base is created based on the medical text represented by the vector.

[0012] In a possible implementation, the acquiring, based on the word segmentation result, the knowledge text matching the medical consultation information in the knowledge base includes:

[0013] Vector representation of the word segmentation result, generating a word segmentation vector, and calculating the similarity between the word segmentation vector and the representation vector corresponding to the knowledge text in the knowledge base;

[0014] Calculating the relevance between the word segmentation result and the knowledge text in the knowledge base;

[0015] The knowledge text matching the medical consultation information is acquired according to the correlation and the similarity.

[0016] In a possible implementation, the medical consultation direction information includes a knowledge graph search result of the medical consultation information, and obtaining the medical consultation direction information corresponding to the medical consultation information includes:

[0017] Obtain the medical entities in the medical information, input the medical entities into a preset knowledge graph, obtain triple information corresponding to the medical entities, and determine the triple information as a knowledge graph retrieval result, wherein the knowledge graph pre-stores different medical entities and the relationships between the different medical entities, and the triple information includes symptom information and disease and department information related to the symptom information.

[0018] In a possible implementation, the construction of the knowledge base and the knowledge graph includes:

[0019] Collecting medical texts, wherein the medical texts include at least one of medical books, medical papers, hospital department catalogs, and department introductions;

[0020] removing noise data from the medical text, segmenting the denoised medical text, and generating text blocks;

[0021] The text block is represented by a vector, a knowledge base is generated based on the text block represented by the vector, and a third model is used to extract relationships from the text block, and the knowledge graph is generated based on the relationship extraction results.

[0022] In a possible implementation, the medical consultation direction information includes similar cases, and obtaining the medical consultation direction information corresponding to the medical consultation information includes:

[0023] Obtaining a representation vector corresponding to the medical information, and obtaining a similarity between the representation vector and the cases in the case database;

[0024] Similar cases corresponding to the medical information are obtained according to the similarity.

[0025] In a possible implementation, the method includes:

[0026] If it is determined that the target visiting department does not meet the preset conditions, the visiting department and the visiting direction information are input into the second large model, and the information that the generated result does not meet the preset conditions is input into the second large model to obtain a new target visiting department. The preset conditions include that the input visiting department does not include the target visiting department.

[0027] According to one aspect of an embodiment of the present application, an electronic device is provided, including a memory, a processor, and a computer program stored in the memory, wherein the processor executes the computer program to implement the steps of the method described above.

[0028] According to one aspect of an embodiment of the present application, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the steps of the method described above are implemented.

[0029] According to one aspect of an embodiment of the present application, a computer program product is provided, including a computer program, and when the computer program is executed by a processor, the steps of the method described above are implemented.

[0030] The beneficial effects of the technical solution provided by the embodiment of the present application are:

[0031] The big model-based medical guidance method provided in the present application includes: inputting the descriptive information of the patient into the first big model to obtain the medical information, which includes at least one of symptom information and medical needs; obtaining medical direction information corresponding to the medical information; inputting the medical department and the medical direction information into the second big model to obtain the target medical department of the patient, the first big model and the second big model are the same or different. The embodiment of the present application can use the big model to obtain the target medical department of the patient, and use the medical direction information to input the big model before the big model determines the target medical department, thereby improving the accuracy of the big model output results, so as to quickly and accurately help the patient know the department that needs to be registered, effectively reduce the medical time of the patient and avoid the patient from registering in the wrong department, reduce treatment costs, and protect the life and health of the patient. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the following is a brief introduction to the drawings required for use in describing the embodiments of the present application.

