Inquiry processing method and device, equipment and storage medium

Abstract

This application provides a method, apparatus, device, and storage medium for online consultation processing. The method includes: receiving symptom information from a client; obtaining a set of candidate diseases corresponding to the symptom information based on a consultation knowledge graph; determining the score of each candidate disease in the candidate disease set through probability analysis; determining the diagnosis and treatment information corresponding to the symptom information collected up to the current round of inquiry based on the scores of multiple candidate diseases in the candidate disease set, wherein the diagnosis and treatment information includes at least one candidate disease with a high score; and finally determining whether to output the diagnosis and treatment information or proceed to the next round of inquiry based on a preset number of inquiries, the information entropy of at least one candidate disease in the diagnosis and treatment information, and any one of the factors in the diagnostic decision model. This approach, combined with a consultation knowledge graph, can effectively reduce the misdiagnosis rate of the consultation system and improve the accuracy of diagnosis and treatment; furthermore, by analyzing the information entropy of candidate diseases to determine whether to stop the inquiry, the inquiry time can be shortened, improving the user's inquiry experience.

Description

Technical Field

[0001] This application relates to the field of artificial intelligence technology, and in particular to a method, apparatus, device and storage medium for diagnosis processing. Background Technology

[0002] With the rapid development of information technology and the vigorous promotion of the "Internet+" concept, a new direction for reform in the medical industry has been explored. In the current context, the development of computer-aided diagnosis systems has become a research hotspot.

[0003] Consultation is a diagnostic method that involves asking patients or their companions for information related to the disease, including the patient's subjective symptoms, in order to understand the patient's various discomforts and the onset, development, and treatment of the disease. In the context of the "Internet+" development, users can interact with the consultation system through smart terminals to obtain preliminary diagnostic results and treatment suggestions from the system.

[0004] However, existing computer-aided diagnosis systems are still immature and lack objectivity, accuracy, and standardization. Summary of the Invention

[0005] This application provides a method, apparatus, device, and storage medium for medical history taking, which improves the accuracy and efficiency of diagnosis and treatment.

[0006] The first aspect of this application provides a medical history taking method, including:

[0007] Receive symptom information from a client, the symptom information including at least one symptom data input by the user;

[0008] Based on a preset medical history knowledge graph, a first set of candidate diseases and a first set of candidate symptoms corresponding to the symptom information are obtained. The first set of candidate diseases includes multiple candidate diseases, and the first set of candidate symptoms includes all candidate symptoms of the multiple candidate diseases.

[0009] The diagnostic and treatment information corresponding to the symptom information up to the current round of inquiry is determined from the first candidate disease set through probability analysis, and the diagnostic and treatment information is used to indicate at least one candidate disease;

[0010] Based on any one of the preset number of inquiries, the information entropy of at least one candidate disease in the diagnosis and treatment information, and the pre-trained diagnostic decision model, the diagnosis and treatment information is output or the next round of inquiries is initiated.

[0011] In an optional embodiment of this application, determining the diagnosis and treatment information corresponding to the symptom information up to the current round of inquiry from the first candidate disease set through probability analysis includes:

[0012] Obtain the score for each candidate disease in the first candidate disease set, and the score is used to indicate the probability value that the user has the candidate disease;

[0013] Based on the scores of the multiple candidate diseases in the first candidate disease set, the diagnosis and treatment information corresponding to the symptom information up to the current round of inquiry is determined. The diagnosis and treatment information includes a preset number of candidate diseases with scores from high to low among the multiple candidate diseases.

[0014] In an optional embodiment of this application, obtaining the score of each candidate disease in the first candidate disease set includes:

[0015] The contribution of each candidate symptom of the first candidate disease to the first candidate disease is obtained. The contribution is used to indicate the statistical probability value of the sample accompanied by the candidate symptom of the first candidate disease. The first candidate disease is any one of the plurality of candidate diseases.

[0016] The score of the first candidate disease is determined based on the contribution of all candidate symptoms to the first candidate disease.

[0017] In an optional embodiment of this application, determining the score of the first candidate disease based on the contribution of all candidate symptoms of the first candidate disease to the first candidate disease includes:

[0018] The score of the first candidate disease is determined based on the contribution of all candidate symptoms to the first candidate disease and the noise parameter.

[0019] In an optional embodiment of this application, determining whether to output the diagnostic information or proceed to the next round of inquiry based on the information entropy of at least one candidate disease in the diagnostic information includes:

[0020] If the sum of the information entropy of all candidate diseases in the at least one candidate disease is less than a preset threshold, the diagnosis and treatment information is determined to be output; or

[0021] If the sum of the information entropy of all candidate diseases in the at least one candidate disease is greater than or equal to the preset threshold, then proceed to the next round of inquiry.

[0022] In one optional embodiment of this application, determining whether to output the medical information or proceed to the next round of inquiries based on the preset number of inquiries includes:

[0023] Determine whether the current round of queries has reached the preset number of queries;

[0024] If the current round of inquiry reaches the preset number of inquiries, the diagnosis and treatment information is output; or, if the current round of inquiry does not reach the preset number of inquiries, the next round of inquiry is initiated.

[0025] In one optional embodiment of this application, determining whether to output the diagnostic information or proceed to the next round of inquiry based on the pre-trained diagnostic decision model includes:

[0026] The symptom information is input into the pre-trained diagnostic decision model, and the diagnostic decision model outputs the diagnostic information or proceeds to the next round of inquiry based on the output value of the model.

[0027] The diagnostic decision model is obtained by training a fully connected neural network using a reinforcement learning algorithm through multiple sample sequences. The sample sequences include at least one symptom data and a decision result, which is used to indicate the output of diagnostic information or to conduct the next round of inquiry.

[0028] In an optional embodiment of this application, determining whether to output the diagnostic information or proceed with the next round of inquiry based on the output value of the diagnostic decision model includes:

[0029] If the diagnostic decision model outputs a first value, the diagnostic information will be output; or

[0030] If the diagnostic decision model outputs a second value, it determines to proceed to the next round of inquiry.

[0031] In an optional embodiment of this application, the method further includes:

[0032] If it is determined to proceed to the next round of inquiry, the target inquiry symptom is determined from the first candidate symptom set, and inquiry information is sent to the client. The inquiry information is used to ask the user whether they have the target inquiry symptom.

[0033] Receive a response from the client, update the diagnosis and treatment information based on the response, and determine whether to output the updated diagnosis and treatment information or proceed with the next round of inquiries.

[0034] In an optional embodiment of this application, determining the target inquiry symptom from the first candidate symptom set includes at least one of the following:

[0035] From the candidate diseases with the highest scores in the first candidate disease set, select the detailed symptoms of at least one candidate symptom confirmed by the user in the current round of inquiry as the target inquiry symptom; or

[0036] From the candidate diseases with the highest scores in the first candidate disease set, select other symptoms besides those already confirmed by the user as the target query symptoms; or

[0037] The candidate symptom that causes the fastest decrease in the overall information entropy of the first candidate disease set is selected from all candidate symptoms of multiple candidate diseases in the first candidate disease set as the target query symptom.

