A method for calculating consultation priority based on TCM disease relationship knowledge graph

Abstract

The present invention provides a method for calculating consultation priority based on a knowledge graph of TCM symptom relationships. By constructing a knowledge graph of TCM symptom relationships, utilizing the correlation between symptoms, between symptoms and diseases, and between disease nature and disease, combined with the user's symptom information, the method automatically calculates and determines the priority of consultation. By prioritizing the highest-priority symptoms, key information can be quickly obtained, thereby improving the efficiency and accuracy of TCM diagnosis, reducing the number of consultations for patients, improving patient experience, and providing patients with more timely and effective medical services. It can avoid omissions in inquiries due to lack of experience or subjective factors of doctors, realize unmanned intelligent consultation, ensure the standardization and normalization of the consultation process, and improve the quality of medical services.

Description

Technical Field

[0001] The present invention relates to the field of intelligent technology of traditional Chinese medicine, and in particular to a method for calculating consultation priority based on a knowledge graph of traditional Chinese medicine symptom relationships. Background Art

[0002] In today's medical diagnosis field, Traditional Chinese Medicine (TCM), as the traditional Chinese medical system, continues to play a vital role in healthcare, leveraging its unique theoretical framework and rich clinical experience. During a TCM diagnosis, doctors use the patient's initial description of symptoms to determine the order of questioning and ensure a correct diagnosis.

[0003] Currently, Traditional Chinese Medicine (TCM) diagnoses rely primarily on the physician's personal experience and expertise, lacking systematic and standardized diagnostic tools. This, to a certain extent, impacts diagnostic accuracy and efficiency. Furthermore, with the increasing number of patients, physicians are faced with the immense amount of symptom information they need to process when making diagnoses. This not only increases their workload but can also prevent some patients in urgent need of treatment from receiving timely diagnosis and treatment.

[0004] Therefore, how to use modern scientific and technological means to improve the efficiency and accuracy of traditional Chinese medicine diagnosis has become an urgent problem to be solved. Summary of the Invention

[0005] To solve the problems existing in the prior art, the present invention provides a method for calculating consultation priority based on a knowledge graph of TCM disease relationships, comprising the following steps:

[0006] A knowledge graph of TCM disease relationships is pre-built, with symptoms, pathologies, and diseases as entities, and relationships between symptoms, between symptoms and diseases, and between pathologies and diseases as inter-entity relationships. Inter-entity relationships are represented by weight values.

[0007] Initialize the priority values ​​of all symptoms to 0, initialize the weight values ​​between all symptoms and diseases, the weight values ​​between symptoms, and the weight values ​​between pathologies and diseases, and set the search path length between symptoms and diseases and the search path length between symptoms;

[0008] Receive symptom information currently input by the user and determine the nature of the user's illness, wherein the symptom information includes symptoms selected by the user and symptoms not selected by the user;

[0009] Calculate the priority of each symptom based on the TCM disease relationship knowledge graph and the received user symptom information and disease nature;

[0010] Make the highest priority symptoms the ones your doctor asks about first during your next visit.

[0011] Optionally, the step of calculating the priority of each symptom based on the TCM disease relationship knowledge graph and the received user symptom information and disease type includes:

[0012] Obtain the weight values ​​between the received user symptoms and diseases and the weight values ​​between the user's pathological nature and diseases from the TCM disease relationship knowledge graph, and calculate the probability value of each disease based on the weight values ​​between the received user symptoms and diseases and the weight values ​​between the user's pathological nature and diseases;

[0013] The priority of each symptom is calculated based on the probability value of each disease, the initial weight values ​​between the symptom and the disease, and the initial weight values ​​between the symptoms.

[0014] Furthermore, the method of obtaining the weight values ​​between the received user symptoms and diseases and the weight values ​​between the user's pathological nature and the diseases from the TCM disease relationship knowledge graph, and calculating the probability value of each disease based on the weight values ​​between the received user symptoms and diseases and the weight values ​​between the user's pathological nature and the diseases, includes:

[0015] Get all user-selected symptoms and these symptoms belong to the disease i;

[0016] Get all symptoms not selected by the user and these symptoms belong to the disease i;

[0017] Get symptoms and the weight between disease i Sum these weight values;

[0018] Get symptoms and the weight between disease i Sum these weight values;

[0019] Get the weight values ​​w of all symptoms of disease i ds , calculate the sum of these weight values;

[0020] Get the weight value between the user's disease type and disease i Sum these weight values;

[0021] Get the weight value w of all pathogenicity dn , calculate the sum of these weight values;

[0022] Calculate the probability value P of disease i i , the specific formula is as follows:

[0023]

[0024] Among them, λ1 and λ2 are pre-set specific gravity values.

