Agent-based Expert Doctor Recommendation Method and System

Abstract

The present invention belongs to the technical field of doctor recommendation. A method and system for recommending expert doctors based on an agent are provided. The method includes: outputting a set of prompt questions to a patient and receiving a first response voice input by the patient in response to the set of prompt questions; using a semantic analysis plugin to extract several sets of first response semantic contents from the first response voice, and if the first quantity of the first response semantic contents is higher than a preset value, determining that the patient is a target patient; generating and outputting a set of inquiry questions to the target patient according to each first response semantic content and receiving a second response voice input by the target patient in response to the set of inquiry questions; extracting several sets of second response semantic contents from the second response voice, screening out a second quantity of expert doctors based on each set of second response semantic contents, and outputting them to the target patient. The present invention can solve the technical problem that it is difficult to recommend expert doctors when patients cannot accurately describe their own conditions.

Description

Technical Field

[0001] The present invention relates to the technical field of doctor recommendation, and more particularly, to an intelligent agent-based expert doctor recommendation method and system. Background Art

[0002] In order to improve the level of medical services for patients, there are now some products that can recommend expert doctors suitable for their conditions to assist patients in selecting appropriate expert doctors. These products generally require patients to input their own conditions, and then select expert doctors suitable for their conditions through preset matching and screening rules.

[0003] However, the above products require patients to be able to accurately describe their own conditions, which is difficult for some patients, especially those with insufficient language summary ability or insufficient understanding of their own conditions. Therefore, how to improve the usability of expert doctor recommendation is a technical problem that needs to be solved urgently at present. Summary of the Invention

[0004] In view of this, the present invention provides an intelligent agent-based expert doctor recommendation method, system, electronic device, computer storage medium, and computer program product to solve the above technical problems.

[0005] The present invention discloses an intelligent agent-based expert doctor recommendation method applied to an intelligent agent. The method includes the following steps:

[0006] Output a set of prompt questions to the patient and receive the first response voice input by the patient for the set of prompt questions;

[0007] Use a semantic analysis plugin to extract several sets of first response semantic contents from the first response voice. If the first quantity of the first response semantic contents is higher than a preset value, determine that the patient is a target patient; wherein, the semantic similarity between each set of the first response semantic contents is lower than a similarity threshold;

[0008] Generate and output a set of inquiry questions to the target patient based on each of the first response semantic contents, and receive the second response voice input by the target patient for the set of inquiry questions;

[0009] Use a semantic analysis plugin to extract several sets of second response semantic contents from the second response voice, screen out a second quantity of expert doctors based on each of the second response semantic contents, and output them to the target patient.

[0010] In some embodiments, the step of outputting a set of prompt questions to the patient and receiving the first response voice input by the patient for the set of prompt questions includes:

[0011] Determine whether the operation selection information of the patient is received. If not, execute the following step S0;

[0012] If so, determine whether the operation selection information is the prompt mode. If so, execute the following step S0; if not, do not execute the following step S0;

[0013] S0: Output a set of prompt questions to the patient and receive the first response voice input by the patient for this set of prompt questions.

[0014] In some embodiments, using the semantic analysis plugin to extract several sets of first response semantic contents from the first response voice, including:

[0015] Use the semantic analysis plugin to extract several response statements from the first response voice, input each of the response statements into the semantic aggregation model based on the large model, and the semantic aggregation model outputs multiple sets of the first response semantic contents; wherein, each set of the first response semantic contents contains multiple of the response statements;

[0016] Among them, the control relationship data between the disease type and the disease symptoms is embedded in the semantic aggregation model based on the large model.

[0017] In some embodiments, the determination that the patient is the target patient includes:

[0018] Extract multiple voice interruption node segments from the first response voice content, calculate the third quantity and the total duration of each voice interruption node segment, and evaluate the voice input continuity according to the third quantity and the total duration;

[0019] Determine the adjustment coefficient value according to the voice input continuity;

[0020] Calculate the first quantity of the first response semantic content output by the semantic aggregation model, multiply the first quantity by the adjustment coefficient value to obtain the adjusted second quantity;

[0021] If the second quantity of the patient is higher than the preset value, determine that the patient is the target patient.

[0022] In some embodiments, the generation of a set of inquiry questions according to each of the first response semantic contents and output to the target patient includes:

[0023] Determine the disease type corresponding to each of the first response semantic contents according to the control relationship data between the disease type and the disease symptoms, generate a set of the inquiry questions based on other disease symptoms corresponding to the disease type, and output them to the target patient.

[0024] In some embodiments, screening out a second quantity of expert doctors based on the second response semantic content of each group includes:

[0025] Performing semantic similarity calculation on the second response semantic content of each group and the medical skill description information corresponding to each expert doctor, and screening out a second quantity of expert doctors with a semantic similarity higher than the semantic similarity threshold.

