Emergency triage question-answering system based on knowledge graph

Abstract

The invention discloses an emergency triage question-answering system based on a knowledge graph, and relates to the technical field of medical information, and the emergency triage question-answering system comprises an information collection module, a question-answering module and a knowledge graph management module, the question and answer module is used for performing interactive question and answer on the specific condition of the disease after the patient logs in and pushing related questions and instrument examination suggestions, and the knowledge graph management module is used for performing optimization in combination with an actual diagnosis result to construct a new disease knowledge graph. The information collection module comprises a patient information input module, a diagnosis recording device, a text processing module, a key information labeling module and an examination result collection module, the diagnosis recording device is electrically connected with the text processing module, and the key information labeling module is used for enabling a doctor to label important symptoms in a transcriptional text. The method has the characteristic of combining actual optimization.

Description

Technical Field

[0001] The present invention relates to the field of medical information technology, and particularly to an emergency triage Q&A system based on a knowledge graph. Background Art

[0002] An emergency triage Q&A system is an intelligent system that uses knowledge graph technology and aims to provide timely and efficient triage guidance and health consultations for emergency patients. It can help patients quickly understand information such as their own symptoms, diseases they belong to, and medical advice, thereby saving medical treatment time and optimizing the allocation of medical resources.

[0003] When the existing emergency triage Q&A system is in use, the update of its knowledge graph often depends on the system operator for unified data entry. During the actual medical treatment process, the judgment results may be inconsistent with the manual judgment results of doctors because doctors often additionally learn some symptoms and give more accurate judgments in combination with the examination results of instruments. These data are difficult to collect, resulting in the inability of the Q&A system to optimize the knowledge graph according to the actual situation. Therefore, it is necessary to design an emergency triage Q&A system based on a knowledge graph that combines actual optimization. Summary of the Invention

[0004] The purpose of the present invention is to provide an emergency triage Q&A system based on a knowledge graph to solve the problems raised in the above background art.

[0005] To solve the above technical problems, the present invention provides the following technical solution: An emergency triage Q&A system based on a knowledge graph includes an information collection module, a Q&A module, and a knowledge graph management module. The information collection module is used to collect doctors' diagnostic data and the examination result data of instruments. The Q&A module is used for interactive Q&A about the specific situation of diseases after patients log in, and for pushing relevant questions and instrument examination suggestions. The knowledge graph management module is used to optimize based on the actual diagnosis results and construct a new disease knowledge graph.

[0006] According to the above technical solution, the information collection module includes a patient information entry module, a diagnostic recording device, a text processing module, a key information annotation module, and an examination result collection module. The diagnostic recording device is electrically connected to the text processing module, and the text processing module is electrically connected to the key information annotation module. The patient information entry module is used to record the basic information and symptoms of patients. The diagnostic recording device is used to record the conversation between doctors and patients during the diagnosis process in real time. The text processing module is used to convert the recording content into text format. The key information annotation module is used for doctors to annotate important symptoms in the transcribed text and manually input diagnostic symptoms not reflected in the language. The examination result collection module is used to collect the results of symptoms detected by instruments by doctors;

[0007] The Q&A module includes a symptom query module, an intelligent analysis module, a related question push module, and an instrument examination suggestion module. The symptom query module is electrically connected to the patient information input module, the intelligent analysis module is electrically connected to the symptom query module, and the related question push module is electrically connected to the key information annotation module. The symptom query module is used for patients to input symptoms in the Q&A system, and the system uses the knowledge graph for preliminary disease reasoning. The intelligent analysis module is used to analyze the patient's symptoms and give a disease diagnosis result and medical advice based on the knowledge graph. The related question push module is used to push other potential symptom judgments related to assist the patient in further describing the symptoms. The instrument examination suggestion module is used to recommend relevant instrument examinations to the patient according to the preliminary diagnosis result;

[0008] The knowledge graph management module includes a knowledge graph construction module, a knowledge update module, and an error optimization module. Both the knowledge graph construction module and the knowledge update module are electrically connected to the intelligent analysis module, and the error optimization module is electrically connected to the knowledge graph construction module. The knowledge graph construction module is used to construct a knowledge graph that deduces from symptoms and instrument examination results to disease diagnosis results. The knowledge update module is used to construct a new knowledge graph branch according to the symptoms additionally learned by the doctor in combination with the symptoms entered by the patient and the instrument examination results. The error optimization module is used to modify the knowledge graph when there is redundancy in the knowledge graph branches of the system.