[0033] Figure 1 A flowchart of a large model-based diagnosis guidance method provided in an embodiment of the present application;

[0034] Figure 2 A flowchart of knowledge base retrieval provided for an embodiment of the present application;

[0035] Figure 3 A flowchart of knowledge graph retrieval provided for an embodiment of the present application;

[0036] Figure 4 A flowchart of case library retrieval provided in an embodiment of the present application;

[0037] Figure 5 A flowchart for constructing the knowledge base and knowledge graph provided in the embodiments of the present application;

[0038] Figure 6 The overall flow chart of the large model-based diagnosis guidance method provided in the embodiment of the present application;

[0039] Figure 7 This is a structural diagram of the electronic device of this application. DETAILED DESCRIPTION

[0040] The embodiments of the present application are described below in conjunction with the drawings in the present application. It should be understood that the implementation methods described below in conjunction with the drawings are exemplary descriptions for explaining the technical solutions of the embodiments of the present application and do not constitute a limitation on the technical solutions of the embodiments of the present application.

[0041] It will be understood by those skilled in the art that, unless specifically stated, the singular forms "one", "said", and "the" used herein may also include plural forms. It should be further understood that the terms "including" and "comprising" used in the embodiments of the present application refer to that the corresponding features can be implemented as the presented features, information, data, steps, operations, elements and / or components, but do not exclude the implementation of other features, information, data, steps, operations, elements, components and / or combinations thereof supported by the present technical field. It should be understood that when we say that an element is "connected" or "coupled" to another element, the one element may be directly connected or coupled to the other element, or it may refer to that the one element and the other element establish a connection relationship through an intermediate element. In addition, the "connection" or "coupling" used herein may include wireless connection or wireless coupling. The term "and / or" used herein indicates at least one of the items defined by the term, for example, "A and / or B" indicates that it is implemented as "A", or is implemented as "A", or is implemented as "A and B".

[0042] In order to make the objectives, technical solutions and advantages of the present application clearer, the implementation methods of the present application will be further described in detail below in conjunction with the accompanying drawings.

[0043] The following describes several exemplary embodiments to illustrate the technical solutions of the embodiments of the present application and the technical effects produced by the technical solutions of the present application. It should be noted that the following embodiments can refer to, draw on or combine with each other, and the same terms, similar features and similar implementation steps in different embodiments will not be described repeatedly.

[0044] The large-model-based diagnosis guidance method, equipment, medium and product provided in this application are intended to solve at least one technical problem existing in the prior art.

[0045] A large model-based medical guidance method is provided in an embodiment of the present application. The device that executes the large model-based medical guidance method may be a mobile phone, a computer, a server, a cloud platform, or other terminal that can obtain the description information of the patient and use the description information to obtain the target medical department.

[0046] Alternatively, if Figure 1-Figure 6 As shown in the figure, the guidance method based on the large model includes:

[0047] S101: Input the description information of the patient into the first large model to obtain the patient information.

[0048] Optionally, the medical consultation information includes at least one of symptom information and medical consultation requirements. The first large model can extract information related to the medical consultation information from the description information of the patient, and describe the information with a medical-related name or structure.

[0049] Optionally, before inputting the description information of the patient into the first large model, it can be detected whether the patient's visit instruction is received; if the visit instruction is received, the description information input by the patient containing the symptom description and / or the visit requirement is obtained. The patient can input the description information by voice input, key input, etc., or after receiving the visit instruction, an input box for the description information can be displayed, and the visit information can be obtained through the input box. A series of questions or multiple-choice questions for obtaining symptom information or visit requirements can also be provided to accurately obtain the visit information that matches the symptoms or visit requirements.

[0050] Optionally, after obtaining the descriptive information, the terminal executing the medical guidance method of the present application inputs the descriptive information into the first large model, so as to use the first large model to extract the symptom information and / or medical treatment needs in the descriptive information, and describe the symptom information and / or medical treatment needs with terms or nouns related to the medical field to improve the matching degree of the subsequently obtained target medical department and the descriptive information of the patient.

[0051] Optionally, the first large model may be an LLM (Large Language Model) that can perform natural language processing, such as ChatGPT, Tongyi Qianwen, Kimi, Gemini, Wenxin Yiyan, Copilot, LLaMA, SAM, SORA, etc.

[0052] Optionally, when using the first large model to obtain medical information, the first large model can be used to extract key information corresponding to the medical information in the description information, and remove the colloquial noise in the description information, or refine and purify the patient's symptom information and medical needs, and summarize them into short professional medical terms. This method can effectively remove invalid text in the doctor-patient dialogue, reduce noise, and refine key information.