[0025] Furthermore, the priority of each symptom is calculated based on the probability value of each disease, the initial weight value between the symptom and the disease, and the initial weight value between the symptoms, including:

[0026] Traverse the symptoms k except those selected and unselected by the user, and obtain the initial weight values ​​between these symptoms k and disease i, and the initial weight values ​​between symptom k and other symptoms;

[0027] Calculate the latest weight value between symptom k and disease i, and update the initial weight value to the latest weight value. The specific formula is as follows:

[0028]

[0029] in, is the initial weight value between symptom k and disease i, P i is the probability value of disease i;

[0030] Calculate the latest weight value between symptom k and other symptoms, and update the initial weight value to the latest weight value. The specific formula is as follows:

[0031]

[0032] Among them, μ is a variable coefficient, which can be set according to needs. is the initial weight value between symptom k and other symptoms;

[0033] Calculate the priority of symptom k j , the specific formula is as follows:

[0034]

[0035] in, and are the search path lengths between symptoms and diseases and the search path lengths between symptoms, respectively. The lengths can be set according to actual conditions. l is the number of diseases adjacent to symptom k. is the latest weight value between the disease adjacent to symptom k and symptom k, m is the number of symptoms selected by users adjacent to symptom k, The latest weight value between the symptoms selected by the users adjacent to symptom k and symptom k.

[0036] After adopting the above technical solution, the present invention has the following beneficial effects:

[0037] 1. Improve the accuracy of medical consultation

[0038] The present invention is based on the knowledge graph of TCM disease relationships. It uses the correlation between symptoms, between symptoms and diseases, and between disease nature and disease to accurately calculate the priority of other symptoms. By first asking about the symptoms with the highest priority, key information can be quickly obtained, thereby improving the accuracy of diagnosis.

[0039] 2. Reduce the number of consultation rounds

[0040] The present invention can effectively reduce the number of consultation rounds through scientific calculation of symptom priority. Traditional intelligent Chinese medicine consultation methods often adopt the method of filling out a questionnaire. Users need to fill in dozens or even hundreds of questions to complete the collection of symptoms. This solution uses the symptom priority calculation method to lock in key symptoms in a shorter time, reduce the number of patient consultations, and improve patient experience.

[0041] 3. Standardized consultation process

[0042] The present invention provides a standardized symptom inquiry process, combined with a priority calculation method, which can avoid omissions in inquiries due to doctors' lack of experience or subjective factors, ensure the standardization and normalization of the consultation process, and improve the quality of medical services.

[0043] 4. Dynamically adjust weights

[0044] In the present invention, the weight values ​​between symptoms and diseases, and between symptoms are dynamically calculated and can be adjusted in real time according to the current symptom selection status and the probability value of the disease. This can more flexibly reflect the changes in the actual condition and improve the sensitivity and applicability of the medical consultation.

[0045] 5. Comprehensive consideration of symptoms and disease

[0046] In the calculation of disease probability, the present invention not only considers the relationship between symptoms and disease, but also considers the impact of the disease nature on disease probability. Combining these two aspects of information can more comprehensively evaluate the possibility of disease and improve the comprehensiveness and accuracy of diagnosis. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0048] Figure 1 A flowchart of a method for calculating consultation priority based on a knowledge graph of TCM disease relationships is provided for an embodiment of the present disclosure. DETAILED DESCRIPTION

[0049] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0050] In recent years, knowledge graphs, an emerging artificial intelligence technology, have shown promising application prospects in fields such as healthcare. Knowledge graphs can systematically organize and represent knowledge, effectively helping doctors quickly find relevant diagnostic evidence within vast amounts of medical information. However, within Traditional Chinese Medicine (TCM), research on symptom analysis and diagnosis using knowledge graphs is still in its early stages, and a mature application system has yet to be established.

[0051] Based on the above background, the inventors proposed a method for calculating consultation priority based on the knowledge graph of TCM disease relationships. By constructing a knowledge graph of TCM disease relationships and combining it with the user's symptom information, the priority of consultations is automatically calculated and determined, thereby improving the efficiency and accuracy of TCM diagnosis and providing patients with more timely and effective medical services. At the same time, this method can reduce the workload of doctors and provide strong support for the standardization and systematization of TCM diagnosis.