[0009] According to the above technical solution, the working method of the system is as follows:

[0010] S1. Use each symptom to construct a label as the explicit cause in the knowledge graph, use the results of each instrument examination of symptoms to construct a label as the implicit cause in the knowledge graph, and use each disease name to construct a label as the diagnosis result in the knowledge graph. Any combination form of an explicit cause and an implicit cause corresponding to a diagnosis result is a knowledge graph branch;

[0011] S2. The patient logs in through the Q&A system and enters basic information and symptoms. The system finds all knowledge graph branches containing the explicit causes corresponding to the symptoms according to the symptoms, and guides the patient to further supplement other symptoms according to the knowledge graph branches to obtain the corresponding diagnosis result and instrument examination suggestion. The doctor views this information in both aspects before the patient's face-to-face consultation to have a preliminary understanding of the patient's condition;

[0012] S3. During the face-to-face consultation, record the conversation between the doctor and the patient in real time, convert the recorded content into an editable text format. At the end of the face-to-face consultation, the doctor inputs the symptoms not reflected in the language into the key information annotation module and annotates the symptoms in the editable text as the explicit causes;

[0013] S4. The patient undergoes instrument examinations according to the instrument examination suggestions given by the Q&A system and the doctor. After analyzing the examination result data, the doctor enters the results of the symptoms detected by the instrument examination as the implicit causes;

[0014] S5. The doctor gives a diagnosis result and compares it with the diagnosis result given by the Q&A system. If the diagnosis result given by the doctor is included in the diagnosis result given by the Q&A system, it means that the existing knowledge graph of the system is comprehensive enough for the disease diagnosis of this patient. If it is not included, it is not comprehensive. When the existing knowledge graph of the system is not comprehensive, a new knowledge graph branch is constructed. When the knowledge graph branch of the system is redundant, the knowledge graph is modified.

[0015] According to the above technical solution, in S2, the specific method for guiding the patient to further supplement other symptoms according to the knowledge graph branch is as follows:

[0016] S2-1. After the patient inputs symptoms, let the patient rate the severity x i of each symptom to obtain {x 1 , x 2 , …, x n}, and input the duration t i of each symptom to obtain

[0017] {t 1 , t 2 , …, t n}, where i is the serial number of the symptom type, i ∈ 1 to n, and compare with the average severity Δx i of the symptoms recorded in the system and the average duration Δt i of the symptoms to obtain a proportional value and calculate the urgency of the disease. The formula is urgency where k 1 is the severity coefficient and k 2 is the duration coefficient;

[0018] S2-2. The system displays the explicit causes involved in other knowledge graph branches to the patient. The patient manually selects whether the displayed explicit causes appear, supplements other symptoms that the patient has not entered, and adjusts the number of times of guiding the patient to further supplement other symptoms according to the urgency A. As the urgency A increases, the number of times of guiding the patient to further supplement other symptoms decreases. Let w be the number of knowledge graph branches corresponding to the symptoms with explicit causes entered by this patient. The number of times M of guiding the patient to further supplement other symptoms is M = [w - αA], M is an integer, and α is the influence coefficient of the urgency A on the number of times M of supplementing other symptoms.

[0019] According to the above technical solution, in S3, when converting the recorded content into an editable text format, all sentences containing symptoms in the text are highlighted, and the doctor selects symptom words that can be used as explicit causes from the highlighted sentences.

[0020] According to the above technical solution, in S2, the matching diagnostic results are all the diagnostic results corresponding to this combination form of dominant causes in the knowledge graph branch after the patient supplements other symptoms; the method for obtaining the instrument examination suggestions is as follows: list all the combination forms of dominant causes that include this combination form of dominant causes, find all the corresponding diagnostic results in the knowledge graph branch, and reverse deduce the results of all symptoms detected by instruments when various diseases occur, and use these instrument examination items as the instrument examination suggestions.

[0021] According to the above technical solution, in S5, when the existing knowledge graph is incomplete, the specific method for constructing a new knowledge graph branch is as follows: combine all the dominant causes given by the doctor and the question-and-answer system during the patient's visit this time, and all the recessive causes detected during the instrument examination, and then correspond them to the diagnostic results to construct the knowledge graph branch under this combination.