[0026] The present invention also discloses an expert doctor recommendation system based on an intelligent agent, which is applied to the intelligent agent. The system includes a prompt module, a first analysis module, an inquiry module, and a second analysis module;

[0027] The prompt module is used to output a set of prompt questions to the patient and receive the first response voice input by the patient for this set of prompt questions;

[0028] The first analysis module is used to extract several groups of first response semantic content from the first response voice using a semantic analysis plugin. If the first quantity of the first response semantic content is higher than a preset value, it is determined that the patient is a target patient; wherein, the semantic similarity between each group of the first response semantic content is lower than the similarity threshold;

[0029] The inquiry module is used to generate and output a set of inquiry questions to the target patient according to each of the first response semantic content and receive the second response voice input by the target patient for this set of inquiry questions;

[0030] The second analysis module uses a semantic analysis plugin to extract several groups of second response semantic content from the second response voice, screens out a second quantity of expert doctors based on each group of the second response semantic content, and outputs them to the target patient.

[0031] The present invention also discloses an electronic device, including: at least one processor, a memory, and a computer program stored in the memory and executable on the at least one processor, and the processor executes the computer program to implement the method as described in any one of the foregoing.

[0032] The present invention also discloses a computer storage medium, and the computer-readable storage medium stores a computer program, and the computer program is executed by a processor to implement the method as described in any one of the foregoing.

[0033] The present invention also discloses a computer program product, which implements the method as described in any one of the foregoing when the computer program product runs on a terminal.

[0034] The present invention dynamically decides whether to guide the patient to describe their condition by asking questions based on the patient's accurate description of their own condition, which can solve the technical problem that it is difficult to recommend expert doctors when the patient cannot accurately describe their own condition, thereby providing a better medical experience for the patient. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the embodiments will be briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present invention and should not be regarded as limiting the scope. For those of ordinary skill in the art, other related drawings can be obtained based on these drawings without creative efforts.

[0036] Figure 1 is a schematic flowchart of a method for recommending expert doctors based on an agent disclosed in an embodiment of the present invention;

[0037] Figure 2 is a schematic flowchart of an implementation scheme for determining that the patient is a target patient disclosed in an embodiment of the present invention;

[0038] Figure 3 is a schematic structural diagram of a system for recommending expert doctors based on an agent disclosed in an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0039] The following specific embodiments illustrate the implementation manners of the present application. Those skilled in the art can easily understand other advantages and effects of the present application from the content disclosed in this specification. Obviously, the described embodiments are part of the embodiments of the present application, rather than all of them. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative efforts belong to the scope protected by the present application.

[0040] In addition, the technical features involved in different implementation manners of the present application described below can be combined with each other as long as they do not conflict with each other.

[0041] As Figure 1 shown, an embodiment of the present invention discloses a method for recommending expert doctors based on an agent, which is applied to an agent. The method includes the following steps:

[0042] Output a set of prompt questions to the patient and receive the first response voice input by the patient in response to the set of prompt questions;

[0043] Use a semantic analysis plugin to extract several groups of first response semantic contents from the first response voice. If the first quantity of the first response semantic contents is higher than a preset value, determine that the patient is the target patient; wherein, the semantic similarity between each group of the first response semantic contents is lower than the similarity threshold;

[0044] Generate and output a group of inquiry questions to the target patient according to each of the first response semantic contents, and receive the second response voice input by the target patient in response to this group of inquiry questions;

[0045] Use a semantic analysis plugin to extract several groups of second response semantic contents from the second response voice, screen out the second quantity of expert doctors based on each group of the second response semantic contents, and output them to the target patient.

[0046] In the solution of the present invention, the intelligent agent first outputs a group of prompt questions to the patient, and these prompt questions are used to prompt the patient to describe their own condition from at least one aspect such as disease name, symptoms, symptom duration, symptom severity, etc. The patient then inputs the first response voice to the intelligent agent in the way of voice input.

[0047] Then, the intelligent agent uses a semantic analysis plugin to extract several groups of first response semantic contents from the first response voice, and the semantic similarity between each group of the first response semantic contents is lower than the similarity threshold. If the first quantity of the first response semantic contents is higher than the preset value, it indicates that the patient cannot accurately describe their condition. At this time, the patient is determined to be the target patient and the strategy needs to be changed.

[0048] Next, the intelligent agent generates and outputs a group of inquiry questions to the target patient according to each of the first response semantic contents. The inquiry questions are that the intelligent agent directly asks the patient whether there are certain symptoms, and the patient then answers this group of inquiry questions one by one in the way of the second response voice, which is beneficial to accurately determine the true disease type of the patient.

[0049] Finally, the intelligent agent uses a semantic analysis plugin to extract several groups of second response semantic contents from the second response voice, screens out the second quantity of expert doctors based on each group of the response semantic contents, and outputs them to the target patient for the target patient to screen.

[0050] It can be seen that the present invention dynamically decides whether to guide the patient to describe their condition by asking questions according to the patient's accurate description of their own condition, which can solve the technical problem that it is difficult to recommend expert doctors when the patient cannot accurately describe their own condition, thereby providing a better medical experience for the patient.

[0051] It should be noted that the agent in the present invention can be embedded and installed in a web page, or can be embedded and installed in various mobile or fixed devices. Among them, mobile devices refer to smartphones, tablets, wearable devices, smart glasses, etc., and fixed devices refer to desktop computers, laptop computers and other devices.