[0052] refer to Figure 1 The present disclosure provides a method for calculating consultation priority based on a knowledge graph of TCM disease relationships, comprising the following steps:

[0053] 1. Construct a knowledge graph of TCM disease relationships

[0054] A knowledge graph of TCM disease relationships is pre-built, with symptoms, pathologies, and diseases as entities, and relationships between symptoms, between symptoms and diseases, and between pathologies and diseases as inter-entity relationships. Inter-entity relationships are represented by weight values.

[0055] 2. Initialization phase

[0056] Initialize the priority values ​​of all symptoms to 0. This is done to uniformly process them during the calculation process and avoid interference from initial values.

[0057] Initialize the weights between all symptoms and diseases, between symptoms, and between disease types and diseases. These weights are derived from the attribute values ​​set based on clinical data when constructing the TCM disease relationship knowledge graph. For example, the initial weight of general symptoms is set to 1, while the weight of certain symptoms that are more important in specific diseases can be set to 5.

[0058] Set the search path length between symptoms and diseases and the search path length between symptoms to ensure that the association calculation between symptoms and diseases is within a reasonable range;

[0059] 3. User symptom input

[0060] Receive the symptom information currently input by the user and determine the type of the user's illness, wherein the symptom information includes symptoms selected by the user and symptoms not selected by the user. For example, the user selects the symptom "low back pain" but does not select the symptom "headache". The types of illness include deficiency, excess, cold, heat, and normal.

[0061] 4. Disease Probability Calculation

[0062] Get all user-selected symptoms and these symptoms belong to the disease i;

[0063] Get all symptoms not selected by the user and these symptoms belong to the disease i;

[0064] Get symptoms and the weight between disease i Sum these weight values. If the initial weight value is between 0 and 1, including 0, then set the weight value to 1;

[0065] Get symptoms and the weight between disease i Sum these weight values. If the initial weight value is between 0 and 1, excluding 0, then set the weight value to 0.

[0066] Get the weight values ​​w of all symptoms of disease i ds , calculate the sum of these weight values;

[0067] Get the weight value between the user's disease type and disease i Sum these weight values;

[0068] Get the weight value w of all pathogenicity dn , calculate the sum of these weight values;

[0069] Calculate the probability value P of disease i i , the specific formula is as follows:

[0070]

[0071] Among them, λ1 and λ2 are pre-set specific gravity values, λ1 is 0.9, and λ2 is 0.1;

[0072] 5. Symptom priority calculation

[0073] Traverse the symptoms k except those selected or not selected by the user, and obtain the initial weight values ​​between these symptoms k and disease i, and the initial weight values ​​between symptom k and other symptoms, for example, the symptom of "weak waist and knees" and its related weight values;

[0074] Calculate the latest weight value between symptom k and disease i, and update the initial weight value to the latest weight value. The specific formula is as follows:

[0075]

[0076] in, is the initial weight value between symptom k and disease i, P i is the probability value of disease i. For example, the probability value of the disease "Kidney Yin Deficiency" calculated in the previous step is 0.8, the probability value of "Xiao Ke" is 0.5, and the initial weight value between the symptom "Weakness of the Waist and Knees" and the diseases "Kidney Yin Deficiency" and "Xiao Ke" is 1. Then, through the above calculation, the latest weight value between the symptom "Weakness of the Waist and Knees" and the disease "Kidney Yin Deficiency" is 0.512, which is significantly higher than the latest weight value of 0.125 between the symptom "Weakness of the Waist and Knees" and the disease "Xiao Ke";

[0077] Calculate the latest weight value between symptom k and other symptoms, and update the initial weight value to the latest weight value. The specific formula is as follows:

[0078]

[0079] Among them, μ is a variable coefficient, which can be set according to needs. is the initial weight value between symptom k and other symptoms;

[0080] Calculate the priority of symptom k j , the specific formula is as follows:

[0081]

[0082] in, and are the search path lengths between symptoms and diseases and the search path lengths between symptoms, respectively. The lengths can be set according to actual conditions. l is the number of diseases adjacent to symptom k. is the latest weight value between the disease adjacent to symptom k and symptom k, m is the number of symptoms selected by users adjacent to symptom k, The latest weight value between the symptom selected by the user adjacent to symptom k and symptom k;

[0083] Through the above steps, the priority of all symptoms can be calculated, and the symptom with the highest priority will be used as the symptom that the doctor will ask about first during the next consultation. For example, if the calculation result shows that the symptom of "weak waist and knees" has the highest priority, the symptom to be asked about next time will be "weak waist and knees".