[0038] In an optional embodiment of this application, the overall information entropy of the first candidate symptom in the first candidate symptom set to the first candidate disease set is determined based on the ratio of the sum of scores of all candidate diseases in the second candidate disease set in the next round of querying to the sum of scores of all candidate diseases in the first candidate disease set;

[0039] The second candidate disease set is determined based on the first candidate disease set and the first candidate symptom, wherein the first candidate symptom is any one of the candidate symptom sets in the first candidate symptom set.

[0040] In one optional embodiment of this application, the medical knowledge graph includes nodes for disease, symptoms, disease causes, disease department information, differential symptoms, complications, and disease course;

[0041] The step of obtaining the first candidate disease set corresponding to the symptom information based on the preset consultation knowledge graph includes: obtaining the candidate disease set corresponding to each symptom data from the consultation knowledge graph according to the at least one symptom data, and obtaining the first candidate disease set.

[0042] A second aspect of this application provides a medical consultation processing device, comprising:

[0043] A receiving module is used to receive symptom information from a client, the symptom information including at least one symptom data input by the user;

[0044] The acquisition module is used to acquire a first candidate disease set and a first candidate symptom set corresponding to the symptom information based on a preset medical history knowledge graph. The first candidate disease set includes multiple candidate diseases, and the first candidate symptom set includes all candidate symptoms of the multiple candidate diseases.

[0045] The processing module is used to determine the diagnosis and treatment information corresponding to the symptom information up to the current round of inquiry from the first candidate disease set through probability analysis, wherein the diagnosis and treatment information is used to indicate at least one candidate disease;

[0046] Based on any one of the preset number of inquiries, the information entropy of at least one candidate disease in the diagnosis and treatment information, and the pre-trained diagnostic decision model, the diagnosis and treatment information is output or the next round of inquiries is initiated.

[0047] A third aspect of this application provides an electronic device, including:

[0048] Memory;

[0049] Processor; and

[0050] Computer programs;

[0051] The computer program is stored in the memory and configured to be executed by the processor to implement the method as described in any one of the first aspects.

[0052] A fourth aspect of this application provides a computer-readable storage medium having a computer program stored thereon, the computer program being executed by a processor to implement the method as described in any one of the first aspects.

[0053] A fifth aspect of this application provides a computer program product, including a computer program that, when executed by a processor, implements the method described in any one of the first aspects.

[0054] This application provides a method, apparatus, device, and storage medium for online consultation processing. The method includes: receiving symptom information from a client; firstly, obtaining a set of candidate diseases corresponding to the symptom information based on a preset online consultation knowledge graph; determining the score of each candidate disease in the candidate disease set through probability analysis, with higher scores indicating a greater probability that the user has the candidate disease; then, determining the diagnostic information corresponding to the symptom information collected up to the current round of inquiry based on the scores of multiple candidate diseases in the candidate disease set, the diagnostic information including at least one candidate disease with a high score; and finally, determining whether to output the diagnostic information or proceed to the next round of inquiry based on a preset number of inquiries, the information entropy of at least one candidate disease in the diagnostic information, and any one of a pre-trained diagnostic decision model. On the one hand, the above solution, combined with an online consultation knowledge graph, can effectively reduce the misdiagnosis rate of the online consultation system and improve the accuracy of diagnosis; on the other hand, determining whether to stop the inquiry by analyzing the information entropy of candidate diseases in the diagnostic information can shorten the inquiry time and improve the user's inquiry experience. Attached Figure Description

[0055] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0056] Figure 1 A schematic diagram of a scenario for the consultation processing method provided in the embodiments of this application;

[0057] Figure 2 Interactive illustration of the consultation processing method provided in the embodiments of this application Figure 1 ;

[0058] Figure 3 Interactive illustration of the consultation processing method provided in the embodiments of this application Figure 2 ;

[0059] Figure 4 A schematic diagram of the structure of the medical consultation processing device provided in the embodiments of this application. Figure 1 ;

[0060] Figure 5 A schematic diagram of the structure of the medical consultation processing device provided in the embodiments of this application. Figure 2 ;

[0061] Figure 6 This is a hardware structure diagram of the electronic device provided in the embodiments of this application.

[0062] The accompanying drawings illustrate specific embodiments of this application, which will be described in more detail below. These drawings and descriptions are not intended to limit the scope of the concept in any way, but rather to illustrate the concept of this application to those skilled in the art through reference to particular embodiments. Detailed Implementation

[0063] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0064] The terms "first," "second," etc., used in the specification, claims, and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this application described herein can be implemented in a sequence other than those illustrated or described herein.

[0065] It should be understood that the terms “comprising” and “having” as used herein, and any variations thereof, are intended to cover non-exclusive inclusion, for example, a process, method, system, product, or apparatus that includes a series of steps or units is not necessarily limited to those steps or units that are expressly listed, but may include other steps or units that are not expressly listed or that are inherent to such process, method, product, or apparatus.

[0066] In the description of the embodiments of this application, the term "correspondence" may indicate that there is a direct or indirect correspondence between two things, or that there is an association between two things, or that there is a relationship of instruction and being instructed, configuration and being configured, etc.

[0067] First, a brief introduction to the relevant terms used in the embodiments of this application will be given.

[0068] Knowledge graphs are a modern theory that combines theories and methods from applied mathematics, computer graphics, information visualization, and information science with bibliometric citation analysis and co-occurrence analysis. They utilize visualized graphs to vividly represent the core structure, development history, cutting-edge fields, and overall knowledge architecture of a discipline, achieving multidisciplinary integration. Knowledge graphs can provide practical and valuable references for disciplinary research.

[0069] Named entities generally refer to entities in text that have specific meaning or strong referentiality. For example, entities in the medical field include personal names, department names, dates and times, and medical proper nouns (such as disease names, abbreviations, and treatment methods).

[0070] Named Entity Recognition (NER) extracts the aforementioned entities from unstructured input text and can identify more entities according to business needs, such as drug names, batch numbers, and prices in the medical field.

[0071] Relation extraction (RE) is used to determine whether a relationship exists between two entities in a sentence and what type of relationship it is. For example, in "I do not have a fever", "do not" is a negative word and "fever" is a symptom word. "Do not" and "fever" have a modification relationship.

[0072] Natural Language Processing (NLP) algorithms include syntax analysis, syntactic analysis, and semantic analysis. Syntactic analysis includes word segmentation, part-of-speech tagging, entity recognition, and spell checking. The basic task of syntactic analysis is to determine the syntactic structure of a sentence or the dependency relationships between words in a sentence. Semantic analysis mainly includes semantic disambiguation and semantic representation.

[0073] Deep Q Network (DQN) is a deep reinforcement learning model.

