A method, system and electronic device for guiding consultation for special diseases

Abstract

The present invention provides a consultation guidance method, system and electronic device for special diseases. The method includes the following steps: obtaining a medical case sample set corresponding to a target special disease, each medical case sample in the medical case sample set including prescription information and symptom information; constructing a multi-classification decision tree model that can reflect the mapping relationship between symptom sequence and prescription according to the medical case sample set; constructing a weighted undirected graph that can reflect the correlation between symptoms according to the multi-classification decision tree model; in response to the current patient symptoms input by the user, obtaining and providing the top N symptoms ranked in descending order of relevance to the current patient symptoms according to the weighted undirected graph. The system includes functional modules corresponding to the implementation of the above steps. The electronic device: the method is implemented when the processor executes the computer program stored in the memory. According to the present invention, the problem that the existing Chinese medicine consultation guidance method has poor flexibility due to the use of a predetermined fixed consultation process for consultation guidance can be solved.

Description

Technical Field

[0001] The present invention belongs to the field of TCM-assisted diagnosis and treatment, and more specifically, relates to a disease-specific consultation guidance method, system and electronic equipment. Background Art

[0002] In the field of traditional Chinese medicine, it is difficult for interns and assistant doctors who have been in the profession for a relatively short time to complete the consultation process with ease. Since they generally do not have the unique consultation process refined based on their own years of consultation experience like senior doctors with relatively long working time, most of them have problems such as lack of direction and unclear termination conditions during consultation. Among them, the problem of lack of direction in consultation has been solved to a certain extent by the consultation guidance technology that has emerged in recent years. However, the existing consultation guidance technology usually guides consultation according to the consultation process preset in the manual consultation form, which has significant disadvantages such as lack of flexibility and low efficiency. The static strategy is slightly cramped when facing complex and changeable real-world level problems, and doctor users will also feel bored when using the same consultation process when facing different patients. Summary of the invention

[0003] The purpose of the present invention is to solve the problem that the existing Chinese medicine consultation guidance method has poor flexibility due to the use of a predetermined fixed consultation process for consultation guidance.

[0004] In order to achieve the above-mentioned purpose, the present invention provides a method, system and electronic equipment for guiding consultation for specific diseases.

[0005] According to a first aspect of the present invention, a method for guiding consultation for a specific disease is provided, the method comprising the following steps:

[0006] Acquire a medical case sample set corresponding to a target disease, wherein each medical case sample in the medical case sample set includes prescription information and symptom information;

[0007] Constructing a multi-classification decision tree model that can reflect the mapping relationship between symptom sequence and prescription according to the medical case sample set;

[0008] Constructing a weighted undirected graph capable of reflecting the correlation between symptoms according to the multi-classification decision tree model;

[0009] In response to the current patient symptoms input by the user, the top N symptoms ranked in descending order of relevance to the current patient symptoms are acquired and provided according to the weighted undirected graph.

[0010] Optionally, obtaining a medical case sample set corresponding to a target disease includes:

[0011] Acquire a real medical case set corresponding to a target disease, each real medical case in the real medical case set including prescription information and symptom information;

[0012] Deduplication is performed on all symptoms included in the real medical case set, and the deduplication symptoms are divided according to their categories to obtain a corresponding number of symptom pools;

[0013] Determine the number of each category of symptoms in each real medical case to determine the range of the number of each category of symptoms in all real medical cases;

[0014] Determine the standard deviation and total value of the weight array corresponding to the symptom combination contained in each real medical case based on the prescription recommendation system, so as to determine the standard deviation range and total value range of the weight array corresponding to the symptom combination of all real medical cases;

[0015] Within the range of the number of symptoms of each category in all real medical records, a random number of symptoms are randomly selected from each symptom pool to obtain the symptom combination to be identified;

[0016] Determine whether the weight array corresponding to the symptom combination to be identified satisfies the standard deviation range and total value range of the weight array corresponding to the symptom combination of all real medical records at the same time; if so, determine whether the symptom combination has been obtained; if not, input the symptom combination into the prescription recommendation system;

[0017] The prescription output by the prescription recommendation system based on the input symptom combination is integrated with the symptom combination into a medical case sample.