[0053] S102: Obtain medical consultation direction information corresponding to the medical consultation information.

[0054] Optionally, the medical consultation information includes at least one of a knowledge base search result, a knowledge graph search result, and similar cases of the medical consultation information. After obtaining the medical consultation information using the first large model, the medical consultation information can be sent to a preset knowledge base, a knowledge graph, and a case library, respectively, so as to obtain the knowledge base search result, the knowledge graph search result, and similar cases, respectively.

[0055] Optionally, the knowledge base, knowledge graph and case library can be stored on the terminal respectively, or on an external storage device connected to the terminal (such as the cloud, server, etc.). After obtaining the medical information, the terminal calls the knowledge base, knowledge graph and case library to obtain medical reference information.

[0056] Optionally, the knowledge base is created based on the medical text after vector representation, and the knowledge graph pre-stores different medical entities and the relationship between different medical entities. The construction of the knowledge base and the knowledge graph includes: collecting medical texts, which include at least one of medical books, medical papers, hospital department catalogs, and department introductions; removing noise data from medical texts, segmenting the denoised medical texts, and generating text blocks; vectorizing the text blocks, generating a knowledge base based on the vectorized text blocks, and using the third largest model to extract relationships from the text blocks, and generating a knowledge graph based on the relationship extraction results. The third largest model can be the same as or different from the first largest model, and both can perform natural language processing.

[0057] In one embodiment, Figure 5As shown, a general medical text is obtained by crawling, data set query, etc. The general medical text includes at least one of medical books, medical papers, and definitions of various departments of general hospitals and specialized hospitals and descriptions of the scope of treatment. After obtaining the medical text, the noise data in the medical text can be removed first, such as using regular matching to remove the noise data in the text. The noise data includes HTML tags, special symbols, irrelevant characters, etc. The proportion of repeated fields in the text can also be calculated, and the text with a repeated field proportion of more than 20% can be removed. After removing the noise data, the text can be segmented by a word segmenter. After being segmented into text blocks that meet fine-grained requirements (such as segmentation based on a 200-word length limit, punctuation or blank characters are used as segmentation marks if the length is less than 200 words, and the text is truncated if the length is more than 200 words), the BGE representation model is used to vectorize the text block, thereby obtaining the representation vector corresponding to the text block, and the representation vector and the corresponding text block are stored in the database to obtain a knowledge base for obtaining retrieval knowledge text (i.e., text block). Moreover, after obtaining the text block, the third largest model can be used to obtain the medical entities in the text block and extract the relationship between the medical entities, and create a knowledge graph based on the medical entities extracted by the third largest model and the relationship between the medical entities. Among them, medical entities include entities related to medical treatment, such as diseases, symptoms, and drugs, and the relationship between medical entities includes causal relationships, therapeutic relationships, correlations, etc., which can be used to represent the information of the association between different medical entities.

[0058] Optionally, the medical information includes a knowledge base search result of the medical information, and medical reference information corresponding to the medical information is obtained, including: obtaining a word segmentation result of the medical information; obtaining knowledge text in the knowledge base that matches the medical information based on the word segmentation result, and determining the knowledge text as the knowledge base search result.

[0059] Optionally, the symptom information or medical needs of the patient in the medical information is used as the text for retrieval, and the text is segmented using a word segmenter, so that the complete text is split into independent words, and the independent words are determined as the segmentation results.

[0060] Optionally, the word segmenter may be BPE, WordPiece, ULM, SentencePiece, or other tools capable of splitting text into independent words.

[0061] Optionally, based on the word segmentation results, knowledge text in the knowledge base that matches the medical information is obtained, including: vector representation of the word segmentation results, generating a word segmentation vector, calculating the similarity between the word segmentation vector and the representation vector corresponding to the knowledge text in the knowledge base; calculating the correlation between the word segmentation results and the knowledge text in the knowledge base; and obtaining the knowledge text that matches the medical information based on the correlation and similarity.