[0074] Information entropy, borrowed from thermodynamics by CEShannon, is used to address the problem of quantifying information. In thermodynamics, thermal entropy represents the degree of disorder in molecular states; Shannon uses the concept of information entropy to describe the uncertainty of information. The greater the information entropy, the greater the uncertainty and the lower the probability.

[0075] Currently, two common methods of consultation and interaction are used to determine diagnostic information: First, relying primarily on the doctor's personal experience, obtaining the user's symptom information through online and offline inquiries to determine diagnostic information; second, using pre-designed questionnaires by experts, with corresponding navigation paths to determine diagnostic information. The first method, relying entirely on manual consultation, is time-consuming, labor-intensive, and wastes significant high-quality medical resources. The second method, using existing similar questionnaires and templates, results in a monotonous consultation path, covers only a limited number of departments, and heavily depends on the quality of the expert-customized templates. Because doctors are typically only familiar with the diagnosis and consultation of diseases within their own department, they cannot consider diseases from all departments during the consultation, potentially leading to inaccurate diagnostic information. Furthermore, the questionnaires developed by doctors are difficult to merge or expand.

[0076] With the development of the internet, the health industry is increasingly integrated with the internet, and online consultations are gradually becoming a convenient way for users to seek health advice. The online consultation scenario differs significantly from in-person doctor consultations. Consultation systems analyze online consultation data to provide diagnostic and treatment suggestions. However, existing consultation systems are still immature, primarily targeting common or mildly symptomatic diseases, and lack objectivity, accuracy, and standardization.

[0077] To address the aforementioned issues, this application proposes a consultation processing method, the main inventive idea of ​​which is as follows: First, a complete consultation knowledge graph is constructed, covering nodes including 200+ common diseases, 400+ symptoms, disease causes, disease department information, differential symptom identification, complications, and disease course. The consultation knowledge graph is used to infer the symptoms input by the user, resulting in a candidate disease set and a candidate symptom set. Second, a probabilistic graphical model is used to determine the score (i.e., probability value) of each candidate disease in the candidate disease set, and a preset number of candidate diseases with high to low scores are obtained as the most likely diseases the user has. Finally, a preset rule or reinforcement learning algorithm is used to determine whether the consultation should end. If the consultation is determined to end, the data on the most likely diseases the user has and other diagnostic information are output; if further inquiry is needed, a target symptom can be selected from the candidate symptom set according to a preset questioning strategy, and the user is asked again.

[0078] The consultation knowledge graph in the above solution covers multidisciplinary data in the medical field. Combining the consultation knowledge graph with candidate diseases and candidate symptoms can effectively reduce the misdiagnosis rate of the consultation system and improve the accuracy of diagnosis and treatment information.

[0079] Before introducing the consultation processing method provided in this application, the application scenarios of the consultation processing method will be briefly introduced below.

[0080] Figure 1 This is a schematic diagram illustrating a scenario for the consultation processing method provided in an embodiment of this application. Figure 1 As shown, the scenario includes a first terminal device 11, a second terminal device 12, and a consultation server 13 (or consultation server, consultation platform). The first terminal device 11 and the second terminal device 12 are respectively connected to the consultation server 13.

[0081] In one optional implementation, the first terminal device 11 and the second terminal device 12 are pre-installed with the application APP of the consultation server 13, and the user of the first terminal device 11 or the second terminal device 12 can access the consultation server 13 through the application.

[0082] In one alternative implementation, users of the first terminal device 11 or the second terminal device 12 can also access the consultation service terminal 13 through web pages, application mini-programs, or other channels.

[0083] As an example, the first terminal device 11 can be a patient-side terminal device, such as a patient's smartphone, tablet, laptop, desktop computer, or a fixed or mobile terminal (such as a fixed or mobile intelligent robot) set up in a hospital's public area.

[0084] As an example, the second terminal device 12 can be a doctor's terminal device, such as a doctor's smartphone, tablet, laptop, desktop computer, or other terminal device. For instance, during a consultation, the doctor can access the consultation server 13 through the second terminal device 12 to obtain the diagnostic information provided by the consultation server 13, which can be used to assist the doctor in making a medical diagnosis.

[0085] In one optional implementation, the consultation server 11 has a built-in processing device for executing the method steps of the embodiments of this application. Optionally, the consultation server 13 has a storage space storing a consultation knowledge graph.

[0086] In an optional implementation, if the first terminal device 11 is an intelligent robot, a processing device can be integrated into the intelligent robot, enabling the intelligent robot to execute the method steps of the embodiments of this application. Optionally, the intelligent robot's storage space stores a medical knowledge graph. For example, users can directly interact with the intelligent robot through voice or text input to obtain medical information or inquiry information.

[0087] Optionally, the intelligent robot, as a regular terminal, interacts with the consultation service (considering the large memory space occupied by the knowledge graph, the intelligent robot may not need to store the consultation knowledge graph) to provide users with consultation information or inquiry information.

[0088] Based on the above scenario, the technical solution provided by the embodiments of this application will be described in detail below through specific examples. The following embodiments use the interaction between a client and a consultation server as an example to illustrate the solution, wherein the client can correspond to... Figure 1 For any of the terminal devices shown, the consultation server can correspond to... Figure 1 The consultation system shown.

[0089] It should be noted that the technical solutions provided in the embodiments of this application may include some or all of the following contents. The following specific embodiments may be combined with each other, and the same or similar concepts or processes may not be described again in some embodiments.

[0090] Figure 2 Interactive illustration of the consultation processing method provided in the embodiments of this application Figure 1 .like Figure 2 As shown, the consultation processing method in this embodiment includes:

[0091] Step 201: Receive symptom information from the client, which includes at least one symptom data entered by the user.

[0092] In this embodiment, the user accesses the consultation server through the client and can describe one or more symptoms via voice or text on the client. Based on the user's voice or text description, the client sends one or more symptom data to the consultation server.

[0093] Optionally, the client can use existing named entity recognition (NER) and relation extraction (RE) algorithms to extract the main symptom data described by the user.

[0094] Step 202: Based on the preset medical history knowledge graph, obtain the first candidate disease set and the first candidate symptom set corresponding to the symptom information.

[0095] The first candidate disease set includes multiple candidate diseases, and the first candidate symptom set includes all candidate symptoms of the multiple candidate diseases.

[0096] In this embodiment, the medical history knowledge graph includes nodes for diseases, symptoms, disease causes, disease department information, differential symptoms, complications, and disease course. Obtaining a first candidate disease set corresponding to symptom information based on the preset medical history knowledge graph specifically includes: based on at least one symptom data, obtaining a first candidate disease set corresponding to each symptom data from the medical history knowledge graph. The first candidate disease set can be denoted as V. cand .

[0097] Optionally, for each disease in the first candidate disease set, all symptom data contained in each candidate disease are obtained based on a preset medical history knowledge graph to generate a first candidate symptom set. The first candidate symptom set can be denoted as P. cand .