[0018] Optionally, constructing a weighted undirected graph capable of reflecting the correlation between symptoms according to the multi-classification decision tree model includes:

[0019] Performing a breadth-first traversal of the multi-classification decision tree model starting from the root node of the multi-classification decision tree model, and taking each symptom node that appears for the first time as each node constituting an undirected graph;

[0020] Adding paths between nodes constituting an undirected graph according to the paths between the first-appearing symptom nodes in the multi-classification decision tree model to form an undirected graph;

[0021] The feature importance of each first-appearing symptom node in the multi-classification decision tree model is used as the weight of each corresponding node in the undirected graph to obtain a weighted undirected graph.

[0022] Optionally, in response to the current patient symptoms input by the user, the top N symptoms ranked in descending order of relevance to the current patient symptoms are obtained and provided according to the weighted undirected graph, including:

[0023] Determine whether the total number of the current patient's symptoms and the previous patient's symptoms is not greater than M. If so, determine the target symptom node in the weighted undirected graph that matches the current patient's symptoms and all the previous patient's symptoms. If not, determine the target symptom node in the weighted undirected graph that matches the current patient's symptoms and the M-1 previous patient's symptoms that are closest to the current patient's symptoms.

[0024] In the weighted undirected graph, determining all adjacent symptom nodes of each target symptom node;

[0025] All adjacent symptom nodes of each target symptom node are sorted in descending order of weight as candidate nodes. When calculating the weight of a candidate node, if a candidate node appears K times, its weight used for sorting is K times its own weight.

[0026] It is determined whether the number of candidate nodes after being sorted in descending order of weights is not less than N. If so, the symptoms corresponding to the top N candidate nodes are used as the top N symptoms ranked in descending order of relevance to the current patient's symptoms.

[0027] Optionally, the determining whether the number of candidate nodes after being sorted in descending order of weights is not less than N further includes:

[0028] If not, increase the number of candidate nodes participating in the descending weight sorting until the number is not less than N;

[0029] The symptoms corresponding to the top N candidate nodes are taken as the top N symptoms ranked in descending order of relevance to the current patient's symptoms;

[0030] The method of increasing the number of candidate nodes participating in the descending weight sorting is specifically:

[0031] For each target symptom node, the peripheral interval symptom node with a distance of 2 from the target symptom node in the weighted undirected graph is taken as a candidate node. If the number of candidate nodes participating in the descending weight sorting is still less than N at this time, the distance of the peripheral interval symptom node as the candidate node is increased by one until the number of candidate nodes participating in the descending weight sorting is not less than N.

[0032] Optionally, after acquiring and providing the top N symptoms ranked in descending order of relevance to the current patient's symptoms according to the weighted undirected graph, the method further includes:

[0033] The top N candidate nodes are deleted from the weighted undirected graph.

[0034] Optionally, after deleting the top N candidate nodes from the weighted undirected graph, the method further includes:

[0035] Determine whether to end the consultation guidance, and if so, end the consultation guidance.

[0036] Optionally, when any of the following conditions is met, the consultation guidance is terminated:

[0037] Condition 1: All symptom nodes in the weighted undirected graph are deleted;

[0038] Condition 2: The total value of the weight array of all patient symptoms currently recorded by the user is not less than the predetermined total value threshold and among the top N symptoms provided most recently in descending order of relevance to the current patient's symptoms, the weight of the first-ranked symptom is greater than the sum of the weights of the next two symptoms.

[0039] According to a second aspect of the present invention, a consultation guidance system for specific diseases is provided, and the consultation guidance system includes the following functional modules:

[0040] A medical case sample set acquisition module is used to acquire a medical case sample set corresponding to a target disease, wherein each medical case sample in the medical case sample set includes prescription information and symptom information;

[0041] A multi-classification decision tree model construction module is used to construct a multi-classification decision tree model that can reflect the mapping relationship between symptom sequence and prescription according to the medical case sample set;

[0042] A weighted undirected graph construction module, used to construct a weighted undirected graph capable of reflecting the correlation between various symptoms according to the multi-classification decision tree model;

[0043] The recommended symptom providing module is used to respond to the current patient symptoms input by the user, and to obtain and provide the top N symptoms ranked in descending order of relevance to the current patient symptoms according to the weighted undirected graph.

