Medical history inquiry method, device, electronic device, and storage medium

Abstract

The present disclosure provides a medical consultation method, apparatus, electronic device, and storage medium, which relate to the field of artificial intelligence technology, and specifically to the fields of intelligent recommendation and deep learning technology. The specific implementation solution is as follows: obtain the medical consultation information of the user seeking medical advice, and determine whether the medical consultation information can be classified into departments; determine the target medical consultation department according to the result of whether it can be classified into departments and the medical consultation information. The embodiments of the present disclosure can improve the accuracy of department classification for triage departments.

Description

Technical Field

[0001] The present disclosure relates to the field of artificial intelligence technology, specifically to the fields of intelligent recommendation and deep learning technology, and particularly to a consultation method, device, electronic device, and storage medium. Background Art

[0002] With the development of computer technology and the Internet, patients can seek medical advice through online consultation, enabling patients to obtain doctor diagnoses and treatments more conveniently.

[0003] The online consultation system determines the department based on the user's chief complaint information and distributes the user to a doctor in the corresponding department for further diagnosis and treatment. Summary of the Invention

[0004] The present disclosure provides a consultation method, device, electronic device, and storage medium.

[0005] According to one aspect of the present disclosure, there is provided a consultation method, including:

[0006] Obtaining the consultation information of the consultation user and determining whether the consultation information can be classified into departments;

[0007] Determining the target consultation department according to the result of whether the consultation information can be classified into departments and the consultation information.

[0008] According to one aspect of the present disclosure, there is provided a consultation device, including:

[0009] A module for determining whether the consultation information can be classified into departments, configured to obtain the consultation information of the consultation user and determine whether the consultation information can be classified into departments;

[0010] A module for determining the triage department, configured to determine the target consultation department according to the result of whether the consultation information can be classified into departments and the consultation information.

[0011] According to another aspect of the present disclosure, there is provided an electronic device, including:

[0012] At least one processor; and

[0013] A memory communicatively connected to the at least one processor; wherein,

[0014] The memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the consultation method according to any embodiment of the present disclosure, or execute the consultation method according to any embodiment of the present disclosure.

[0015] According to another aspect of the present disclosure, there is provided a non-transitory computer-readable storage medium storing computer instructions, wherein the computer instructions are used to cause the computer to execute the interrogation method described in any embodiment of the present disclosure, or execute the interrogation method described in any embodiment of the present disclosure.

[0016] According to another aspect of the present disclosure, there is provided a computer program product including a computer program, which when executed by a processor implements the interrogation method described in any embodiment of the present disclosure, or executes the interrogation method described in any embodiment of the present disclosure.

[0017] Embodiments of the present disclosure can improve the accuracy of the font generated by the interrogation model.

[0018] It should be understood that the content described in this part is not intended to identify the key or important features of the embodiments of the present disclosure, nor is it used to limit the scope of the present disclosure. Other features of the present disclosure will become easily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] The drawings are used to better understand the solution and do not constitute a limitation to the present disclosure. Among them:

[0020] Figure 1 is a schematic diagram of an interrogation method provided according to an embodiment of the present disclosure;

[0021] Figure 2 is a schematic diagram of an interrogation method provided according to an embodiment of the present disclosure;

[0022] Figure 3 is a schematic diagram of an interrogation method provided according to an embodiment of the present disclosure;

[0023] Figure 4 is a schematic diagram of an interrogation method provided according to an embodiment of the present disclosure;

[0024] Figure 5 is a schematic diagram of an interrogation device provided according to an embodiment of the present disclosure;

[0025] Figure 6 is a block diagram of an electronic device for implementing the interrogation method of the embodiment of the present disclosure. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0026] The following makes an explanation of the exemplary embodiments of the present disclosure in conjunction with the drawings. Various details of the embodiments of the present disclosure are included to help understanding, and they should be considered merely exemplary. Therefore, those of ordinary skill in the art should recognize that various changes and modifications can be made to the embodiments described herein without departing from the scope and spirit of the present disclosure. Similarly, for the sake of clarity and conciseness, the description of well-known functions and structures is omitted below.

[0027] Figure 1 It is a flowchart of a medical consultation method disclosed according to an embodiment of the present disclosure. This embodiment can be applicable to the situation where a user obtains the triage department corresponding to the medical consultation information. The method of this embodiment can be executed by a medical consultation device, which can be implemented in a software and / or hardware manner and is specifically configured in an electronic device with certain data operation capabilities. The electronic device can be a client device or a server device. The client device can be, for example, a mobile phone, a tablet computer, a vehicle-mounted terminal, and a desktop computer, etc.

[0028] S101, obtain the medical consultation information of the medical consultation user, and determine whether the medical consultation information can be classified.

[0029] The embodiments of the present disclosure can be applied to the application scenario of online medical consultation, or can also be applied to the application scenario of self-service triage of a triage device. The medical consultation user refers to a user who needs triage, specifically a user who needs to judge the severity and urgency of the condition and its affiliated specialty based on the main symptoms and signs, and reasonably arrange the diagnosis and treatment information such as the department and doctor for the user to visit. The medical consultation information can be the content describing the disease to be diagnosed of the medical consultation user, and can include at least one of the following: user attribute information, symptom information, and historical medical treatment information. The user attribute information can include name, gender, age, etc. The historical medical treatment information can include the lesion location, previous diseases, historical medications, and historical medical treatment materials (such as doctor's diagnosis records and medical imaging materials, etc.). The medical consultation information is used to triage the medical consultation user, specifically to classify and allocate doctors. Among them, classifying refers to determining the medical consultation department to which the medical consultation user belongs, and allocating a doctor refers to determining the recommended doctor for the medical consultation user in the previously determined medical consultation department.