[0098] Step 203: Determine the diagnosis and treatment information corresponding to the symptom information up to the current round of inquiry from the first candidate disease set through probability analysis. The diagnosis and treatment information is used to indicate at least one candidate disease.

[0099] In one optional embodiment, diagnostic information can be determined using the following probability analysis method:

[0100] Step 2031: Obtain the score of each candidate disease in the first candidate disease set. The score of each candidate disease is used to indicate the probability value that the user has the candidate disease.

[0101] Step 2032: Based on the scores of multiple candidate diseases in the first candidate disease set, determine the diagnosis and treatment information corresponding to the symptom information up to the current round of inquiry.

[0102] The diagnostic information includes a preset number of candidate diseases ranked from highest to lowest score among multiple candidate diseases. It should be noted that this embodiment does not impose a specific limit on the preset number and can be reasonably set according to actual needs.

[0103] For example, assuming the preset number is 10, the top 10 disease data with the highest scores are obtained based on the scores of multiple candidate diseases in the first candidate disease set. The diagnosis and treatment information corresponding to the symptom information in the current query round includes at least the top 10 disease data with the highest scores (i.e., diagnosis data).

[0104] Optionally, the diagnostic information may also include treatment recommendations for each of the preset number of candidate diseases with high scores.

[0105] In an optional embodiment, step 2031 specifically includes: obtaining the score of each candidate disease in the first candidate disease set through a probabilistic graphical model.

[0106] The probabilistic graphical model can be the Noisy-or probabilistic graphical model. It should be noted that this embodiment does not specifically limit the probabilistic graphical model; other probabilistic graphical models besides the Noisy-or model can also be used to determine the score for each candidate disease.

[0107] To facilitate understanding, the following section uses the Noisy-or probabilistic graphical model as an example to explain in detail how to obtain the score of each candidate disease in the first candidate disease set.

[0108] As an example, obtaining the score for each candidate disease in the first candidate disease set using a probabilistic graphical model includes the following steps:

[0109] Step 1: Obtain the contribution of each symptom of the first candidate disease to the first candidate disease. The contribution is used to indicate the statistical probability value of the sample of the first candidate disease accompanied by the candidate symptom. The first candidate disease is any one of multiple candidate diseases.

[0110] The sample statistical probability value is determined based on a large number of statistical samples.

[0111] Specifically, the contribution λ of the j-th symptom (Sym) to the i-th disease (Dis) can be determined using the following formula. j :

[0112]

[0113] Among them, #occurrence(Sym j #co_occurrence(Dis) represents the number of times the j-th symptom occurs in the statistical sample. i Sym j ) represents the number of times the j-th symptom and the i-th disease occur simultaneously in the statistical sample.

[0114] The following example illustrates the calculation of sample statistical probability values.

[0115] Suppose that in 20,000 medical records (i.e., medical records samples), "cough" appears 100 times, and "cough" and "cold" appear together (co_occurrence) 10 times in a medical record. Then the contribution of "cough" to "cold" is: P(cold|cough) = 10 / 100 = 0.1.

[0116] Step 2: Determine the score of the first candidate disease based on the contribution of all candidate symptoms to the first candidate disease.

[0117] Specifically, the score of the first candidate disease is determined based on the contribution of all candidate symptoms to the first candidate disease and the noise parameter.

[0118] This step can also be represented as: inputting the contribution of all candidate symptoms of the first candidate disease to the first candidate disease into the Noisy-or probabilistic graphical model to obtain the score of the first candidate disease.

[0119] The Noisy-or probabilistic graphical model can be represented by the following formula:

[0120]

[0121] in, p represents the score of the first candidate disease v1. j Let λj represent the j-th symptom of the first candidate disease, j∈[1,k], where k is a positive integer, and λ0 represent the noise parameter. j This represents the statistical probability value of a sample where the first candidate disease is accompanied by the j-th symptom.

[0122] Step 204: Based on the preset number of inquiries, the information entropy of at least one candidate disease in the diagnosis and treatment information, and any one of the pre-trained diagnostic decision models, determine whether to output the diagnosis and treatment information or proceed to the next round of inquiries.

[0123] In this step, determining to output treatment information can be understood as determining to stop the inquiry and output the treatment information. It should be understood that if it is determined to proceed to the next round of inquiry, then it is not necessary to output the treatment information determined in the current round of inquiry.

[0124] The following describes in detail whether to continue the inquiry through several specific implementation methods.

[0125] In one optional implementation, determining whether to output diagnostic information or proceed to the next round of inquiry based on the information entropy of at least one candidate disease in the diagnostic information includes: determining to output diagnostic information if the sum of the information entropy of all candidate diseases in at least one candidate disease is less than a preset threshold; or determining to proceed to the next round of inquiry if the sum of the information entropy of all candidate diseases in at least one candidate disease is greater than or equal to a preset threshold.

[0126] As an example, the system obtains the scores of a preset number of candidate diseases from multiple candidate diseases, ranked from highest to lowest; determines the information entropy of each candidate disease within the preset number of candidate diseases; if the sum of the information entropies of all candidate diseases within the preset number of candidate diseases is less than a preset threshold, it determines to output diagnostic information; or, if the sum of the information entropies of all candidate diseases within the preset number of candidate diseases is greater than or equal to the preset threshold, it determines to proceed to the next round of inquiry.

[0127] In this embodiment, the information entropy of each candidate disease can be determined by the following formula:

[0128] entropy(v) = -S v *logS v

[0129] Where entropy(v) represents the information entropy of candidate disease v, and S v This represents the score for candidate disease v. Information entropy is used to describe the uncertainty of information.

[0130] For example, assuming a preset number of 10, after obtaining the scores of these 10 candidate diseases, the information entropy of each of these 10 candidate diseases is determined. If the sum of the information entropies of these 10 candidate diseases is less than a preset threshold, it indicates that the uncertainty of the data for these 10 candidate diseases determined in step 204 is low (or that the accuracy is high), and therefore, treatment information including the data for these 10 candidate diseases can be output. If the sum of the information entropies of these 10 candidate diseases is greater than or equal to the preset threshold, it indicates that the instability of the data for these 10 candidate diseases determined in step 204 is high (or that the accuracy is low), and therefore, a next round of querying is required.

[0131] In one optional implementation, determining whether to output medical information or proceed to the next round of inquiries based on a preset number of inquiries includes: determining whether the current round of inquiries has reached the preset number of inquiries; if the current round of inquiries has reached the preset number of inquiries, determining to output medical information; or, if the current round of inquiries has not reached the preset number of inquiries, determining to proceed to the next round of inquiries.

[0132] For example, assuming a preset number of queries is 5, after the client interacts with the question-and-answer server 5 times, the server outputs the confirmed diagnosis information from the fifth query. It should be understood that the server determines the diagnosis information for each round of queries, and the accuracy of the diagnosis information continuously improves as the number of queries increases.

[0133] In one optional implementation, determining whether to output diagnostic information or proceed to the next round of inquiry based on a pre-trained diagnostic decision model includes: inputting symptom information into the pre-trained diagnostic decision model, and determining whether to output diagnostic information based on the output value of the diagnostic decision model.