[0044] According to a third aspect of the present invention, there is provided an electronic device comprising a processor and a memory, wherein the processor implements any one of the above-mentioned consultation guidance methods for specific diseases when executing a computer program stored in the memory.

[0045] The beneficial effects of the present invention are:

[0046] The consultation guidance method for specific diseases of the present invention first obtains a medical case sample set corresponding to a target specific disease, wherein each medical case sample in the medical case sample set includes prescription information and symptom information; secondly, a multi-classification decision tree model capable of reflecting the mapping relationship between symptom sequence and prescription is constructed based on the medical case sample set; thirdly, a weighted undirected graph capable of reflecting the correlation between symptoms is constructed based on the multi-classification decision tree model; finally, in response to the current patient symptoms input by a user, the top N symptoms ranked in descending order of relevance to the current patient symptoms are obtained and provided based on the weighted undirected graph.

[0047] By adopting the disease-specific consultation guidance method of the present invention, in the consultation scenario of the target disease, when the physician user inputs the recorded current patient symptoms, the top N symptoms ranked in descending order of relevance to the current patient symptoms can be obtained and provided for selection by the physician user based on the weighted undirected graph, thereby achieving flexible consultation guidance that is closer to the patient's actual situation, thereby effectively solving the problem of poor flexibility of the existing TCM consultation guidance method due to the use of a predetermined fixed consultation process for consultation guidance.

[0048] The disease-specific consultation guidance system and electronic device of the present invention belong to a general inventive concept as well as the disease-specific consultation guidance method mentioned above, and have at least the same beneficial effects as the disease-specific consultation guidance method mentioned above, and the beneficial effects will not be repeated here.

[0049] Other features and advantages of the present invention will be described in detail in the following detailed description. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] The present invention may be better understood by referring to the following description taken in conjunction with the accompanying drawings, in which the same or similar reference numerals are used throughout the drawings to represent the same or similar components.

[0051] Figure 1 A flowchart of the implementation of the consultation guidance method for a specific disease according to an embodiment of the present invention is shown;

[0052] Figure 2 A structural block diagram of a consultation guidance system for specific diseases according to an embodiment of the present invention is shown. DETAILED DESCRIPTION

[0053] In order to enable those skilled in the art to more fully understand the technical solution of the present invention, the exemplary embodiments of the present invention will be described in more comprehensive and detailed in conjunction with the accompanying drawings below. Obviously, the one or more embodiments of the present invention described below are only one or more of the specific ways in which the technical solution of the present invention can be implemented, and are not exhaustive. It should be understood that other ways belonging to a general inventive concept can be used to implement the technical solution of the present invention, and should not be limited by the exemplary embodiments described. Based on one or more embodiments of the present invention, all other embodiments obtained by ordinary technicians in the field without making creative work should fall within the scope of protection of the present invention.

[0054] Example: Figure 1 The flowchart of the implementation of the method for guiding consultation for a specific disease according to an embodiment of the present invention is shown. Figure 1 The method for guiding consultation for a specific disease according to an embodiment of the present invention comprises the following steps:

[0055] Step S100, obtaining a medical case sample set corresponding to a target disease, wherein each medical case sample in the medical case sample set includes prescription information and symptom information;

[0056] Step S200: constructing a multi-classification decision tree model that can reflect the mapping relationship between symptom sequence and prescription according to the medical case sample set;

[0057] Step S300, constructing a weighted undirected graph capable of reflecting the correlation between symptoms according to the multi-classification decision tree model;

[0058] Step S400: In response to the current patient symptoms input by the user, the top N symptoms ranked in descending order of relevance to the current patient symptoms are acquired and provided according to the weighted undirected graph.

[0059] Specifically, in the embodiment of the present invention, the correlation with the current patient's symptoms refers to the degree of association with the current patient's symptoms in reaching the final prescription recommendation.