[0030] The consultation information can be input in at least one way such as text, voice, image or video. The voice can be converted into text, and image recognition can be performed to recognize the text in the image, or some medical images can be recognized based on some image recognition models to obtain recognition results. Among them, for the directly obtained input text or the recognized text, further parsing and intent recognition can be performed to obtain parsing results and intent recognition results. The text can be input into a parsing model to obtain the feature information in the consultation information as the parsing result. For example, at least one piece of information such as text length, user attribute information, underlying diseases, symptoms and medications. Among them, the parsing model includes a bidirectional long short-term memory network (Long Short-Term Memory, LSTM) and a softmax network, or includes a bidirectional LSTM network and a conditional random field network (Conditional Random Field, CRF). The text is also used to be input into an intent recognition model to obtain the intent recognition result corresponding to the text. Exemplarily, the text is "What is the best hospital for treating diabetes?", and the intent recognition result includes that the underlying disease is diabetes and the consultation department with a high ranking for treating diabetes is queried. In addition, the parsing result and the intent recognition result can be input into a pre-established knowledge graph for inverted indexing to obtain associated department information. The knowledge graph includes the corresponding relationship between the parsing result, the intent recognition result and the department information. The parsing result, the intent recognition result and the associated department information can be associated with the consultation information.

[0031] Whether it is possible to conduct department classification judgment means judging whether the consultation user can be classified according to the consultation information. In fact, the clearer the main problem in the consultation information input by the consultation user, the more conducive it is to conduct department classification. The more vague and unclear the problem is, the less conducive it is to department classification. Through the department classification judgment, it is possible to pre-judge whether department classification can be carried out before determining the department classification result. According to the department classification judgment result, further department classification can be carried out to improve the accuracy of department classification. As mentioned above, the department classification judgment can take the parsing result, the intent recognition result and the associated department information corresponding to the consultation information as inputs to obtain the department classification judgment result.

[0032] Whether it can be judged by department can be as follows: input the consultation information into a pre-trained model for judging whether it can be divided into departments, and obtain the result of whether it can be divided into departments output by the model for judging whether it can be divided into departments. The model for judging whether it can be divided into departments can be a classification model for outputting results of being able to be divided into departments or not being able to be divided into departments. The model for judging whether it can be divided into departments can be a machine learning model, specifically a deep learning model. Exemplarily, the model for judging whether it can be divided into departments includes the current logistic regression model and neural network model (Neural Network, NN). The training samples can include consultation information and the corresponding result of whether it can be divided into departments. Or, whether it can be judged by department can be as follows: according to a pre-configured template including various parameters, detect whether the consultation information includes the content corresponding to each parameter, and determine the result of whether it can be divided into departments according to the parameters lacking content. Exemplarily, the parameters lacking content include parameter A and parameter B, and it is determined that the result of whether it can be divided into departments is that it cannot be divided into departments; the parameter lacking content includes parameter A, and it is determined that the result of whether it can be divided into departments is that it can be divided into departments.

[0033] S102, determine the target consultation department according to the result of whether it can be divided into departments and the said consultation information.

[0034] The target consultation department can refer to the department to which the consultation user belongs. The target consultation department is used for the consultation user to select a doctor under this department for registration and conduct disease diagnosis.

[0035] In the case where the result of whether it can be divided into departments is that it can be divided into departments, determine the target consultation department according to the consultation information; in the case where the result of whether it can be divided into departments is that it cannot be divided into departments, prompt the user to add information, update the consultation information, and determine the target consultation department based on the updated consultation information. Or, continue to judge whether it can be divided into departments based on the updated consultation information, and in the case of not being able to be divided into departments, prompt the user to add information, continuously update the consultation information until the result of whether it can be divided into departments is that it can be divided into departments, and then determine the target consultation department according to the currently updated consultation information.

[0036] In the case where the result of whether it can be divided into departments is that it can be divided into departments, determining the target consultation department according to the consultation information can be inputting the consultation information into a pre-trained department division model to obtain the department division result output by the department division model, that is, determining the target consultation department, where the department division model can be a neural network model; or pre-establishing the corresponding relationship between the consultation information and the consultation department, and querying the target consultation department corresponding to the consultation information.

[0037] In the prior art, during the online consultation process, the department division result is determined according to the consultation information input by the patient. However, if the consultation information provided by the patient is vague, unclear or incomplete, etc., it will lead to inaccurate department division results.

[0038] According to the technical solution of the present disclosure, by determining whether the consultation information can be classified, and based on the result of the determination of whether it can be classified and the consultation information, the target consultation department is determined. When the consultation information can be classified, the target consultation department can be determined, improving the accuracy of the consultation information, making the consultation information clearer, and thus improving the accuracy rate of classification.

[0039] Figure 2 FIG. 4 is a flowchart of another consultation method disclosed according to an embodiment of the present disclosure, which is further optimized and extended based on the above technical solution and can be combined with each of the above optional embodiments. Determining the target consultation department according to the result of the determination of whether it can be classified and the consultation information is specifically: when the result of the determination of whether it can be classified is that it cannot be classified, performing human-computer interaction with the consultation user to obtain interaction information and update the consultation information; determining the target consultation department according to the updated consultation information.