[0022] According to the above technical solution, in S5, the specific method for modifying the knowledge graph is as follows: when the new knowledge graph branch is constructed, first retrieve whether there is a combination form with fewer dominant causes and recessive causes corresponding to this diagnostic result. When there is a combination form with fewer causes, define the redundant dominant causes and recessive causes as concurrent symptoms, and count all the cases with this disease as the diagnostic result subsequently, and record the number of times a without concurrent symptoms 1 and the number of times a with concurrent symptoms 2 , and obtain When is larger, the probability of this concurrent symptom occurring together with this disease is greater. When guiding the patient to further supplement other symptoms, this concurrent symptom is preferentially displayed to the patient for their judgment.

[0023] Compared with the prior art, the beneficial effects achieved by the present invention are as follows: in the present invention, after the patient uses the question-and-answer system to obtain a preliminary judgment result, the doctor further diagnoses in combination with the judgment result. By recording the entire process of the doctor's diagnosis, recording it in text form and then manually marking the key sentences by the doctor, and collecting the instrument examination result data at the same time. If the system makes a wrong judgment, a new disease knowledge graph will be constructed, and when other patients use the question-and-answer system for similar situations, the system will automatically push relevant questions and instrument examination suggestions, which can be optimized in combination with the actual diagnostic results, and gradually improve the judgment accuracy with the accumulation of the system application time, further saving the doctor's workload. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] The drawings are used to provide a further understanding of the present invention, and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation to the present invention. In the drawings:

[0025] Figure 1 It is a schematic diagram of the overall module structure of the present invention. Specific embodiments

[0026] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0027] Please refer to Figure 1 , the present invention provides a technical solution: an emergency triage Q&A system based on a knowledge graph, including an information collection module, a Q&A module, and a knowledge graph management module. The information collection module is used to collect doctors' diagnostic data and the inspection result data of instruments. The Q&A module is used for interactive Q&A about the specific conditions of diseases after patients log in, and push relevant questions and instrument inspection suggestions. The knowledge graph management module is used to optimize in combination with the actual diagnosis results to construct a new disease knowledge graph;

[0028] The information collection module includes a patient information entry module, a diagnostic recording device, a text processing module, a key information annotation module, and an inspection result collection module. The diagnostic recording device is electrically connected to the text processing module, and the text processing module is electrically connected to the key information annotation module. The patient information entry module is used to record the basic information and symptoms of patients. The diagnostic recording device is used to record the conversation between doctors and patients during the diagnosis process in real time. The text processing module is used to convert the recording content into a text format. The key information annotation module is used for doctors to annotate important symptoms in the transcribed text and manually input diagnostic symptoms not reflected in the language. The inspection result collection module is used to collect the results of the symptoms detected by the doctor through the instrument;

[0029] The Q&A module includes a symptom query module, an intelligent analysis module, a relevant question push module, and an instrument inspection suggestion module. The symptom query module is electrically connected to the patient information entry module, the intelligent analysis module is electrically connected to the symptom query module, and the relevant question push module is electrically connected to the key information annotation module. The symptom query module is used for patients to input symptoms in the Q&A system, and the system uses the knowledge graph for preliminary disease reasoning. The intelligent analysis module is used to analyze the symptoms of patients and give the disease diagnosis results and medical treatment suggestions based on the knowledge graph. The relevant question push module is used to push other potential symptom judgments related to assist patients in further describing symptoms. The instrument inspection suggestion module is used to recommend relevant instrument inspections to patients according to the preliminary diagnosis results;

[0030] The knowledge graph management module includes a knowledge graph construction module, a knowledge update module, and an error optimization module. The knowledge graph construction module and the knowledge update module are both electrically connected to the intelligent analysis module, and the error optimization module is electrically connected to the knowledge graph construction module. The knowledge graph construction module is used to construct a knowledge graph that deduces disease diagnosis results from symptoms and instrument examination results. The knowledge update module is used to construct new branches of the knowledge graph based on the symptoms additionally learned by the doctor in combination with the symptoms and instrument examination results entered by the patient. The error optimization module is used to modify the knowledge graph when there are redundant branches in the system knowledge graph;

[0031] The working method of this system is as follows:

[0032] S1. Use the construction of each symptom as a label as the explicit cause in the knowledge graph, the result of each instrument examination symptom as a label as the implicit cause in the knowledge graph, and each disease name as a label as the diagnosis result in the knowledge graph. Any combination form of an explicit cause and an implicit cause corresponding to a diagnosis result is a branch of the knowledge graph;