[0060] Furthermore, the step S100 of the embodiment of the present invention of obtaining a medical record sample set corresponding to a target disease includes:

[0061] Acquire a real medical case set corresponding to a target disease, each real medical case in the real medical case set including prescription information and symptom information;

[0062] Deduplication is performed on all symptoms included in the real medical case set, and the deduplication symptoms are divided according to their categories to obtain a corresponding number of symptom pools;

[0063] Determine the number of each category of symptoms in each real medical case to determine the range of the number of each category of symptoms in all real medical cases;

[0064] Determine the standard deviation and total value of the weight array corresponding to the symptom combination contained in each real medical case based on the prescription recommendation system, so as to determine the standard deviation range and total value range of the weight array corresponding to the symptom combination of all real medical cases;

[0065] Within the range of the number of symptoms of each category in all real medical records, a random number of symptoms are randomly selected from each symptom pool to obtain the symptom combination to be identified;

[0066] Determine whether the weight array corresponding to the symptom combination to be identified satisfies the standard deviation range and total value range of the weight array corresponding to the symptom combination of all real medical records at the same time; if so, determine whether the symptom combination has been obtained; if not, input the symptom combination into the prescription recommendation system;

[0067] The prescription output by the prescription recommendation system based on the input symptom combination is integrated with the symptom combination into a medical case sample.

[0068] Specifically, in the embodiment of the present invention, the standard deviation and total value of the weight array corresponding to the symptom combination included in each real medical case determined by the prescription recommendation system include:

[0069] Based on the prescription recommendation system, the weight array corresponding to the symptom combination contained in each real medical case is determined.

[0070] The corresponding standard deviation and total value are determined according to the weight array corresponding to the symptom combination contained in each real medical case.

[0071] Specifically, in the embodiment of the present invention, the prescription recommendation system as the front-end system is an expert system for recommending prescriptions based on a symptom weight scoring mechanism. For the currently input symptom, if it hits the applicable symptom of a prescription under the target disease, the prescription can be scored, and the prescription with the highest score will be regarded as the actual prescription. In actual application, other prescription recommendation systems with similar implementation logic can also be used, such as a prescription recommendation system based on a neural network.

[0072] Specifically, in the embodiment of the present invention, the randomly obtained symptom combination used to form the medical case sample must meet four conditions, namely, the number range of each symptom category included matches the number range of each category symptom of all real medical cases; the standard deviation range of the corresponding weight array matches the standard deviation range of the weight array corresponding to the symptom combination of all real medical cases; the total value range of the corresponding weight array matches the total value range of the weight array corresponding to the symptom combination of all real medical cases; and it has not been randomly obtained before. Such a setting can make the generated medical case sample closer to the real medical case and non-repetitive.

[0073] Furthermore, in step S200 of the embodiment of the present invention, constructing a weighted undirected graph capable of reflecting the correlation between the symptoms according to the multi-classification decision tree model includes:

[0074] Performing a breadth-first traversal of the multi-classification decision tree model starting from the root node of the multi-classification decision tree model, and taking each symptom node that appears for the first time as each node constituting an undirected graph;

[0075] Adding paths between nodes constituting an undirected graph according to the paths between the first-appearing symptom nodes in the multi-classification decision tree model to form an undirected graph;

[0076] The feature importance of each first-appearing symptom node in the multi-classification decision tree model is used as the weight of each corresponding node in the undirected graph to obtain a weighted undirected graph.

[0077] Specifically, in an embodiment of the present invention, the topological structure of the weighted undirected graph can characterize the internal correlation of the symptom set for the target disease, and the weight of each symptom node can characterize the importance of the disease for achieving the final prescription recommendation.

[0078] Furthermore, in step S400 of the embodiment of the present invention, in response to the current patient symptoms input by the user, the top N symptoms ranked in descending order of relevance to the current patient symptoms are obtained and provided according to the weighted undirected graph, including:

[0079] Determine whether the total number of the current patient's symptoms and the previous patient's symptoms is not greater than M. If so, determine the target symptom node in the weighted undirected graph that matches the current patient's symptoms and all the previous patient's symptoms. If not, determine the target symptom node in the weighted undirected graph that matches the current patient's symptoms and the M-1 previous patient's symptoms that are closest to the current patient's symptoms.