[0040] S201, obtain the consultation information of the consultation user and determine whether the consultation information can be classified.

[0041] S202, when the result of the determination of whether it can be classified is that it cannot be classified, perform human-computer interaction with the consultation user to obtain interaction information and update the consultation information.

[0042] The result of the determination of whether it can be classified being that it cannot be classified indicates that the consultation information is insufficient to determine the target consultation department. Performing human-computer interaction with the consultation user is used to obtain more and more detailed consultation information from the consultation user to enrich the consultation information. Specifically, the human-computer interaction process is: providing questions to the consultation user, and the consultation user providing answers to the questions. The interaction information is used to update the consultation information. The interaction information is preset questions and the answers input by the consultation user for the preset questions. Updating the consultation information may mean adding the interaction information to the consultation information and correcting the consultation information.

[0043] In fact, the interaction information includes preset questions and corresponding answer texts, which usually do not conform to the user's language habits, while the inquiry information is usually the information input in the user's language habits. Updating the inquiry information according to the interaction information may refer to converting the interaction information into information corresponding to the language type of the inquiry information and fusing it with the inquiry information to update the inquiry information. Exemplarily, updating the inquiry information may include: inputting the interaction information into a pre-trained text generation model to obtain the inquiry text output by the text generation model, and fusing the inquiry text with the inquiry information, where the text generation model may be a neural network model for converting the initial information into information in the user's language habits. Or, updating the inquiry information may include: based on a preset text template, extracting keywords from the interaction information and adding them to the corresponding positions of the text template to generate an inquiry text and fuse it with the inquiry information. Among them, when the inquiry information is text, fusion means splicing the inquiry text with the inquiry information; when the inquiry information is non-text, the inquiry text can be converted into information of the type corresponding to the inquiry information and spliced with the inquiry information. For example, the interaction information is: the past disease is diabetes. The inquiry information is: I have had a headache for 3 consecutive days. The updated inquiry information is: I have diabetes and have had a headache for 3 consecutive days.

[0044] Optionally, the human-computer interaction with the inquiring user to obtain interaction information includes: obtaining question information and performing human-computer interaction with the inquiring user according to the question information; obtaining the answer information provided by the inquiring user based on the question information; and determining the interaction information according to the question information and the answer information.

[0045] The question information is used to provide to the inquiring user and obtain the answer information provided by the inquiring user to determine the interaction information. The answer information refers to the content replied by the inquiring user to the question information. At least one question can be selected from multiple preset questions as the question information and provided to the inquiring user respectively, and correspondingly, the answer to each question is obtained as the answer information. The question information may include multiple associated questions. When the inquiring user inputs answer information for a question, the question associated with this question is selected as the next question. For example, question 1 is: Do you have a fever? The associated question 2 is: Do you have symptoms of a sore throat? Or, it can also be set that there are multiple associated question information for the question information, and the next question is determined from multiple associated questions according to the answer information. No limitation is made on this.

[0046] Determine the interaction information based on at least one question and the corresponding answer. It can be understood that some questions are yes / no judgment type questions. When the question and the answer are combined, the symptoms or past diseases of the inquiring user can be determined. For example, the question is: Do you have the symptom of headache? The answer is: Yes. At this time, the information provided by the inquiring user is that there is a symptom of headache. If only based on the answer, the true meaning of the inquiring user cannot be determined.

[0047] Through pre-configured questions, conduct human-computer interaction with the inquiring user, obtain the answer information provided by the inquiring user for the question information, and determine the interaction information from the question information and the answer information, so as to accurately and completely obtain the true inquiring content and true past diseases of the inquiring user, etc. Update the inquiring information based on the interaction information, enrich the content of the inquiring information, improve the representativeness of the inquiring information, and thus improve the accuracy of department classification.

[0048] Optionally, the obtaining of the question information includes at least one of the following: query the corresponding associated department according to the inquiring information, and determine the corresponding question information according to the accompanying symptoms of the associated department; query the corresponding target scenario according to the inquiring information, and query the question information corresponding to the target scenario according to the corresponding relationship between the scenario and the question information; and generate the question information based on a pre-trained doctor-patient interaction model.

[0049] The associated department can refer to the department that can diagnose and treat the inquiring information. The accompanying symptom can refer to the symptom that the associated department can diagnose and treat. Querying the corresponding associated department according to the inquiring information can be to analyze and identify the inquiring information, input the analysis result and the intention recognition result into a pre-established knowledge graph for indexing, obtain the associated department information, determine the associated department, and extract the accompanying symptoms of the associated department from the associated department information. Determining the corresponding question information according to the accompanying symptoms of the associated department can be to generate at least one confirmation question for the accompanying symptom as the question information. For example, use the accompanying symptoms corresponding to different associated departments as candidates, such as generating a question "Do you have any accompanying symptoms" to confirm with the inquiring user to obtain the corresponding answer.

[0050] Different scenarios can be configured corresponding to the situation where department classification cannot be performed. Multiple inquiring information can be collected in advance, and the analysis result and intention recognition result of each inquiring information can be obtained, and the inquiring information can be classified, either manually or automatically. Each category is defined as a scenario. Querying the corresponding target scenario according to the inquiring information can be to determine the type of the inquiring information, and specifically, a clustering algorithm or a classification model, etc. can be used. Manually edit the question information involved in the inquiry process of different scenarios, establish the corresponding relationship between the scenario and the question information, and select the question information involved in different inquiry processes through the matching of the inquiring information and the scenario.