[0062] Optionally, a representation model can be used to vectorize the word segmentation results, and for similarity, cosine similarity can be used as a measure of the similarity between the word segmentation vector and the representation vector. For relevance, it can be calculated using the BM25 score. Other algorithms or tools for calculating similarity (such as dot product) and relevance (such as TF-IDF) can also be used to obtain knowledge text that matches the medical information.

[0063] In one embodiment, Figure 2 As shown, the BGE representation model can be used to vectorize the obtained word segmentation results and transform them into word segmentation vectors. Cosine similarity is used as a similarity measurement indicator to calculate the cosine similarity of the representation vectors corresponding to the knowledge text (i.e., text blocks) in the knowledge base. The formula for calculating cosine similarity is as follows:

[0064]

[0065] Among them, A represents the word segmentation vector, B represents the representation vector, A*B is the inner product of vectors A and B, and ||A|| and ||B|| are the Euclidean norms (i.e., the lengths of the vectors) of vectors A and B respectively.

[0066] Use the BM25 algorithm to calculate the BM25 score for the word segmentation results. The BM25 score calculation formula is as follows:

[0067]

[0068] Among them, D represents the knowledge text in the knowledge base, Q represents the word segmentation result, and q i represents the i-th word in the word segmentation result, n represents the number of words in the word segmentation result, f(q i ,D) represents word q i The frequency of the word in the document, |D| represents the length of the document, avgdl represents the average length of all documents, k1 controls the saturation of the word frequency, and the value range is 1.2 to 2.0, and b controls the impact of the document length on the word frequency, and the value can be 0.75.

[0069] IDF represents word q i The inverse document frequency of is: N represents the total number of documents, n(q i ) means it contains word q i The number of documents.

[0070] Optionally, after obtaining the similarity and correlation, the matching degree between each knowledge text in the knowledge base and the medical information is calculated in combination with the similarity and correlation, and the knowledge text is filtered according to the matching degree. Specifically, the calculated similarity and BM25 score can be scored and fused, and the relevant formula is as follows:

[0071]

[0072] Among them, r(d) represents the ranking of the cosine similarity and BM25 score of knowledge text d among all knowledge texts in the knowledge base (if the similarity and BM25 score rankings are different, the average can be taken), and k represents a hyperparameter, which can be set to 60. The higher the score, the more relevant the knowledge text d should be to the medical information.

[0073] Optionally, after obtaining the matching degree between each knowledge text and the medical information, a preset number of knowledge texts with the highest matching degree are taken as the knowledge base search results. Specifically, the 10 (or 8, 12, 15 or other values) knowledge texts with the highest scores can be taken as the knowledge base search results.

[0074] Optionally, when the medical consultation direction information includes the knowledge graph search result, obtaining the medical consultation direction information corresponding to the medical consultation information includes: obtaining the medical entity in the medical consultation information, inputting the medical entity into a preset knowledge graph, obtaining the triple information corresponding to the medical entity, and determining the triple information as the knowledge graph search result, the knowledge graph pre-stores different medical entities and the relationship between different medical entities, and the triple information includes symptom information and disease and department information related to the symptom information. Among them, the medical terms related to the symptom information in the medical consultation information can be determined as the medical entities in the medical consultation information and extracted.

[0075] In one embodiment, Figure 3 As shown, the first large model is used to obtain the medical entity used to describe the symptom information in the medical information. The medical entity is used to search in the knowledge graph to obtain the triple information corresponding to the medical entity. The triple information may include the symptom information corresponding to the medical entity, and the disease or department information related to the symptom information.

[0076] Optionally, when the medical consultation direction information includes similar cases, obtaining the medical consultation direction information corresponding to the medical consultation information includes: obtaining a representation vector corresponding to the medical consultation information, obtaining the similarity between the representation vector and the cases in the case library; and obtaining similar cases corresponding to the medical consultation information based on the similarity.

[0077] Optionally, when obtaining the representation vector corresponding to the medical information, the text of the medical information is first segmented, and then the segmented text is vectorized to obtain the representation vector corresponding to the medical information.