[0080] In the weighted undirected graph, determining all adjacent symptom nodes of each target symptom node;

[0081] All adjacent symptom nodes of each target symptom node are sorted in descending order of weight as candidate nodes. When calculating the weight of a candidate node, if a candidate node appears K times, its weight used for sorting is K times its own weight.

[0082] It is determined whether the number of candidate nodes after being sorted in descending order of weights is not less than N. If so, the symptoms corresponding to the top N candidate nodes are used as the top N symptoms ranked in descending order of relevance to the current patient's symptoms.

[0083] Specifically, in the embodiment of the present invention, if there is a first-level symptom in the symptoms corresponding to the first N symptom nodes and the second-level symptoms of the first-level symptom have been provided in the previous symptom provision link, the symptom node corresponding to the first-level symptom is removed and the symptom node is filled;

[0084] If there is a weak symptom among the symptoms corresponding to the first N symptom nodes, and the strong symptom corresponding to the weak symptom has been provided in the previous symptom provision link, then the symptom node corresponding to the weak symptom is removed and the symptom node is filled in;

[0085] Among them, secondary symptoms are sub-symptoms of primary symptoms, and strong symptoms and weak symptoms are two manifestations of different severity of the same symptom.

[0086] Specifically, in the embodiment of the present invention, the previous patient symptoms are patient symptoms acquired before the current patient symptoms;

[0087] M is preferably 5.

[0088] Furthermore, in the embodiment of the present invention, the determining whether the number of candidate nodes after being sorted in descending order of weights is not less than N further includes:

[0089] If not, increase the number of candidate nodes participating in the descending weight sorting until the number is not less than N;

[0090] The symptoms corresponding to the top N candidate nodes are taken as the top N symptoms ranked in descending order of relevance to the current patient's symptoms;

[0091] The method of increasing the number of candidate nodes participating in the descending weight sorting is specifically:

[0092] For each target symptom node, the peripheral interval symptom node with a distance of 2 from the target symptom node in the weighted undirected graph is taken as a candidate node. If the number of candidate nodes participating in the descending weight sorting is still less than N at this time, the distance of the peripheral interval symptom node as the candidate node is increased by one until the number of candidate nodes participating in the descending weight sorting is not less than N.

[0093] Furthermore, in an embodiment of the present invention, after obtaining and providing the top N symptoms ranked in descending order of relevance to the current patient's symptoms according to the weighted undirected graph, the method further includes:

[0094] The top N candidate nodes are deleted from the weighted undirected graph.

[0095] Furthermore, in an embodiment of the present invention, after deleting the top N candidate nodes from the weighted undirected graph, the method further includes:

[0096] Determine whether to end the consultation guidance, and if so, end the consultation guidance.

[0097] Furthermore, in the embodiment of the present invention, when any of the following conditions is met, it is determined that the consultation guidance is terminated:

[0098] Condition 1: All symptom nodes in the weighted undirected graph are deleted;

[0099] Condition 2: The total value of the weight array of all patient symptoms currently recorded by the user is not less than the predetermined total value threshold and among the top N symptoms provided most recently in descending order of relevance to the current patient's symptoms, the weight of the first-ranked symptom is greater than the sum of the weights of the next two symptoms.

[0100] Accordingly, based on the disease-specific consultation guidance method of the embodiment of the present invention, the embodiment of the present invention also proposes a disease-specific consultation guidance system.

[0101] Figure 2FIG. 1 is a block diagram showing a structure of a disease-specific consultation guidance system according to an embodiment of the present invention. Figure 2 The disease-specific consultation guidance system of the embodiment of the present invention includes the following functional modules:

[0102] A medical case sample set acquisition module is used to acquire a medical case sample set corresponding to a target disease, wherein each medical case sample in the medical case sample set includes prescription information and symptom information;

[0103] A multi-classification decision tree model construction module is used to construct a multi-classification decision tree model that can reflect the mapping relationship between symptom sequence and prescription according to the medical case sample set;

[0104] A weighted undirected graph construction module, used to construct a weighted undirected graph capable of reflecting the correlation between various symptoms according to the multi-classification decision tree model;

[0105] The recommended symptom providing module is used to respond to the current patient symptoms input by the user, and to obtain and provide the top N symptoms ranked in descending order of relevance to the current patient symptoms according to the weighted undirected graph.