[0051] The first two methods determine the corresponding question information based on the consultation information. The last method generates questions based on a doctor-patient interaction model, and there is no corresponding relationship between the generated questions and the consultation information. Historical doctor-patient conversations can be collected in advance, with the doctor's questions as the input and the patient's responses as the output to train the doctor-patient interaction model. The doctor-patient interaction model learns the doctor's questioning logic and generates candidates, i.e., questions, as question information. Exemplarily, the doctor-patient interaction model includes an end-to-end pre-trained dialogue generation model (Plato) based on the latent space, a sequence-to-sequence model (Seq2Seq), or a feature generation pre-trained model (DiagGPT), etc.

[0052] At least one of these methods can be selected to generate questions. In the case of using at least two methods, the question information obtained by different methods can be fused, and similar and duplicate questions can be removed to obtain the question information.

[0053] Generating questions through multiple methods can enrich the content of the questions, increase the coverage of the questions, improve the representativeness of the questions, thereby improving the representativeness of the answers provided by the consultation users and enriching the consultation information.

[0054] S203. Determine the target consultation department according to the updated consultation information.

[0055] Determining the target consultation department according to the updated consultation information can add content to the consultation information to make the consultation information richer and more accurate, and determining the target consultation department can improve the accuracy of determining the consultation department.

[0056] Optionally, the determining the target consultation department according to the updated consultation information includes: inputting the updated consultation information into a pre-trained department classification model to obtain a first classification result; and / or determining a second classification result according to the updated consultation information in the pre-established correspondence between the standard information and the departments; determining the target consultation department according to the first classification result and the second classification result.

[0057] The first classification result is the target consultation department determined based on the department classification model. The second classification result is the target consultation department determined based on the retrieval method.

[0058] The department classification model is used to determine the target consultation department corresponding to the consultation information according to the consultation information. The training samples include the consultation information and the corresponding consultation department. For example, the department classification model is a pre-trained neural network model. Exemplarily, the department classification model includes a pre-trained model and a softmax model, a pre-trained model and a Text Convolutional Neural Networks (Textcnn), or a pre-trained model and an LSTM model, etc.

[0059] The standard information is used for similarity retrieval with the consultation information to determine the target consultation department corresponding to the consultation information. The standard information can be information that abstractly defines the consultation information. The standard information can be understood as information obtained by classifying multiple pieces of consultation information, extracting common features in each class, and abstractly defining them as the standard information for that class. The corresponding relationship between the standard information and the department is used to determine the target consultation department corresponding to the consultation information. Through a retrieval method, the standard information corresponding to the consultation information is queried, and the target consultation department corresponding to the corresponding standard information is determined as the target consultation department corresponding to the consultation information. The retrieval method can be an Approximate Nearest Neighbor (ANN) method. Exemplarily, the standard information and the department can be manually constructed and stored in a database, and the ANN similarity index retrieval is performed between the consultation information and the standard information to obtain the department corresponding to the similar standard information.

[0060] In the case of updated consultation information, the aforementioned consultation information can be replaced by the updated consultation information.

[0061] In the case where only the first classification result is obtained and the second classification result is empty, the target consultation department is determined according to the first classification result and the second classification result. In fact, the first classification result is determined as the target consultation department. In the case where only the second classification result is obtained and the first classification result is empty, the target consultation department is determined according to the first classification result and the second classification result. In fact, the second classification result is determined as the target consultation department. In the case where the first classification result and the second classification result are obtained, the first classification result and the second classification result can be fused to determine the target consultation department. Among them, the fusion method can be to correct the first classification result based on the second classification result. The priority of the second classification result is higher than that of the first classification result, or the second classification result can also be directly determined as the target consultation department.

[0062] By calculating the first classification result and / or the second classification result, and determining the target consultation department according to the first classification result and the second classification result, multiple methods are used for department classification, and the classification results are fused to determine the target consultation department, which can improve the accuracy of department classification.

[0063] According to the technical solution of the present disclosure, when the result of the department classification judgment is that department classification is not possible, a human-computer interaction is carried out with the inquiring user to guide the inquiring user to more completely describe their own situation, obtain interaction information, and update the inquiring information, which can increase the content of the inquiring information, improve the integrity of the inquiring information, make the inquiring information more accurate, and determine the target inquiring department based on the updated inquiring information, which can improve the accuracy of the target inquiring department, thereby improving the accuracy of department classification.

[0064] Figure 3 It is a flowchart of another inquiring method disclosed according to an embodiment of the present disclosure, which is further optimized and extended based on the above technical solution and can be combined with each of the above optional implementation manners. The inquiring method is optimized to: obtain at least one alternative doctor information corresponding to the target inquiring department; sort each of the alternative doctor information according to the inquiring information and each of the alternative doctor information; determine the recommended doctor information according to the sorting result.

[0065] S301, obtain the inquiring information of the inquiring user and perform a department classification judgment on the inquiring information.

[0066] S302, determine the target inquiring department according to the department classification judgment result and the inquiring information.

[0067] S303, obtain at least one alternative doctor information corresponding to the target inquiring department.

[0068] The alternative doctor information may refer to the information of doctors affiliated with the target inquiring department.