[0084] Through the above implementation, the present invention effectively uses the TCM disease relationship knowledge graph to calculate the priority of symptoms. For example, when the user inputs the symptom as "low back pain", through the above steps, the symptom with the highest priority may be "weakness in the waist and knees", and the doctor is advised to give priority to asking about the "weakness in the waist and knees" symptom during the next consultation.

[0085] Although the present invention has been disclosed above by way of embodiments, they are not intended to limit the present invention. Any person skilled in the art may make slight changes and modifications without departing from the spirit and scope of the present invention. Therefore, the scope of protection of the present invention shall be determined by the claims.

Claims

1. A method for calculating consultation priority based on a knowledge graph of TCM disease relationships, characterized in that: The following steps are involved: A knowledge graph of TCM disease relationships is pre-built, with symptoms, pathologies, and diseases as entities, and relationships between symptoms, between symptoms and diseases, and between pathologies and diseases as inter-entity relationships. Inter-entity relationships are represented by weight values. Initialize the priority values ​​of all symptoms to 0, initialize the weight values ​​between all symptoms and diseases, the weight values ​​between symptoms, and the weight values ​​between pathologies and diseases, and set the search path length between symptoms and diseases and the search path length between symptoms; Receive symptom information currently input by the user and determine the nature of the user's illness, wherein the symptom information includes symptoms selected by the user and symptoms not selected by the user; Calculate the priority of each symptom based on the TCM disease relationship knowledge graph and the received user symptom information and disease nature; Make the highest priority symptom the one your doctor asks about first during the next consultation; The step of calculating the priority of each symptom based on the TCM disease relationship knowledge graph and the received user symptom information and disease nature includes: Obtain the weight values ​​between the received user symptoms and diseases and the weight values ​​between the user's pathological nature and diseases from the TCM disease relationship knowledge graph, and calculate the probability value of each disease based on the weight values ​​between the received user symptoms and diseases and the weight values ​​between the user's pathological nature and diseases; Calculate the priority of each symptom based on the probability value of each disease, the initial weight value between the symptom and the disease, and the initial weight value between the symptoms; The method of obtaining the weight values ​​between the received user symptoms and diseases and the weight values ​​between the user's pathological nature and the diseases from the TCM disease relationship knowledge graph, and calculating the probability value of each disease based on the weight values ​​between the received user symptoms and diseases and the weight values ​​between the user's pathological nature and the diseases, includes: Get all user-selected symptoms and these symptoms belong to the disease i; Get all symptoms not selected by the user and these symptoms belong to the disease i; Get symptoms and the weight between disease i Sum these weight values; Get symptoms and the weight between disease i Sum these weight values; Get the weight values ​​w of all symptoms of disease i ds , calculate the sum of these weight values; Get the weight value between the user's disease type and disease i Sum these weight values; Get the weight value w of all pathogenicity dn , calculate the sum of these weight values; Calculate the probability value P of disease i i , the specific formula is as follows: Among them, λ1 and λ2 are pre-set specific gravity values; The step of calculating the priority of each symptom based on the probability value of each disease, the initial weight values ​​between the symptom and the disease, and the initial weight values ​​between the symptoms includes: Traverse the symptoms k except those selected and unselected by the user, and obtain the initial weight values ​​between these symptoms k and disease i, and the initial weight values ​​between symptom k and other symptoms; Calculate the latest weight value between symptom k and disease i, and update the initial weight value to the latest weight value. The specific formula is as follows: in, is the initial weight value between symptom k and disease i, P i is the probability value of disease i; Calculate the latest weight value between symptom k and other symptoms, and update the initial weight value to the latest weight value. The specific formula is as follows: Among them, μ is a variable coefficient, which can be set according to needs. is the initial weight value between symptom k and other symptoms; Calculate the priority of symptom k j , the specific formula is as follows: in, and are the search path lengths between symptoms and diseases and the search path lengths between symptoms, respectively. The lengths can be set according to actual conditions. l is the number of diseases adjacent to symptom k. is the latest weight value between the disease adjacent to symptom k and symptom k, m is the number of symptoms selected by users adjacent to symptom k, The latest weight value between the symptoms selected by the users adjacent to symptom k and symptom k.

Images