[0106] Accordingly, based on the consultation guidance method for specific diseases in the embodiment of the present invention, the embodiment of the present invention also proposes an electronic device, which includes a processor and a memory. When the processor executes the computer program stored in the memory, the consultation guidance method for specific diseases proposed in the embodiment of the present invention is implemented.

[0107] Although one or more embodiments of the present invention are described above, it should be known to those skilled in the art that the present invention can be implemented in any other form without departing from its subject matter and scope. Therefore, the embodiments described above are illustrative rather than restrictive, and many modifications and substitutions are obvious to those skilled in the art without departing from the spirit and scope of the present invention as defined in the appended claims.

Claims

1. A method for guiding consultation for a specific disease, characterized in that: include: Acquire a medical case sample set corresponding to a target disease, wherein each medical case sample in the medical case sample set includes prescription information and symptom information; Constructing a multi-classification decision tree model that can reflect the mapping relationship between symptom sequence and prescription according to the medical case sample set; Constructing a weighted undirected graph capable of reflecting the correlation between symptoms according to the multi-classification decision tree model; In response to the current patient symptoms input by the user, the top N symptoms ranked in descending order of relevance to the current patient symptoms are acquired and provided according to the weighted undirected graph; The obtaining of a medical case sample set corresponding to a target disease includes: Acquire a real medical case set corresponding to a target disease, each real medical case in the real medical case set including prescription information and symptom information; Deduplication is performed on all symptoms included in the real medical case set, and the deduplication symptoms are divided according to their categories to obtain a corresponding number of symptom pools; Determine the number of each category of symptoms in each real medical case to determine the range of the number of each category of symptoms in all real medical cases; Determine the standard deviation and total value of the weight array corresponding to the symptom combination contained in each real medical case based on the prescription recommendation system, so as to determine the standard deviation range and total value range of the weight array corresponding to the symptom combination of all real medical cases; Within the range of the number of symptoms of each category in all real medical records, a random number of symptoms are randomly selected from each symptom pool to obtain the symptom combination to be identified; Determine whether the weight array corresponding to the symptom combination to be identified satisfies the standard deviation range and total value range of the weight array corresponding to the symptom combination of all real medical records at the same time; if so, determine whether the symptom combination has been obtained; if not, input the symptom combination into the prescription recommendation system; The prescription output by the prescription recommendation system based on the input symptom combination is integrated with the symptom combination into a medical case sample.

2. The method for guiding consultation for a specific disease according to claim 1, characterized in that: The method of constructing a weighted undirected graph capable of reflecting the correlation between symptoms according to the multi-classification decision tree model includes: Performing a breadth-first traversal of the multi-classification decision tree model starting from the root node of the multi-classification decision tree model, and taking each symptom node that appears for the first time as each node constituting an undirected graph; Adding paths between nodes constituting an undirected graph according to the paths between the first-appearing symptom nodes in the multi-classification decision tree model to form an undirected graph; The feature importance of each first-appearing symptom node in the multi-classification decision tree model is used as the weight of each corresponding node in the undirected graph to obtain a weighted undirected graph.

3. The method for guiding consultation for a specific disease according to claim 2, characterized in that: In response to the current patient symptoms input by the user, the top N symptoms ranked in descending order of relevance to the current patient symptoms are obtained and provided according to the weighted undirected graph, including: Determine whether the total number of the current patient's symptoms and the previous patient's symptoms is not greater than M. If so, determine the target symptom node in the weighted undirected graph that matches the current patient's symptoms and all the previous patient's symptoms. If not, determine the target symptom node in the weighted undirected graph that matches the current patient's symptoms and the M-1 previous patient's symptoms that are closest to the current patient's symptoms. In the weighted undirected graph, determining all adjacent symptom nodes of each target symptom node; All adjacent symptom nodes of each target symptom node are sorted in descending order of weight as candidate nodes. When calculating the weight of a candidate node, if a candidate node appears K times, its weight used for sorting is K times its own weight. It is determined whether the number of candidate nodes after being sorted in descending order of weights is not less than N. If so, the symptoms corresponding to the top N candidate nodes are used as the top N symptoms ranked in descending order of relevance to the current patient's symptoms.