[0069] The alternative doctor information is used to screen the recommended doctor information and provide it to the inquiring user to suggest that the inquiring user register specifically, and finally achieve the purpose of triage. Among them, the alternative doctor information is not limited to one hospital. It is possible to select alternative departments that are the same or similar to the target inquiring department from multiple hospitals, determine at least one alternative doctor affiliated with the alternative department, collect the information of each doctor in each alternative department, and form a corresponding relationship between the department and the doctor information. The doctor information may include at least one of the following: the professional department of the doctor, the diseases they are good at, the disease treatment methods, and the service hours, etc. A corresponding relationship between the doctor information and the department can be established according to the professional department in the doctor information. According to the target inquiring department, query the corresponding department and determine the corresponding doctor information as the alternative doctor information.

[0070] S304, sort each of the alternative doctor information according to the inquiring information and each of the alternative doctor information.

[0071] Based on the consultation information and the alternative doctor information, the matching degree between the consultation information and each alternative doctor information can be determined, and the alternative doctor information can be sorted according to the matching degree. In the case where department division is not possible, the alternative doctor information can be sorted according to the updated consultation information and the alternative doctor information. By updating the consultation information, the consultation intention of the consultation user is clarified, and the updated consultation information is copied and combined with each alternative doctor information respectively to obtain at least one input information, which is input into a pre-trained linear regression model to obtain the score output by the linear regression model, and the alternative doctor information in the corresponding input information is sorted according to the score. The linear regression model is used to simulate the linear relationship between the consultation information and the doctor information and the matching degree between them. Among them, the training samples include the consultation feature information extracted from the consultation information, the doctor feature information extracted from the doctor information, and the matching score between the consultation information and the doctor information. Among them, the feature information can be represented in the form of a vector. The consultation feature information can include the aforementioned parsing results and intention recognition results, etc. The doctor feature information can include keywords in the doctor information. Among them, the doctor feature information can be extracted from the doctor information based on a pre-trained text recognition model, or the doctor feature information can be extracted from the doctor information according to a preset template.

[0072] S305. According to the sorting result, determine the recommended doctor information.

[0073] The recommended doctor information is used to be recommended to the consultation user to suggest that the consultation user choose the recommended doctor for diagnosis and treatment, so as to achieve accurate triage and improve the diagnosis and treatment effect. According to the sorting result, it can be to screen out the alternative doctor information with a higher ranking and determine it as the recommended doctor information. Exemplarily, the first alternative doctor information can be determined as the recommended doctor information; or the 1st - 3rd alternative doctor information can be determined as the recommended doctor information. In addition, there are other situations, which are not specifically limited.

[0074] Optionally, the sorting of each alternative doctor information according to the consultation information and each alternative doctor information includes: querying the severity level corresponding to the consultation information in the pre-established correspondence between the standard information and the severity level; sorting each alternative doctor information according to the consultation information, the corresponding severity level and each alternative doctor information.

[0075] The standard information can be referred to the foregoing description. The severity refers to the severity of the condition of the user being questioned, and the severity can include severe and not severe; or it can also include extremely severe, severely severe, moderately severe, mildly severe, and not severe. The correspondence between the standard information and the severity is used to determine the severity corresponding to the questioned information. Through a retrieval method, the standard information corresponding to the questioned information is queried, and the severity corresponding to the corresponding standard information is determined as the severity corresponding to the questioned information. The retrieval method can be ANN. Exemplarily, the standard information and the severity can be manually constructed and stored in a database, and the ANN similarity index retrieval is performed between the questioned information and the standard information to obtain the severity corresponding to the similar standard information.

[0076] In the case where the department cannot be divided, by updating the questioned information, the intention of the user being questioned to seek medical treatment and the severity of the condition are clarified, and the updated questioned information and the severity are copied and combined with each alternative doctor information respectively to obtain at least one input information, which is input into a pre-trained linear regression model, the score output by the linear regression model is obtained, and according to the score, the alternative doctor information in the corresponding input information is sorted. Specifically, the severity is added to the questioned feature information, that is, the questioned feature information can include the foregoing parsing results, intention recognition results, and severity, etc. The questioned feature information is copied and combined with the doctor feature information of each alternative doctor information respectively to obtain at least one input information, which is input into a pre-trained linear regression model, the score output by the linear regression model is obtained, and according to the score, the alternative doctor information in the corresponding input information is sorted.

[0077] In addition, in the correspondence between the standard information and the severity, when querying the severity corresponding to the questioned information and the query result is empty, a doctor of a lower level can be pre-assigned to the user being questioned to conduct a medical inquiry and a preliminary diagnosis of the disease, and judge the severity of the condition to obtain the severity determined by the doctor; or, the severity input by the user being questioned can also be obtained. Among them, the lower-level doctor is relative to the higher-level doctor. The higher-level doctor can refer to an expert-level doctor; the lower-level doctor can be a clinic-level doctor; or, the higher-level doctor can refer to a doctor in a top-three hospital; the lower-level doctor can be a doctor in a second-class hospital.

[0078] By determining the severity of the user being questioned, the doctor can be recommended based on the severity of the condition of the user being questioned, and on the basis of department division, the hierarchical diagnosis and treatment can be further realized, the accuracy of triage can be improved, the recommended doctor can be provided, the accuracy of triage recommendation can be improved, and the user experience can be improved.