[0033] S2. The patient logs in through the Q&A system and enters basic information and symptoms. The system finds all branches of the knowledge graph containing the explicit causes corresponding to the symptoms according to the symptoms, and guides the patient to further supplement other symptoms according to the branches of the knowledge graph, and obtains the corresponding diagnosis results and instrument examination suggestions. The doctor views this information in both aspects before the patient's face-to-face consultation to have a preliminary understanding of the patient's condition;

[0034] S3. During the face-to-face consultation, the conversation between the doctor and the patient is recorded in real time, and the recorded content is converted into an editable text format. At the end of the face-to-face consultation, the doctor inputs the symptoms not reflected in the language to the key information annotation module, and annotates the symptoms in the editable text as the explicit causes;

[0035] S4. The patient undergoes instrument examinations according to the Q&A system and the instrument examination suggestions given by the doctor. After analyzing the examination result data, the doctor enters the results of the symptoms detected by the instrument as the implicit causes;

[0036] S5. The doctor gives a diagnosis result and compares it with the diagnosis result given by the Q&A system. If the diagnosis result given by the doctor is included in the diagnosis result given by the Q&A system, it means that the existing knowledge graph of the system is comprehensive enough for the disease diagnosis of this patient. If it is not included, it is not comprehensive. When the existing knowledge graph of the system is not comprehensive, new branches of the knowledge graph are constructed, and when there are redundant branches in the system knowledge graph, the knowledge graph is modified;

[0037] In S2, the specific method of guiding the patient to further supplement other symptoms according to the branches of the knowledge graph is as follows:

[0038] S2-1. After the patient inputs symptoms, let the patient rate the severity x i of each symptom to obtain {x 1 , x 2 , …, x n}, and input the duration t i of each symptom to obtain

[0039] {t 1 , t 2 , …, t n}, where i is the serial number of symptom types, i ∈ 1~n, and compare with the average severity Δx i of symptoms recorded in the system and the average duration Δt i of symptoms to obtain a proportional value and calculate the urgency of the disease. The formula is Urgency where k 1 is the severity coefficient and k 2 is the duration coefficient;

[0040] S2-2. The system displays the explicit causes involved in other knowledge graph branches to the patient. The patient manually selects whether the displayed explicit causes occur, supplements other symptoms not entered by the patient, and adjusts the number of times to guide the patient to further supplement other symptoms according to the urgency A. As the urgency A increases, the number of times to guide the patient to further supplement other symptoms decreases. Let w be the number of knowledge graph branches containing the explicit causes corresponding to the symptoms entered by the patient this time. The number of times M to guide the patient to further supplement other symptoms is M = [w - αA], M is an integer, and α is the influence coefficient of the urgency A on the number of times M to supplement other symptoms; in order to avoid too long waiting time for patients in emergency situations, when the urgency of the patient's symptoms is high, the number of times to supplement other symptoms is adaptively reduced, while patients with low urgency have more sufficient times to supplement other possible symptoms, making the system judgment more comprehensive and reducing the diagnosis time of the patient at the doctor's place.

[0041] In S3, when converting the recorded content into an editable text format, all sentences containing symptoms in the text are highlighted, and the doctor selects symptom words that can be used as explicit causes from the highlighted sentences;

[0042] In S2, the conforming diagnostic results are all the diagnostic results corresponding to this combination form of explicit causes in the knowledge graph branch after the patient supplements other symptoms; the method for obtaining the instrument examination suggestions is: list all the combination forms of explicit causes containing this combination form of explicit causes, find all the corresponding diagnostic results in the knowledge graph branch, and reverse deduce the results of all symptoms detected by instruments when various diseases occur from the diagnostic results, and use these instrument examination items as the instrument examination suggestions;

[0043] In S5, when the existing knowledge graph is incomplete, the specific method for constructing a new branch of the knowledge graph is as follows: Combine all the explicit causes given by the doctor and the question-and-answer system during the patient's current visit, as well as all the implicit causes detected during the instrument examination, and then correspond them to the diagnosis results to construct a branch of the knowledge graph under this combination;