[0078] In one embodiment, Figure 4As shown, the patient's symptoms or medical needs in the medical information are taken as text, and the text is segmented using a word segmenter to divide the complete text into independent words. The result after word segmentation is vectorized using a representation model and converted into a representation vector. Cosine similarity is used as a similarity measurement indicator to calculate the cosine similarity between the representation vector and the representation vector corresponding to the symptom description of all cases in the case library. The top 5 (or 6, 8, or other set numbers) cases with the highest cosine similarity are selected as similar cases corresponding to the medical information.

[0079] S103: Input the treatment department and treatment direction information into the second largest model to obtain the target treatment department of the patient.

[0080] Optionally, the visiting department includes information about the department of the medical institution where the patient is currently visiting, and the visiting department information can be obtained according to the patient's description information or the registration system currently used by the patient or the application logged in.

[0081] Optionally, the first large model and the second large model may be the same or different. The information of the first large model and the second large model to be used may be determined according to the selection instruction of the patient.

[0082] Optionally, when inputting the visiting department and visiting direction information into the second largest model, a prompt instructing the second largest model to select the target visiting department from the input visiting departments can also be input simultaneously to prompt the large model to select the most suitable visiting department from the given visiting departments.

[0083] Optionally, in order to improve the accuracy of the output results, the second largest model can also be used to detect whether the output target treatment department matches the input treatment department (that is, it is included in the input treatment department). If it is determined that the target treatment department does not match the input treatment department, the treatment department and the treatment direction information are input into the second largest model again, and information that the generated result does not meet the preset conditions is input into the second largest model. The preset conditions include that the input treatment department does not include the target treatment department.

[0084] Optionally, when the target visiting department does not meet the preset conditions, the step of inputting the visiting department and the visiting direction information into the second largest model may be performed multiple times until the outputted target visiting department meets the preset conditions or the number of cycles reaches the maximum number. If the number of cycles reaches the maximum number, the information of the recommended department is output.

[0085] In one embodiment, Figure 6As shown, the steps of the guidance method of this application include: 1. Use the first large model to refine and purify the patient's description information, obtain the patient's symptom information and medical needs, and summarize them into short professional medical terms. This approach can effectively remove invalid text in doctor-patient dialogues, reduce noise, and refine key information. 2. Use the knowledge base to retrieve knowledge and obtain the 10 most similar knowledge texts. 3. Use the knowledge graph to retrieve knowledge and obtain triple knowledge related to symptom information. 4. Use the case library to retrieve similar cases and obtain 5 similar cases. 5. The obtained knowledge text, triple knowledge, similar case symptom descriptions and treatment departments are input into the second large model together to strengthen the second large model's understanding of the medical knowledge background of the patient's symptoms, and give the names of all departments covered by the hospital, prompting the large model to select the most suitable treatment department in the given department. 6. Use the second large model to self-check the generated department to see if it exists in the limited department. If yes, the output is the final result, otherwise the step 5 is called repeatedly, and it is prompted that the result generated by the second largest model does not meet the conditions, until the generated result meets the conditions or exceeds the maximum number of calls (such as 5 times). 7. The department that meets the conditions is determined as the final recommended department, and the final recommended department is output. If the self-check fails, the information that there is no recommended department is output. Among them, the first large model and the second large model can be the same large model.

[0086] The guidance method system based on the big model in this application can quickly process a large amount of descriptive information input by patients, and quickly recommend suitable departments or doctors based on the symptoms and historical data related to the descriptive information, which greatly reduces the waiting time for patients and improves the work efficiency of the hospital. The text retrieval enhancement, knowledge graph retrieval enhancement, case library retrieval enhancement and big model self-inspection methods covered in this application can effectively inject medical knowledge and hospital-specific department information into the big model. The big model enhanced with professional medical knowledge can provide patients with efficient and accurate guidance services. This method can also be adapted to various types of hospitals, including general hospitals and specialized hospitals. Updating the proprietary departments and case library use cases of various hospitals to the big model knowledge base can realize the generalization of this method to multiple hospitals and realize hospital-specific intelligent guidance services.