[0134] In one scenario, if the diagnostic decision model outputs a first value, the output diagnostic information is determined.

[0135] In one scenario, if the diagnostic decision model outputs a second value, it determines to proceed with the next round of questioning.

[0136] For example, when the first value is 1, it is determined to output the diagnosis and treatment information; when the second value is 0, it is determined to proceed to the next round of inquiry. This embodiment does not specifically limit the specific values ​​of the first and second values, as long as the two decision results can be distinguished.

[0137] The diagnostic decision model is obtained by training a fully connected neural network using a reinforcement learning algorithm through multiple sample sequences. The sample sequences include at least one symptom data and a decision result, which is used to indicate the output of diagnostic information or to conduct the next round of inquiry.

[0138] The construction of the sample sequence for the diagnostic decision model includes: using NLP algorithms to perform structured processing on the dialogue data between the client and the question-answering server to obtain dialogue samples {"symptom 1", "symptom 2", ..., "symptom k"}, labeling the decision action "decision 1" corresponding to the dialogue samples, and obtaining a sample sequence {"symptom 1", "symptom 2", ..., "symptom k", "decision 1"}.

[0139] Once enough sample sequences have been collected, the DQN algorithm can be used to build a sequence decision model:

[0140] a t =MLP(s) t )

[0141] Among them, a t In this embodiment, a represents the action at time t. t This can be understood as the decision action in the current round, with an action space of 2, for example, a. t =0 means "continue asking questions", a t =1 indicates "diagnosis". s tThe state at time t is represented by a one-hot vector with dimension D. In this embodiment, the size of D is the total number of candidate symptoms in the sample sequence. MLP is a fully connected neural network.

[0142] When the predicted value a output by the MLP network t When the accuracy reaches a preset threshold, the training process of the diagnostic decision model ends.

[0143] The consultation processing method illustrated in this embodiment receives symptom information from the client. First, it obtains a set of candidate diseases corresponding to the symptom information based on a preset consultation knowledge graph. Then, it determines the score of each candidate disease in the set through probability analysis; a higher score indicates a greater probability that the user has the candidate disease. Next, based on the scores of multiple candidate diseases in the set, it determines the diagnostic information corresponding to the symptom information collected up to the current round of inquiry. This diagnostic information includes at least one candidate disease with a high score. Finally, based on a preset number of inquiries, the information entropy of at least one candidate disease in the diagnostic information, and any one of the pre-trained diagnostic decision-making models, it determines whether to output the diagnostic information or proceed to the next round of inquiry. On the one hand, the above scheme, combined with a consultation knowledge graph, can effectively reduce the misdiagnosis rate of the consultation system and improve the accuracy of diagnosis. On the other hand, by analyzing the information entropy of candidate diseases in the diagnostic information to determine whether to stop the inquiry, the inquiry time can be shortened, improving the user's consultation experience.

[0144] Figure 3 Interactive illustration of the consultation processing method provided in the embodiments of this application Figure 2 .exist Figure 2 Based on the illustrated embodiments, as Figure 3 As shown, the consultation processing method in this embodiment further includes:

[0145] Step 301: If it is determined to proceed to the next round of questioning, determine the target questioning symptom from the first candidate symptom set.

[0146] In this embodiment, if it is determined to proceed to the next round of questioning, the target candidate symptom can be determined from the first candidate symptom set through any of the following implementation methods:

[0147] In one optional implementation, the detailed symptoms of at least one candidate symptom confirmed by the user in the current round of inquiry are selected from the candidate diseases with the highest scores among multiple candidate diseases in the first candidate disease set as the target inquiry symptom.

[0148] For example, if candidate disease 1 has the highest score among the 10 candidate diseases with the highest scores identified in the current round of inquiry, indicating that the user has the highest probability of having candidate disease 1, the consultation service can further ask the user for detailed information on one or more symptoms that the user has confirmed for candidate disease 1. For example, if candidate disease 1 is "common cold" and the user has confirmed "runny nose," the consultation service can further ask the user, for example, "What color is the nasal discharge?" (i.e., detailed symptoms of runny nose). By receiving the user's response data, the service can determine and send the corresponding diagnosis and / or treatment suggestions, or continue to ask questions.

[0149] In one alternative implementation, the symptoms other than those already confirmed by the user are selected from the candidate diseases with the highest scores among multiple candidate diseases in the first candidate disease set as target query symptoms.

[0150] For example, if candidate disease 1 has the highest score among the 10 candidate diseases with the highest scores identified in the current round of inquiry, indicating that the user has the highest probability of having candidate disease 1, the consultation service can further ask the user about one or more symptoms of candidate disease 1 that the user has not yet confirmed. For example, if candidate disease 1 is "common cold" and the user has confirmed symptoms such as "runny nose," other symptoms of "common cold" include, for example, "fever," "dizziness," and "loss of appetite," the consultation service can further ask the user, for example, "Do you have a fever?" (i.e., other symptoms of the common cold that the user has not confirmed). By receiving the user's response data, the service can determine and send the corresponding diagnosis and / or treatment suggestions, or continue to ask questions.

[0151] In an optional implementation, the candidate symptom that causes the fastest decrease in the overall information entropy of the first candidate disease set is selected from all candidate symptom data of multiple candidate diseases in the first candidate disease set as the target query symptom.

[0152] In this embodiment, the overall information entropy of the first candidate symptom in the first candidate symptom set to the first candidate disease set is determined by the ratio of the sum of scores of all candidate diseases in the second candidate disease set in the next round of querying to the sum of scores of all candidate diseases in the first candidate disease set. The second candidate disease set is determined based on the first candidate disease set and the first candidate symptom, and the first candidate symptom is any one of the candidate symptoms in the first candidate symptom set.

[0153] As an example, the overall information entropy of each candidate symptom of each candidate disease in the first candidate disease set to the first candidate disease set can be determined by the following formula:

[0154] entropy(p)=-prob(p)*log(prob(p))

[0155]

[0156] Where entropy(p) represents the candidate symptom p (i.e., the first candidate symptom) against the candidate disease set V. cand The overall information entropy of (i.e., the first set of candidate diseases); prob(p) represents the new set of candidate diseases Vcand∩V if candidate symptom p is the target query symptom (i.e., the symptom of the next round of query). p The sum of the scores of (i.e., the second candidate disease set) and the candidate disease set V cand The ratio of the total scores; S v This represents the score of candidate disease v.

[0157] For example, if 10 candidate diseases with high scores are identified in the current round of inquiry, candidate symptoms for these 10 diseases can be obtained. For each obtained candidate symptom, the overall information entropy of each candidate symptom relative to the current set of candidate diseases (i.e., these 10 candidate diseases) can be determined. It should be understood that the lower the information entropy, the lower the uncertainty, and the higher the probability value. Therefore, from all candidate symptoms of the 10 candidate diseases with high scores, the candidate symptom with the lowest information entropy can be selected as the target symptom to be asked of the user in the next round of inquiry.