[0079] According to the technical solution of the present disclosure, by obtaining at least one alternative doctor information corresponding to the consultation information and screening out the recommended doctor information from the alternative doctor information based on the consultation information, medical resources can be reasonably utilized to provide a suitable doctor for the consultation user, achieve hierarchical diagnosis and treatment, and improve the accuracy of doctor assignment.

[0080] Figure 4 It is a flowchart of another consultation method disclosed according to an embodiment of the present disclosure and a specific application scenario of a consultation method.

[0081] S401, Obtain the consultation information of the consultation user.

[0082] S402, Determine whether the consultation information can be classified. If it can be classified, execute S403; otherwise, execute S404.

[0083] S403, Determine the target consultation department according to the consultation information and execute S405.

[0084] S404, Conduct human-computer interaction with the consultation user, obtain interaction information, and update the consultation information.

[0085] The user answers and splices and fuses with the consultation information to obtain the updated consultation information.

[0086] S405, Query the severity corresponding to the consultation information in the pre-established correspondence between standard information and severity, and determine whether the severity can be queried; if it can, execute S406; otherwise, execute S407.

[0087] S406, Assign a low-level doctor to the consultation user and obtain the severity through the inquiry and diagnosis of the low-level doctor.

[0088] S407, Determine the recommended doctor information as the recommended authoritative top-three hospital doctor according to the severity, the consultation information, and at least one alternative doctor information corresponding to the target consultation department.

[0089] Sort each of the alternative doctor information according to the consultation information, the corresponding severity, and each of the alternative doctor information. Determine the recommended doctor information according to the sorting result. In the case where the severity is severe, the recommended doctor information is the authoritative top-three hospital doctor in the target consultation department.

[0090] By pre-judging whether it is possible to classify departments, in the case where it is not possible to classify departments, through human-computer interaction, the consultation information is updated to guide the user to describe their basic situation more completely, which can increase the content of the consultation information, improve the integrity of the consultation information, make the consultation information more accurate, improve the accuracy of department classification, and by obtaining the severity of the consultation information and at least one alternative doctor information corresponding to the target consultation department, screening and recommending doctor information, and providing it to the user of the consultation, it is possible to rationally utilize medical resources, provide suitable doctors for the user of the consultation, achieve hierarchical diagnosis and treatment, and improve the accuracy of minute diagnosis.

[0091] According to an embodiment of the present disclosure, Figure 5 is a structural diagram of a consultation device in an embodiment of the present disclosure. The embodiment of the present disclosure is applicable to the situation where a user obtains a triage department corresponding to consultation information. The device is implemented by software and / or hardware and is specifically configured in an electronic device with certain data operation capabilities.

[0092] As Figure 5 shown, a consultation device 500 includes: a department classification judgment module 501 and a

[108] triage department determination module 502; wherein,

[0093] The department classification judgment module 501 is configured to obtain the consultation information of the user of the consultation and judge whether the consultation information can be classified into departments;

[0094] The triage department determination module 502 is configured to determine a target consultation department according to the result of the department classification judgment and the consultation information.

[0095] According to the technical solution of the present disclosure, by judging whether the consultation information can be classified into departments and determining the target consultation department according to the result of the department classification judgment and the consultation information, it is possible to determine the target consultation department in the case where the consultation information can be classified into departments, improve the accuracy of the consultation information, make the consultation information clearer, and thus improve the accuracy of department classification.

[0096] Further, the triage department determination module 502 includes: a consultation information update unit configured to, in the case where the result of the department classification judgment is that it is not possible to classify departments, perform human-computer interaction with the user of the consultation, obtain interaction information, and update the consultation information; a triage department re-determination unit configured to determine a target consultation department according to the updated consultation information.

[0097] Further, the consultation information update unit includes: a human-computer interaction subunit configured to obtain question information and perform human-computer interaction with the user of the consultation according to the question information; an answer information acquisition subunit configured to obtain answer information provided by the user of the consultation based on the question information; an interaction information determination subunit configured to determine interaction information according to the question information and the answer information.

[0098] Further, the human-computer interaction sub-unit is configured to perform at least one of the following: query a corresponding associated department according to the medical interview information, and determine corresponding problem information according to the accompanying symptoms of the associated department; query a corresponding target scenario according to the medical interview information, and query the problem information corresponding to the target scenario according to the corresponding relationship between the scenario and the problem information; and generate problem information based on a pre-trained doctor-patient interaction model.

[0099] Further, the medical interview device 500 further includes: a department doctor acquisition module, configured to acquire at least one alternative doctor information corresponding to the target medical interview department; a doctor ranking module, configured to rank each of the alternative doctor information according to the medical interview information and each of the alternative doctor information; and a doctor recommendation module, configured to determine recommended doctor information according to the ranking result.

[0100] Further, the doctor ranking module includes: a severity determination unit, configured to query the severity corresponding to the medical interview information in a pre-established correspondence between standard information and severity; and a severity ranking unit, configured to rank each of the alternative doctor information according to the medical interview information, the corresponding severity, and each of the alternative doctor information.

[0101] Further, the triage department determination module 502 includes: a model classification unit, configured to input the updated medical interview information into a pre-trained department classification model to obtain a first classification result; and / or a search classification unit, configured to determine a second classification result according to the updated medical interview information in a pre-established correspondence between standard information and departments; and a classification fusion unit, configured to determine the target medical interview department according to the first classification result and the second classification result.