[0087] In an optional embodiment, an electronic device is provided, such as Figure 7 As shown, Figure 7The electronic device 4000 shown includes: a processor 4001 and a memory 4003. The processor 4001 and the memory 4003 are connected, such as through a bus 4002. Optionally, the electronic device 4000 may also include a transceiver 4004, which may be used for data interaction between the electronic device and other electronic devices, such as data transmission and / or data reception. It should be noted that in actual applications, the transceiver 4004 is not limited to one, and the structure of the electronic device 4000 does not constitute a limitation on the embodiments of the present application.

[0088] Processor 4001 may be a CPU (Central Processing Unit), a general-purpose processor, a DSP (Digital Signal Processor), an ASIC (Application Specific Integrated Circuit), an FPGA (Field Programmable Gate Array), or other programmable logic devices, transistor logic devices, hardware components, or any combination thereof. It may implement or execute various exemplary logic blocks, modules, and circuits described in conjunction with the disclosure of this application. Processor 4001 may also be a combination that implements computing functions, such as a combination of one or more microprocessors, a combination of a DSP and a microprocessor, and the like.

[0089] The bus 4002 may include a path to transmit information between the above components. The bus 4002 may be a PCI (Peripheral Component Interconnect) bus or an EISA (Extended Industry Standard Architecture) bus. The bus 4002 may be divided into an address bus, a data bus, a control bus, etc. For ease of representation, only one thick line is used in the figure, but it does not mean that there is only one bus or one type of bus.

[0090] Memory 4003 can be ROM (Read Only Memory) or other types of static storage devices that can store static information and instructions, RAM (Random Access Memory) or other types of dynamic storage devices that can store information and instructions, or it can be EEPROM (Electrically Erasable Programmable Read Only Memory), CD-ROM (Compact Disc Read Only Memory) or other optical disk storage, optical disk storage (including compressed optical disk, laser disc, optical disk, digital versatile disc, Blu-ray disc, etc.), magnetic disk storage medium, other magnetic storage devices or any other medium that can be used to carry or store computer programs and can be read by a computer, without limitation here.

[0091] The memory 4003 is used to store the computer program for executing the embodiment of the present application, and the execution is controlled by the processor 4001. The processor 4001 is used to execute the computer program stored in the memory 4003 to implement the steps shown in the above method embodiment.

[0092] Among them, the electronic device can be any electronic product that can interact with an object, such as a personal computer, a tablet computer, a smart phone, a personal digital assistant (PDA), a game console, an interactive network television (IPTV), a smart wearable device, etc.

[0093] The electronic device may also include a network device and / or an object device, wherein the network device includes, but is not limited to, a single network server, a server group consisting of multiple network servers, or a cloud consisting of a large number of hosts or network servers for cloud computing.

[0094] The network where the electronic device is located includes but is not limited to the Internet, a wide area network, a metropolitan area network, a local area network, a virtual private network (Virtual Private Network, VPN), etc.

[0095] An embodiment of the present application provides a computer-readable storage medium, which includes a computer program. When the computer program is executed by a processor, the steps and corresponding contents of the aforementioned method embodiment can be implemented.

[0096] An embodiment of the present application provides a computer program product, which includes a computer program. When the computer program is executed by a processor, the steps and corresponding contents of the aforementioned method embodiment can be implemented.

[0097] The terms "first", "second", "third", "fourth", "1", "2", etc. (if any) in the specification and claims of this application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the numbers used in this way can be interchanged where appropriate, so that the embodiments of the present application described herein can be implemented in an order other than that shown or described in the drawings.

[0098] It should be understood that, although each operation step is indicated by arrows in the flowchart of the embodiment of the present application, the implementation order of these steps is not limited to the order indicated by the arrows. Unless clearly stated herein, in some implementation scenarios of the embodiment of the present application, the implementation steps in each flowchart can be performed in other orders according to demand. In addition, some or all of the steps in each flowchart may include multiple sub-steps or multiple stages based on actual implementation scenarios. Some or all of these sub-steps or stages may be executed at the same time, and each sub-step or stage in these sub-steps or stages may also be executed at different times respectively. In different scenarios of execution time, the execution order of these sub-steps or stages may be flexibly configured according to demand, and the embodiment of the present application does not limit this.