[0158] In this embodiment, by identifying the candidate symptom with the fastest decrease in information entropy, the most informative symptom can be located quickly and accurately, thereby improving the efficiency of diagnosing a disease and accelerating the consultation speed of the consultation service.

[0159] Step 302: Send an inquiry message to the client. The inquiry message is used to ask the user if they have the target inquiry symptoms.

[0160] Step 303: Receive the reply information from the client.

[0161] In this step, the user's response may include confirmation or non-confirmation. For example, if the inquiry from the online consultation service is "Do you have a fever?", the user's response may be "Yes" or "No".

[0162] Step 304: Update the diagnosis and treatment information based on the response information, and determine whether to output the updated diagnosis and treatment information or to conduct the next round of inquiries.

[0163] In this step, symptom information is updated based on response information. A set of candidate diseases and candidate symptoms corresponding to the updated symptom information is obtained based on a pre-defined medical knowledge graph. Through probability analysis, the corresponding treatment information for the symptom information up to the current round of inquiry is determined from the updated candidate disease set. This determines whether to output the treatment information or continue the inquiry. The specific implementation process is as follows... Figure 2Steps 202 to 204 of the illustrated embodiment are similar and can be referred to the above embodiment, and will not be repeated here.

[0164] It should be noted that the updated symptom information is determined based on the user's response information. The updated symptom information may involve adding at least one new symptom to the original symptom data, or it may involve excluding at least one symptom from the original symptom data. The consultation server obtains a set of candidate diseases corresponding to the updated symptom information based on a pre-defined consultation knowledge graph. It should be understood that as the symptom data changes, the number of candidate diseases in the candidate disease set may increase, decrease, or remain unchanged. A probabilistic graphical model is used to determine the score of each candidate disease in the updated candidate disease set. Based on the score of each candidate disease in the updated candidate disease set, the diagnosis and treatment information corresponding to the symptom information up to the current round of inquiry (i.e., the updated diagnosis and treatment information) is determined. Using the pre-defined rules or reinforcement learning algorithm described in the above embodiment, it is determined whether to output the updated diagnosis and treatment information.

[0165] The consultation processing method shown in this embodiment, based on the previous embodiment, if it is determined that a next round of inquiry is needed, selects the target symptom and asks the user again based on multiple candidate diseases in the current candidate disease set using a preset questioning strategy. A new round of data processing and analysis is performed based on the client's response information, and a final determination is made as to whether to output diagnostic information or continue asking questions. Based on the preset questioning strategy in this embodiment, targeted inquiries can simultaneously improve the consultation efficiency and accuracy of the consultation system.

[0166] The above describes the consultation processing method provided in the embodiments of this application. The consultation processing device provided in the embodiments of this application will be described below.

[0167] This application embodiment can divide the consultation processing device into functional modules according to the above method embodiment. For example, each function can be divided into a separate functional module, or two or more functions can be integrated into one processing module. The integrated module can be implemented in hardware or in the form of software functional modules.

[0168] It should be noted that the module division in this embodiment is illustrative and represents only one logical functional division. In actual implementation, other division methods may be used. The following explanation uses the division of functional modules according to their respective functions as an example.

[0169] Figure 4 A schematic diagram of the structure of the medical consultation processing device provided in the embodiments of this application. Figure 1 .like Figure 4As shown, the consultation processing device 400 of this embodiment includes: a receiving module 401, an acquisition module 402, and a processing module 403.

[0170] The receiving module 401 is used to receive symptom information from the client, the symptom information including at least one symptom data input by the user;

[0171] The acquisition module 402 is used to acquire a first candidate disease set and a first candidate symptom set corresponding to the symptom information based on a preset medical knowledge graph. The first candidate disease set includes multiple candidate diseases, and the first candidate symptom set includes all candidate symptoms of the multiple candidate diseases.

[0172] Processing module 403 is used to determine the diagnosis and treatment information corresponding to the symptom information up to the current round of inquiry from the first candidate disease set through probability analysis, wherein the diagnosis and treatment information is used to indicate at least one candidate disease;

[0173] Based on any one of the preset number of inquiries, the information entropy of at least one candidate disease in the diagnosis and treatment information, and the pre-trained diagnostic decision model, the diagnosis and treatment information is output or the next round of inquiries is initiated.

[0174] In an optional embodiment of this example, the acquisition module 402 is used to acquire the score of each candidate disease in the first candidate disease set, and the score is used to indicate the probability value of the user having the candidate disease;

[0175] Processing module 403 is used to determine the diagnosis and treatment information corresponding to the symptom information up to the current query round based on the scores of the plurality of candidate diseases in the first candidate disease set. The diagnosis and treatment information includes a preset number of candidate diseases with scores from high to low among the plurality of candidate diseases.

[0176] In an optional embodiment of this example, the acquisition module 402 is used to acquire the contribution of each candidate symptom of the first candidate disease to the first candidate disease, wherein the contribution is used to indicate the sample statistical probability value of the first candidate disease accompanied by the candidate symptom, and the first candidate disease is any one of the plurality of candidate diseases;

[0177] Processing module 403 is used to determine the score of the first candidate disease based on the contribution of all candidate symptoms of the first candidate disease to the first candidate disease.

[0178] In an optional embodiment of this example, the processing module 403 is used to determine the score of the first candidate disease based on the contribution of all candidate symptoms of the first candidate disease to the first candidate disease and noise parameters.

[0179] In an optional embodiment of this example, the processing module 403 is configured to:

[0180] If the sum of the information entropy of all candidate diseases in the at least one candidate disease is less than a preset threshold, the diagnosis and treatment information is determined to be output; or

[0181] If the sum of the information entropy of all candidate diseases in the at least one candidate disease is greater than or equal to the preset threshold, then proceed to the next round of inquiry.

[0182] In an optional embodiment of this example, the processing module 403 is configured to:

[0183] Determine whether the current round of queries has reached the preset number of queries;

[0184] If the current round of inquiry reaches the preset number of inquiries, the diagnosis and treatment information is output; or, if the current round of inquiry does not reach the preset number of inquiries, the next round of inquiry is initiated.

[0185] In an optional embodiment of this example, the processing module 403 is configured to:

[0186] The symptom information is input into the pre-trained diagnostic decision model, and the diagnostic decision model outputs the diagnostic information or proceeds to the next round of inquiry based on the output value of the model.

[0187] The diagnostic decision model is obtained by training a fully connected neural network using a reinforcement learning algorithm through multiple sample sequences. The sample sequences include at least one symptom data and a decision result, which is used to indicate the output of diagnostic information or to conduct the next round of inquiry.

[0188] In an optional embodiment of this example, the processing module 403 is configured to:

[0189] If the diagnostic decision model outputs a first value, the diagnostic information will be output; or

[0190] If the diagnostic decision model outputs a second value, it determines to proceed to the next round of inquiry.