[0044] In S5, the specific method for modifying the knowledge graph is as follows: When the construction of a new branch of the knowledge graph is completed, first retrieve whether there is a combination form with fewer explicit causes and implicit causes corresponding to this diagnosis result. When a combination form with fewer causes appears, define the redundant explicit causes and implicit causes as concurrent symptoms, and count all subsequent cases with this disease as the diagnosis result, and record the number of times a without concurrent symptoms 1 and the number of times b with concurrent symptoms 2 , and obtain When is larger, the probability that this concurrent symptom occurs together with this disease is higher. When guiding the patient to further supplement other symptoms, this concurrent symptom is more likely to be displayed to the patient for judgment first. Since many diseases have many concurrent symptoms, and these concurrent symptoms vary from person to person and do not necessarily occur with the appearance of this disease, but the occurrence probability is also very high. If concurrent symptoms are not considered when constructing the knowledge graph, many probabilities of correctly diagnosing the disease will be missed. Therefore, considering the occurrence probability of this complication, the higher the probability, the more likely it will appear in the process of guiding the patient to supplement symptoms of the disease, helping the patient to determine which specific disease they have more quickly.

[0045] It should be noted that in this article, relational terms such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the term "comprising", "including" or any other variant thereof is intended to cover non-exclusive inclusion, so that a process, method, article or device comprising a series of elements not only includes those elements, but also includes other elements not expressly listed, or elements inherent to such process, method, article or device.

[0046] Finally, it should be noted that the above are only the preferred embodiments of the present invention and are not used to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments, or perform equivalent replacements on some of the technical features. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.

Claims

1. An emergency triage question-answering system based on knowledge graph, characterized by: It includes an information collection module, a question-and-answer module, and a knowledge graph management module. The information collection module is used to collect doctors' diagnostic data and instrument examination result data. The question-and-answer module is used for patients to interactively ask questions about the specific conditions of the disease after logging in, as well as to push related questions and instrument examination suggestions. The knowledge graph management module is used to optimize based on actual diagnostic results and construct a new disease knowledge graph.

2. An emergency triage question-answering system based on knowledge graph according to claim 1, characterized in that: The information collection module includes a patient information entry module, a diagnosis recording device, a text processing module, a key information marking module, and an examination result collection module. The diagnosis recording device is electrically connected to the text processing module, and the text processing module is electrically connected to the key information marking module. The patient information entry module is used to record the patient's basic information and symptoms. The diagnosis recording device is used to record the doctor's conversation with the patient during the diagnosis process in real time. The text processing module is used to convert the recording content into a text format. The key information marking module is used to allow the doctor to mark important symptoms in the transcribed text, and manually input diagnostic symptoms that are not reflected in the language. The examination result collection module is used to collect the doctor's results of the symptoms detected by the instrument; The question-answering module includes a symptom query module, an intelligent analysis module, a related question push module, and an instrument inspection suggestion module. The symptom query module is electrically connected to the patient information entry module, the intelligent analysis module is electrically connected to the symptom query module, and the related question push module is electrically connected to the key information annotation module. The symptom query module is used for patients to input symptoms in the question-answering system, and the system uses the knowledge graph to perform preliminary disease reasoning. The intelligent analysis module is used to analyze the patient's symptoms and give the disease diagnosis results and medical advice based on the knowledge graph. The related question push module is used to push other related potential symptom judgments to assist patients in further describing symptoms. The instrument inspection suggestion module is used to recommend related instrument inspections to patients based on the preliminary diagnosis results; The knowledge graph management module includes a knowledge graph construction module, a knowledge update module, and an error optimization module. The knowledge graph construction module and the knowledge update module are both electrically connected to the intelligent analysis module. The error optimization module is electrically connected to the knowledge graph construction module. The knowledge graph construction module is used to construct a knowledge graph that derives disease diagnosis results from symptoms and instrument examination results. The knowledge update module is used to construct a new knowledge graph branch based on the additional symptoms learned by the doctor combined with the symptoms and instrument examination results entered by the patient. The error optimization module is used to modify the knowledge graph when the system knowledge graph branches are redundant.