[0099] The above is only an optional implementation method for some implementation scenarios of the present application. It should be pointed out that for ordinary technicians in this technical field, without departing from the technical concept of the solution of the present application, other similar implementation methods based on the technical ideas of the present application are also within the protection scope of the embodiments of the present application.

Claims

1. A diagnosis guidance method based on a large model, characterized in that: The method comprises: Inputting description information of the patient into the first large model to obtain medical information, wherein the medical information includes at least one of symptom information and medical needs; Obtaining medical consultation information corresponding to the medical consultation information, wherein the medical consultation information includes at least one of a knowledge base search result, a knowledge graph search result, and similar cases of the medical consultation information; The treatment department and the treatment direction information are input into the second large model to obtain the target treatment department of the treatment object. The first large model and the second large model are the same or different.

2. The large model-based diagnosis guidance method according to claim 1, characterized in that: The medical consultation information includes a knowledge base search result of the medical consultation information, and obtaining medical consultation direction information corresponding to the medical consultation information includes: Obtaining the word segmentation result of the medical information; Based on the word segmentation result, a knowledge text in the knowledge base that matches the medical information is obtained, and the knowledge text is determined as a retrieval result of the knowledge base, wherein the knowledge base is created based on the medical text represented by the vector.

3. The large model-based diagnosis guidance method according to claim 2, characterized in that: The step of acquiring the knowledge text matching the medical consultation information in the knowledge base based on the word segmentation result includes: Vector representation of the word segmentation result, generating a word segmentation vector, and calculating the similarity between the word segmentation vector and the representation vector corresponding to the knowledge text in the knowledge base; Calculating the relevance between the word segmentation result and the knowledge text in the knowledge base; The knowledge text matching the medical consultation information is acquired according to the correlation and the similarity.

4. The large model-based diagnosis guidance method according to claim 2, characterized in that: The medical consultation direction information includes a knowledge graph search result of the medical consultation information, and obtaining the medical consultation direction information corresponding to the medical consultation information includes: Obtain the medical entities in the medical information, input the medical entities into a preset knowledge graph, obtain triple information corresponding to the medical entities, and determine the triple information as a knowledge graph retrieval result, wherein the knowledge graph pre-stores different medical entities and the relationships between the different medical entities, and the triple information includes symptom information and disease and department information related to the symptom information.

5. The large model-based diagnosis guidance method according to claim 4, characterized in that: The construction of the knowledge base and the knowledge graph includes: Collecting medical texts, wherein the medical texts include at least one of medical books, medical papers, hospital department catalogs, and department introductions; removing noise data from the medical text, segmenting the denoised medical text, and generating text blocks; The text block is represented by a vector, a knowledge base is generated based on the text block represented by the vector, and a third model is used to extract relationships from the text block, and the knowledge graph is generated based on the relationship extraction results.

6. The large model-based diagnosis guidance method according to claim 1, characterized in that: The medical consultation information includes similar cases, and obtaining the medical consultation information corresponding to the medical consultation information includes: Obtaining a representation vector corresponding to the medical information, and obtaining a similarity between the representation vector and the cases in the case database; Similar cases corresponding to the medical information are obtained according to the similarity.

7. The large model-based diagnosis guidance method according to claim 1, characterized in that: The method comprises: If it is determined that the target visiting department does not meet the preset conditions, the visiting department and the visiting direction information are input into the second large model, and the information that the generated result does not meet the preset conditions is input into the second large model to obtain a new target visiting department. The preset conditions include that the input visiting department does not include the target visiting department.

8. An electronic device comprising a memory, a processor and a computer program stored in the memory, characterized in that: The processor executes the computer program to implement the steps of the method according to any one of claims 1 to 7.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 7 are implemented.

10. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 7 are implemented.

Images