[0191] Figure 5 A schematic diagram of the structure of the medical consultation processing device provided in the embodiments of this application. Figure 2 .exist Figure 4 Based on the device shown, such as Figure 5 As shown, the consultation processing device 400 of this embodiment includes: a sending module 404.

[0192] Processing module 403 is used to determine the target inquiry symptom from the first candidate symptom set if it is determined to proceed to the next round of inquiry;

[0193] Sending module 404 is used to send query information to the client, the query information being used to ask the user whether they have the target query symptoms;

[0194] The receiving module 401 is used to receive reply information from the client, and the processing module 403 is used to update the diagnosis and treatment information based on the reply information, and determine whether to output the updated diagnosis and treatment information or to conduct the next round of inquiry.

[0195] In an optional embodiment of this example, the processing module 403 is configured to perform at least one of the following:

[0196] From the candidate diseases with the highest scores in the first candidate disease set, select the detailed symptoms of at least one candidate symptom confirmed by the user in the current round of inquiry as the target inquiry symptom; or

[0197] From the candidate diseases with the highest scores in the first candidate disease set, select other symptoms besides those already confirmed by the user as the target query symptoms; or

[0198] The candidate symptom that causes the fastest decrease in the overall information entropy of the first candidate disease set is selected from all candidate symptoms of multiple candidate diseases in the first candidate disease set as the target query symptom.

[0199] In an optional embodiment of this example, the overall information entropy of the first candidate symptom in the first candidate symptom set to the first candidate disease set is determined based on the ratio of the sum of scores of all candidate diseases in the second candidate disease set in the next round of querying to the sum of scores of all candidate diseases in the first candidate disease set.

[0200] The second candidate disease set is determined based on the first candidate disease set and the first candidate symptom, wherein the first candidate symptom is any one of the candidate symptom sets in the first candidate symptom set.

[0201] In one optional embodiment of this example, the medical history knowledge graph includes nodes for disease, symptoms, disease causes, disease department information, differential symptoms, complications, and disease course;

[0202] The acquisition module 402 is used to acquire a set of candidate diseases corresponding to each symptom data from the consultation knowledge graph based on the at least one symptom data, and obtain the first set of candidate diseases.

[0203] The consultation processing device provided in this embodiment can execute the technical solutions of any of the above method embodiments. Its implementation principle and technical effect are similar, and will not be described again here.

[0204] Figure 6 This is a hardware structure diagram of the electronic device provided in an embodiment of this application. For example... Figure 6 As shown, the electronic device 500 provided in this embodiment includes:

[0205] Memory 501;

[0206] Processor 502; and

[0207] Computer programs;

[0208] The computer program is stored in memory 501 and configured to be executed by processor 502 to implement the technical solution of any of the above method embodiments. The implementation principle and technical effect are similar, and will not be repeated here.

[0209] Optionally, the memory 501 can be either standalone or integrated with the processor 502. When the memory 501 is a device independent of the processor 502, the electronic device 500 also includes a bus 503 for connecting the memory 501 and the processor 502.

[0210] This application also provides a computer-readable storage medium storing a computer program thereon, which is executed by a processor 502 to implement the technical solutions of any of the foregoing method embodiments.

[0211] This application provides a computer program product, including a computer program, which, when executed by a processor, implements the technical solutions of any of the foregoing method embodiments.

[0212] This application also provides a chip, including: a processing module and a communication interface, wherein the processing module is capable of executing the technical solutions of any of the foregoing method embodiments.

[0213] Furthermore, the chip also includes a storage module (e.g., a memory), which is used to store instructions, and a processing module is used to execute the instructions stored in the storage module. The execution of the instructions stored in the storage module causes the processing module to execute the technical solution of any of the aforementioned method embodiments.

[0214] It should be understood that the aforementioned processor can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), etc. A general-purpose processor can be a microprocessor or any conventional processor. The steps of the method disclosed in this invention can be directly manifested as execution by a hardware processor, or execution by a combination of hardware and software modules within the processor.

[0215] The memory may include high-speed RAM, and may also include non-volatile storage (NVM), such as at least one disk storage device, and may also be a USB flash drive, external hard drive, read-only memory, disk or optical disc, etc.

[0216] The bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, or an Extended Industry Standard Architecture (EISA) bus, etc. Buses can be categorized as address buses, data buses, control buses, etc. For ease of illustration, the buses shown in the accompanying drawings are not limited to a single bus or a single type of bus.

[0217] The aforementioned storage medium can be implemented from any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk. The storage medium can be any available medium accessible to general-purpose or special-purpose computers.

[0218] An exemplary storage medium is coupled to a processor, enabling the processor to read information from and write information to the storage medium. Alternatively, the storage medium can be an integral part of the processor. Both the processor and the storage medium can reside in an application-specific integrated circuit (ASIC). Alternatively, the processor and storage medium can exist as discrete components in an electronic device.

[0219] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features therein. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of this application.

Claims

1. A medical interview processing method characterized by comprising: include: Receive symptom information from a client, the symptom information including at least one symptom data input by the user; Based on a preset medical history knowledge graph, a first set of candidate diseases and a first set of candidate symptoms corresponding to the symptom information are obtained. The first set of candidate diseases includes multiple candidate diseases, and the first set of candidate symptoms includes all candidate symptoms of the multiple candidate diseases. Obtain a score for each candidate disease in the first candidate disease set, the score being used to indicate the probability value that the user has the candidate disease; wherein, the score of the candidate disease is determined based on the contribution of all candidate symptoms of the first candidate disease to the first candidate disease; the contribution is used to indicate the sample statistical probability value of the first candidate disease accompanied by the candidate symptoms, and the first candidate disease is any one of a plurality of candidate diseases; Based on the scores of the multiple candidate diseases in the first candidate disease set, the diagnosis and treatment information corresponding to the symptom information up to the current round of inquiry is determined. The diagnosis and treatment information includes a preset number of candidate diseases with scores from high to low among the multiple candidate diseases. Based on any one of the preset number of inquiries, the information entropy of at least one candidate disease in the medical information, and the pre-trained diagnostic decision model, the medical information is output or a next round of inquiries is initiated; wherein, the information entropy is determined by the following formula: ;in, The information entropy of candidate disease v is represented by its value. This represents the score of candidate disease v; Based on the information entropy of at least one candidate disease in the diagnostic information, determine whether to output the diagnostic information or proceed to the next round of inquiry, including: If the sum of the information entropy of all candidate diseases in the at least one candidate disease is less than a preset threshold, the diagnosis and treatment information is determined to be output; or If the sum of the information entropy of all candidate diseases in the at least one candidate disease is greater than or equal to the preset threshold, then proceed to the next round of inquiry. The step of obtaining the score for each candidate disease in the first candidate disease set includes: Obtain the contribution of each candidate symptom of the first candidate disease to the first candidate disease; wherein the contribution is determined based on the following formula: ; in, This represents the contribution of the j-th symptom to the i-th disease; This represents the number of times the j-th symptom appears in the statistical sample. This represents the number of times the j-th symptom and the ith disease occur simultaneously in the statistical sample. The contribution of all candidate symptoms to the first candidate disease is input into the probabilistic graphical model to obtain the score of the first candidate disease; the probabilistic graphical model is represented by the following formula: in, Indicates the first candidate disease The score, This represents the j-th symptom of the first candidate disease, where j∈[1,k] and k is a positive integer. Indicates noise parameters, This represents the statistical probability value of a sample where the first candidate disease is accompanied by the j-th symptom.