3. An emergency triage question-answering system based on knowledge graph according to claim 2, characterized in that: The system works as follows: S1. Labels are constructed with each symptom as explicit causes in the knowledge graph, labels are constructed with the results of each instrument checking the symptoms as implicit causes in the knowledge graph, labels are constructed with each disease name as the diagnosis result in the knowledge graph, and any combination of explicit causes and implicit causes corresponding to a diagnosis result is a knowledge graph branch; S2. The patient logs in through the question-and-answer system and enters basic information and symptoms. The system finds all knowledge graph branches that contain the explicit causes corresponding to the symptoms based on the symptoms, and guides the patient to further supplement other symptoms based on the knowledge graph branches, and obtains appropriate diagnosis results and instrument examination recommendations. The doctor checks these two aspects of information before the patient's face-to-face consultation to gain a preliminary understanding of the patient's condition; S3. During the face-to-face consultation, the conversation between the doctor and the patient is recorded in real time and converted into an editable text format. At the end of the face-to-face consultation, the doctor inputs the symptoms that are not reflected in the language into the key information annotation module and annotates the symptoms in the editable text as explicit causes; S4. The patient undergoes an instrumental examination according to the question-answering system and the doctor's recommendations. The doctor analyzes the examination result data and enters the results of the instrumental examination symptoms as hidden causes. S5. The doctor gives a diagnosis result, which is compared with the diagnosis result given by the question-answering system. If the diagnosis result given by the doctor is included in the diagnosis result given by the question-answering system, it means that the existing knowledge graph of the system is comprehensive enough for the diagnosis of the patient's disease. If it is not included, it is incomplete. When the existing knowledge graph of the system is incomplete, a new knowledge graph branch is constructed. When the system knowledge graph branch is redundant, the knowledge graph is modified.

4. An emergency triage question-answering system based on knowledge graph according to claim 3, characterized in that: In S2, the specific method of guiding patients to further supplement other symptoms based on the knowledge graph branch is: S2-1. After the patient inputs the symptoms, the patient is asked to rate the severity of each symptom x i Rating, get {x1, x2, …, x n }, and the duration of each symptom t i Enter and get {t1, t2, …, t n }, where i is the symptom type number, i∈1~n, and is related to the average symptom severity Δx recorded in the system i and the average duration of symptoms Δt i Compare and find the ratio and calculate the urgency of the disease. The formula is: Where k1 is the severity coefficient and k2 is the duration coefficient; S2-2. The system displays the explicit causes involved in other knowledge graph branches to the patient. The patient manually chooses whether to display the explicit causes and supplements other symptoms that the patient has not entered. The number of times the patient is guided to further supplement other symptoms is adjusted according to the urgency A. As the urgency A increases, the number of times the patient is guided to further supplement other symptoms becomes lower. Let w be the number of all knowledge graph branches containing the explicit causes corresponding to the symptoms entered by the patient this time. The number of times the patient is guided to further supplement other symptoms is M = [w-αA], where M is an integer and α is the influence coefficient of the urgency A on the number of times other symptoms are supplemented M.

5. An emergency triage question-answering system based on knowledge graph according to claim 4, characterized in that: In S3, when the audio recording is converted into an editable text format, all sentences containing symptoms in the text are highlighted, and the doctor selects symptom words that can serve as explicit causes from the highlighted sentences.

6. An emergency triage question-answering system based on knowledge graph according to claim 5, characterized in that: In S2, the matching diagnosis results are all the diagnosis results corresponding to this combination of explicit causes in the knowledge graph branch after the patient has added other symptoms; the method for obtaining instrument examination suggestions is: list all the explicit cause combinations that contain this explicit cause combination, and after finding all the corresponding diagnosis results in the knowledge graph branch, infer the results of all instrument examinations of symptoms when various diseases occur from the diagnosis results, and use these instrument examination items as instrument examination suggestions.

7. An emergency triage question-answering system based on knowledge graph according to claim 6, characterized in that: In S5, the specific method of constructing a new knowledge graph branch when the existing knowledge graph is incomplete is: combine all the explicit causes given by the doctor and the question-and-answer system when the patient visits the hospital, as well as all the implicit causes detected during the instrument examination, and then correspond them with the diagnosis results to construct a knowledge graph branch under this combination.

8. An emergency triage question-answering system based on knowledge graph according to claim 7, characterized in that: In S5, the specific method for modifying the knowledge graph is as follows: when the new knowledge graph branch is constructed, first search whether the diagnosis result corresponds to a combination of less explicit causes and less implicit causes. When there are fewer combinations, define the redundant explicit causes and implicit causes as complications, and count all subsequent cases with the diagnosis result of this disease, record the number of times without complications a1 and the number of times with complications a2, and get when The larger it is, the greater the probability that this complication symptom will occur together with the disease. When guiding the patient to further supplement other symptoms, this complication symptom will be presented to the patient for his or her judgment.

Images