[0102] The above-mentioned medical interview device can execute the medical interview method provided in any embodiment of the present disclosure, and has corresponding functional modules and beneficial effects for executing the medical interview method.

[0103] In the technical solution of the present disclosure, the processing of collection, storage, use, processing, transmission, provision, and disclosure of user personal information involved all comply with the provisions of relevant laws and regulations and do not violate public order and good customs.

[0104] According to an embodiment of the present disclosure, the present disclosure also provides an electronic device, a readable storage medium, and a computer program product.

[0105] Figure 6FIG. shows a schematic block diagram of an exemplary electronic device 600 that can be used to implement embodiments of the present disclosure. The electronic device is intended to represent various forms of digital computers, such as, for example, laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as, for example, personal digital processors, cellular telephones, smart phones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely exemplary and are not intended to limit the implementations of the present disclosure described and / or claimed herein.

[0106] As Figure 6 shown, the device 600 includes a computing unit 601 that can perform various appropriate actions and processes in accordance with a computer program stored in a read-only memory (ROM) 602 or a computer program loaded from a storage unit 608 into a random access memory (RAM) 603. In the RAM 603, various programs and data required for the operation of the device 600 can also be stored. The computing unit 601, the ROM 602, and the RAM 603 are connected to each other via a bus 604. An input / output (I / O) interface 605 is also connected to the bus 604.

[0107] A plurality of components in the device 600 are connected to the I / O interface 605, including: an input unit 606, such as a keyboard, a mouse, etc.; an output unit 607, such as various types of displays, speakers, etc.; a storage unit 608, such as a magnetic disk, an optical disk, etc.; and a communication unit 609, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 609 allows the device 600 to exchange information / data with other devices via a computer network such as the Internet and / or various telecommunication networks.

[0108] The computing unit 601 can be various general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of the computing unit 601 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various dedicated artificial intelligence (AI) computing chips, various computing units running machine learning model algorithms, a digital signal processor (DSP), and any suitable processor, controller, microcontroller, etc. The computing unit 601 executes the various methods and processes described above, such as the interrogation method or the interrogation method. For example, in some embodiments, the interrogation method or the interrogation method can be implemented as a computer software program tangibly embodied in a machine-readable medium, such as the storage unit 608. In some embodiments, part or all of the computer program can be loaded and / or installed onto the device 600 via the ROM 602 and / or the communication unit 609. When the computer program is loaded into the RAM 603 and executed by the computing unit 601, one or more steps of the interrogation method or the interrogation method described above can be executed. Alternatively, in other embodiments, the computing unit 601 can be configured to execute the interrogation method or the interrogation method by any other suitable means (e.g., by means of firmware).

[0109] Various embodiments of the systems and techniques described above in this document can be implemented in digital electronic circuitry, integrated circuit systems, field-programmable gate arrays (FPGA), application-specific integrated circuits (ASIC), application-specific standard products (ASSP), systems-on-a-chip (SOC), complex programmable logic devices (CPLD), computer hardware, firmware, software, and / or combinations thereof. These various embodiments can include: implemented in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which can be a dedicated or general-purpose programmable processor that receives data and instructions from a storage system, at least one input device, and at least one output device, and transmits the data and instructions to the storage system, the at least one input device, and the at least one output device.

[0110] The program code for implementing the methods of the present disclosure can be written in any combination of one or more programming languages. These program codes can be provided to a processor or controller of a general-purpose computer, a special-purpose computer, or other programmable data processing device, such that when the program codes are executed by the processor or controller, the functions / operations specified in the flowcharts and / or block diagrams are implemented. The program code can be executed entirely on the machine, partially on the machine, as a stand-alone software package partially on the machine and partially on a remote machine, or entirely on a remote machine or server.

[0111] In the context of this disclosure, a machine-readable medium can be a tangible medium that can contain or store a program for use by or in connection with an instruction execution system, apparatus, or device. A machine-readable medium can be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium can include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination of the foregoing. More specific examples of a machine-readable storage medium would include an electrical connection based on one or more wires, a portable computer diskette, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or Flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0112] To provide for interaction with a user, the systems and techniques described herein can be implemented on a computer having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and a pointing device (e.g., a mouse or a trackball) by which the user can provide input to the computer. Other kinds of devices can also be used to provide for interaction with the user; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including acoustic, speech, or tactile input).

[0113] The systems and techniques described herein can be implemented in a computing system that includes back-end components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes front-end components (e.g., a user computer having a graphical user interface or a web browser through which the user can interact with an implementation of the systems and techniques described herein), or a computing system that includes any combination of such back-end components, middleware components, or front-end components. The components of the system can be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: a local area network (LAN), a wide area network (WAN), and the Internet.

[0114] A computer system can include a client and a server. The client and the server are generally remote from each other and typically interact through a communication network. The relationship of client and server arises by virtue of computer programs running on the respective computers and having a client-server relationship to each other. The server can be a cloud server, a server of a distributed system, or a server incorporating a blockchain.

[0115] It should be understood that the various forms of processes shown above can be used, with steps reordered, added or deleted. For example, the steps described in the present disclosure can be executed in parallel, sequentially or in different orders, as long as the desired results of the technical solution disclosed in the present disclosure can be achieved, and no limitation is imposed herein.