2. The method according to claim 1, characterized in that, Based on the preset number of inquiries, determine whether to output the medical information or proceed to the next round of inquiries, including: Determine whether the current round of queries has reached the preset number of queries; If the current round of inquiry reaches the preset number of inquiries, the diagnosis and treatment information is output; or, if the current round of inquiry does not reach the preset number of inquiries, the next round of inquiry is initiated.

3. The method according to claim 1, characterized in that, Based on the pre-trained diagnostic decision model, determine whether to output the diagnostic information or proceed to the next round of inquiry, including: The symptom information is input into the pre-trained diagnostic decision model, and the diagnostic decision model outputs the diagnostic information or proceeds to the next round of inquiry based on the output value of the model. The diagnostic decision model is obtained by training a fully connected neural network using a reinforcement learning algorithm through multiple sample sequences. The sample sequences include at least one symptom data and a decision result, which is used to indicate the output of diagnostic information or to conduct the next round of inquiry.

4. The method according to claim 3, characterized in that, The step of determining whether to output the diagnostic information or proceed to the next round of inquiry based on the output value of the diagnostic decision model includes: If the diagnostic decision model outputs a first value, the diagnostic information will be output; or If the diagnostic decision model outputs a second value, it determines to proceed to the next round of inquiry.

5. The method according to any one of claims 1-2, characterized in that, The method further includes: If it is determined to proceed to the next round of inquiry, the target inquiry symptom is determined from the first candidate symptom set, and inquiry information is sent to the client. The inquiry information is used to ask the user whether they have the target inquiry symptom. Receive a response from the client, update the diagnosis and treatment information based on the response, and determine whether to output the updated diagnosis and treatment information or proceed with the next round of inquiries.

6. The method according to claim 5, characterized in that, The step of determining the target inquiry symptom from the first candidate symptom set includes at least one of the following: From the candidate diseases with the highest scores in the first candidate disease set, select the detailed symptoms of at least one candidate symptom confirmed by the user in the current round of inquiry as the target inquiry symptom; or From the candidate diseases with the highest scores in the first candidate disease set, select other symptoms besides those already confirmed by the user as the target query symptoms; or The candidate symptom that causes the fastest decrease in the overall information entropy of the first candidate disease set is selected from all candidate symptoms of multiple candidate diseases in the first candidate disease set as the target query symptom.

7. The method according to claim 6, characterized in that, The overall information entropy of the first candidate symptom in the first candidate symptom set to the first candidate disease set is determined based on the ratio of the sum of scores of all candidate diseases in the second candidate disease set in the next round of querying to the sum of scores of all candidate diseases in the first candidate disease set. The second candidate disease set is determined based on the first candidate disease set and the first candidate symptom, wherein the first candidate symptom is any one of the candidate symptom sets in the first candidate symptom set.

8. The method according to any one of claims 1-2, characterized in that, The medical history knowledge graph includes nodes for diseases, symptoms, causes of diseases, disease departments, differential symptoms, complications, and disease course. The step of obtaining the first candidate disease set corresponding to the symptom information based on the preset consultation knowledge graph includes: obtaining the candidate disease set corresponding to each symptom data from the consultation knowledge graph according to the at least one symptom data, and obtaining the first candidate disease set.

9. A medical history processing device, characterized in that, include: A receiving module is used to receive symptom information from a client, the symptom information including at least one symptom data input by the user; The acquisition module is used to acquire a first candidate disease set and a first candidate symptom set corresponding to the symptom information based on a preset medical history knowledge graph. The first candidate disease set includes multiple candidate diseases, and the first candidate symptom set includes all candidate symptoms of the multiple candidate diseases. The processing module is used to determine the diagnosis and treatment information corresponding to the symptom information up to the current round of inquiry from the first candidate disease set through probability analysis, and the diagnosis and treatment information is used to indicate at least one candidate disease; Based on any one of the preset number of inquiries, the information entropy of at least one candidate disease in the medical information, and the pre-trained diagnostic decision model, the medical information is output or a next round of inquiries is initiated; wherein, the information entropy is determined by the following formula: ;in, The information entropy of candidate disease v is represented by its value. This represents the score of candidate disease v; The processing module is specifically used to obtain the score of each candidate disease in the first candidate disease set, and the score is used to indicate the probability value that the user has the candidate disease; wherein, the score of the candidate disease is determined based on the contribution of all candidate symptoms of the first candidate disease to the first candidate disease; the contribution is used to indicate the sample statistical probability value of the first candidate disease accompanied by the candidate symptoms, and the first candidate disease is any one of multiple candidate diseases; Based on the scores of the multiple candidate diseases in the first candidate disease set, the diagnosis and treatment information corresponding to the symptom information up to the current round of inquiry is determined. The diagnosis and treatment information includes a preset number of candidate diseases with scores from high to low among the multiple candidate diseases. The processing module is specifically configured to determine and output the diagnostic information if the sum of the information entropies of all candidate diseases among the at least one candidate disease is less than a preset threshold; or If the sum of the information entropy of all candidate diseases in the at least one candidate disease is greater than or equal to the preset threshold, then proceed to the next round of inquiry. The processing module is specifically used to obtain the contribution of each candidate symptom of the first candidate disease to the first candidate disease; wherein the contribution is determined based on the following formula: ; in, This represents the contribution of the j-th symptom to the i-th disease; This represents the number of times the j-th symptom appears in the statistical sample. This represents the number of times the j-th symptom and the ith disease occur simultaneously in the statistical sample. The contribution of all candidate symptoms to the first candidate disease is input into the probabilistic graphical model to obtain the score of the first candidate disease; the probabilistic graphical model is represented by the following formula: in, Indicates the first candidate disease The score, This represents the j-th symptom of the first candidate disease, where j∈[1,k] and k is a positive integer. Indicates noise parameters, This represents the statistical probability value of a sample where the first candidate disease is accompanied by the j-th symptom.

10. An electronic device, characterized in that, include: Memory; processor; as well as Computer programs; The computer program is stored in the memory and configured to be executed by the processor to implement the method as described in any one of claims 1-8.

11. A computer-readable storage medium, characterized in that, It stores a computer program, which is executed by a processor to implement the method as described in any one of claims 1-8.

12. A computer program product, characterized in that, Includes a computer program that, when executed by a processor, implements the method of any one of claims 1-8.

Images