4. The method for guiding consultation for a specific disease according to claim 3, characterized in that: The step of judging whether the number of candidate nodes sorted in descending order of weight is not less than N further includes: If not, increase the number of candidate nodes participating in the descending weight sorting until the number is not less than N; The symptoms corresponding to the top N candidate nodes are taken as the top N symptoms ranked in descending order of relevance to the current patient's symptoms; The method of increasing the number of candidate nodes participating in the descending weight sorting is specifically: For each target symptom node, the peripheral interval symptom node with a distance of 2 from the target symptom node in the weighted undirected graph is taken as a candidate node. If the number of candidate nodes participating in the descending weight sorting is still less than N at this time, the distance of the peripheral interval symptom node as the candidate node is increased by one until the number of candidate nodes participating in the descending weight sorting is not less than N.

5. The method for guiding consultation for a specific disease according to claim 4, characterized in that: After obtaining and providing the top N symptoms ranked in descending order of relevance to the current patient's symptoms according to the weighted undirected graph, the method further includes: The top N candidate nodes are deleted from the weighted undirected graph.

6. The method for guiding consultation for a specific disease according to claim 5, characterized in that: After deleting the top N candidate nodes from the weighted undirected graph, the method further includes: Determine whether to end the consultation guidance, and if so, end the consultation guidance.

7. The method for guiding consultation for a specific disease according to claim 6, characterized in that: When any of the following conditions is met, the consultation guidance is terminated: Condition 1: All symptom nodes in the weighted undirected graph are deleted; Condition 2: The total value of the weight array of all patient symptoms currently recorded by the user is not less than the predetermined total value threshold and among the top N symptoms provided most recently in descending order of relevance to the current patient's symptoms, the weight of the first-ranked symptom is greater than the sum of the weights of the next two symptoms.

8. A consultation guidance system for special diseases, characterized in that: include: A medical case sample set acquisition module is used to acquire a medical case sample set corresponding to a target disease, wherein each medical case sample in the medical case sample set includes prescription information and symptom information; A multi-classification decision tree model construction module is used to construct a multi-classification decision tree model that can reflect the mapping relationship between symptom sequence and prescription according to the medical case sample set; A weighted undirected graph construction module, used to construct a weighted undirected graph capable of reflecting the correlation between various symptoms according to the multi-classification decision tree model; A recommended symptom providing module, configured to, in response to the current patient symptoms input by the user, obtain and provide the top N symptoms ranked in descending order of relevance to the current patient symptoms according to the weighted undirected graph; The obtaining of a medical case sample set corresponding to a target disease includes: Acquire a real medical case set corresponding to a target disease, each real medical case in the real medical case set including prescription information and symptom information; Deduplication is performed on all symptoms included in the real medical case set, and the deduplication symptoms are divided according to their categories to obtain a corresponding number of symptom pools; Determine the number of each category of symptoms in each real medical case to determine the range of the number of each category of symptoms in all real medical cases; Determine the standard deviation and total value of the weight array corresponding to the symptom combination contained in each real medical case based on the prescription recommendation system, so as to determine the standard deviation range and total value range of the weight array corresponding to the symptom combination of all real medical cases; Within the range of the number of symptoms of each category in all real medical records, a random number of symptoms are randomly selected from each symptom pool to obtain the symptom combination to be identified; Determine whether the weight array corresponding to the symptom combination to be identified satisfies the standard deviation range and total value range of the weight array corresponding to the symptom combination of all real medical records at the same time; if so, determine whether the symptom combination has been obtained; if not, input the symptom combination into the prescription recommendation system; The prescription output by the prescription recommendation system based on the input symptom combination is integrated with the symptom combination into a medical case sample.

9. An electronic device, characterized in that: It comprises a processor and a memory, and when the processor executes the computer program stored in the memory, it implements the consultation guidance method for specific diseases as described in any one of claims 1 to 7.

Images