[0116] The above specific embodiments do not constitute a limitation on the protection scope of the present disclosure. Those skilled in the art should understand that various modifications, combinations, sub - combinations and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present disclosure shall be included within the protection scope of the present disclosure.

Claims

1. A consultation method, comprising: Obtaining consultation information of a consultation user, inputting the consultation information into a pre-trained model for determining whether to classify departments, and obtaining a determination result of whether to classify departments output by the model for determining whether to classify departments; In the case that the determination result of whether to classify departments is that departments cannot be classified, performing human-computer interaction with the consultation user, obtaining interaction information, and updating the consultation information; wherein, the way of updating the consultation information includes: inputting the interaction information into a pre-trained text generation model, obtaining a consultation text output by the text generation model, and fusing the consultation text with the consultation information to obtain updated consultation information; Determining a target consultation department according to the updated consultation information, including: inputting the updated consultation information into a pre-trained department classification model to obtain a first classification result; determining a second classification result according to the updated consultation information in a pre-established correspondence between standard information and departments; determining a target consultation department according to the first classification result and the second classification result; Obtaining at least one piece of alternative doctor information corresponding to the target consultation department; wherein, the alternative doctor information is used to select an alternative department that is the same as or similar to the target consultation department from multiple hospitals, so as to determine at least one alternative doctor affiliated with the alternative department; Querying the severity level corresponding to the updated consultation information in a pre-established correspondence between standard information and severity level; Copying the updated consultation information and the severity level, respectively combining them with each piece of alternative doctor information to obtain at least one piece of input information, inputting the input information into a pre-trained linear regression model, obtaining a score output by the linear regression model, and sorting the alternative doctor information in the corresponding input information according to the score; Determining recommended doctor information according to the sorting result.

2. The method according to claim 1, wherein, The performing human-computer interaction with the consultation user and obtaining interaction information includes: Obtaining question information and performing human-computer interaction with the consultation user according to the question information; Obtaining answer information provided by the consultation user based on the question information; Determining interaction information according to the question information and the answer information.

3. The method according to claim 2, wherein The obtaining question information includes at least one of the following: Querying a corresponding associated department according to the consultation information, and determining corresponding question information according to the accompanying symptoms of the associated department; Querying a corresponding target scenario according to the consultation information, and querying the question information corresponding to the target scenario according to the correspondence between the scenario and the question information; And Generating question information based on a pre-trained doctor-patient interaction model.

4. A consultation device, comprising: A module for determining whether to classify departments, configured to obtain consultation information of a consultation user, input the consultation information into a pre-trained model for determining whether to classify departments, and obtain a determination result of whether to classify departments output by the model for determining whether to classify departments; A module for determining a triage department, including: an update unit for consultation information and a re-determination unit for triage departments; The medical history information update unit is used to, when the result of the department division judgment is that department division is not possible, perform human-computer interaction with the medical history user, obtain interaction information, and update the medical history information; wherein, the method of updating the medical history information includes: inputting the interaction information into a pre-trained text generation model to obtain the medical history text output by the text generation model, and fusing the medical history text with the medical history information to obtain the updated medical history information; The triage department re-determination unit is used to determine the target medical history department according to the updated medical history information; The triage department re-determination unit is specifically used to input the updated medical history information into a pre-trained department classification model to obtain a first classification result; determine a second classification result according to the updated medical history information in the pre-established correspondence between standard information and departments; and determine the target medical history department according to the first classification result and the second classification result; The alternative doctor information acquisition module is used to acquire at least one piece of alternative doctor information corresponding to the target medical history department; wherein, the alternative doctor information is used to select alternative departments that are the same as or similar to the target medical history department from multiple hospitals, so as to determine at least one alternative doctor affiliated with the alternative department; The doctor ranking module includes: a severity determination unit and a severity ranking unit; The severity determination unit is used to query the severity corresponding to the updated medical history information in the pre-established correspondence between standard information and severity; The severity ranking unit is used to copy the updated medical history information and the severity, and combine them with each piece of alternative doctor information respectively to obtain at least one piece of input information, input the input information into a pre-trained linear regression model, obtain the score output by the linear regression model, and rank the alternative doctor information in the corresponding input information according to the score; The doctor recommendation module is used to determine the recommended doctor information according to the ranking result; 5. The apparatus according to claim 4, wherein The medical history information update unit includes: The human-computer interaction sub-unit is used to obtain question information and perform human-computer interaction with the medical history user according to the question information; The answer information acquisition sub-unit is used to acquire the answer information provided by the medical history user based on the question information; The interaction information determination unit is used to determine the interaction information according to the question information and the answer information; 6. The apparatus according to claim 5, wherein, The human-computer interaction sub-unit is used for at least one of the following: querying the corresponding associated department according to the medical history information, and determining the corresponding question information according to the accompanying symptoms of the associated department; querying the corresponding target scenario according to the medical history information, and querying the question information corresponding to the target scenario according to the correspondence between the scenario and the question information; and generating question information based on a pre-trained doctor-patient interaction model; 7. An electronic device, comprising: At least one processor; And A memory communicatively connected to the at least one processor; wherein, The memory stores instructions that are executed by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to execute the interrogation method according to any one of claims 1-3.

8. A non-transitory computer-readable storage medium storing computer instructions, wherein, The computer instructions are used to cause the computer to execute the interrogation method according to any one of claims 1-3.

9. A computer program product, comprising a computer program which, when executed by a processor, implements the interrogation method according to any one of claims 1-3.

Images