[0052] In some embodiments, outputting a set of prompt questions to the patient and receiving a first response voice input by the patient in response to the set of prompt questions includes:

[0053] Determine whether operation selection information from the patient is received. If not, perform the following step S0;

[0054] If so, determine whether the operation selection information is the prompt mode. If so, perform the following step S0; if not, do not perform the following step S0;

[0055] S0: Output a set of prompt questions to the patient and receive a first response voice input by the patient in response to the set of prompt questions.

[0056] In this embodiment, the present invention sets up multiple disease input modes, namely the normal mode and the prompt mode. In the normal mode, the agent does not output prompt questions or inquiry questions to the patient, but the patient directly inputs a description voice of their own condition to the agent. This mode is suitable for patients who have a high degree of mastery of their own condition. In the prompt mode, the agent needs to output prompt questions and inquiry questions to the patient to guide the patient to clearly describe their own condition. This mode is suitable for patients who have a low degree of mastery of their own condition, such as patients who have not been diagnosed yet.

[0057] Among them, a mode selection button is configured in the web page or mobile device or fixed device in which the agent is embedded and installed. By detecting which button is selected, it can be determined whether the operation selection information is the prompt mode or the normal mode.

[0058] In some embodiments, using a semantic analysis plugin to extract several sets of first response semantic contents from the first response voice includes:

[0059] Use a semantic analysis plugin to extract several response statements from the first response voice, and input each of the response statements into a semantic aggregation model based on a large model. The semantic aggregation model outputs multiple sets of the first response semantic contents; wherein, each set of the first response semantic contents contains multiple of the response statements;

[0060] Among them, data on the control relationship between disease types and disease symptoms is embedded in the semantic aggregation model based on the large model.

[0061] In this embodiment, after receiving the first response voice input by the patient, the intelligent agent first uses the semantic analysis plug-in to extract several response sentences from the first response voice. At the same time, a semantic aggregation model based on a large model is pre-built and trained, and a set of small sample training data is used for fine-tuning training. The semantic aggregation model can cluster the multiple response sentences obtained by the above analysis with the assistance of the embedded disease type and disease symptom comparison relationship data, that is, attribute them to the corresponding disease type, thereby obtaining multiple groups of first response semantic content. The first number of these groups of first response semantic content represents the number of disease types described by the patient in the first response voice. Obviously, the more the first number, the more disease types are involved, which means that the patient's description of his own condition is less accurate, and vice versa, the description is more accurate.

[0062] It should be noted that the large models mentioned above refer to general large models based on advanced neural networks such as Transformer, such as GPT model, BERT model, etc. These models far exceed conventional machine learning models in terms of training data, model parameter volume, and computing resources, and can handle data from multiple tasks and fields, with good adaptability and flexibility.

[0063] In some embodiments, determining that the patient is a target patient comprises:

[0064] Extracting a plurality of voice interruption node segments from the first answer voice content, calculating a third number and a total duration of each voice interruption node segment, and evaluating the voice input continuity according to the third number and the total duration;

[0065] Determine an adjustment coefficient value according to the continuity of the speech input;

[0066] Calculating the first quantity of the first response semantic content output by the semantic aggregation model, and multiplying the first quantity by the adjustment coefficient value to obtain an adjusted second quantity;

[0067] If the second number of the patient is higher than a preset value, the patient is determined to be a target patient.

[0068] In this embodiment, according to the aforementioned scheme, when the number of groups of the first response semantic content in the patient's first response speech, i.e., the first quantity, is higher than a preset value, it can be determined that the patient cannot accurately describe his or her condition, and at this time, the patient is determined to be a target patient who needs to be induced to describe his or her condition by asking questions. However, some patients may need to be diagnosed and treated for multiple diseases at a time, or their diseases do involve more symptoms, and these symptoms are similar to other types of diseases. In this case, the aforementioned scheme is likely to mistakenly determine the patient as a target patient.

[0069] In view of the above situation, as Figure 2 shown, the present invention extracts a plurality of speech interruption node segments (i.e., node segments without input of content with actual meaning, such as silent node segments, node segments similar to "um... um...") from the first response speech content, calculates the third quantity and the total duration of each speech interruption node segment, and evaluates the speech input continuity based on the third quantity and the total duration. Obviously, the speech input continuity can characterize the proficiency of the patient in inputting the disease description information. If the patient has a higher speech input continuity, it means that the patient has a higher understanding of his own condition, and the probability that he belongs to the target patient is lower. At this time, the adjustment coefficient value is set to a smaller value (such as 0.8, 0.9, 1.0), and this adjustment coefficient value is used to adjust the first quantity of the first response semantic content output by the foregoing semantic aggregation model to a smaller second quantity; on the contrary, if the patient has a lower speech input continuity, it means that the patient has a lower understanding of his own condition, and the probability that he belongs to the target patient is higher. At this time, the adjustment coefficient value is set to a larger value (such as 1.1, 1.2), and this adjustment coefficient value is used to adjust the first quantity of the first response semantic content output by the foregoing semantic aggregation model to a larger second quantity. Obviously, the adjustment coefficient value and the speech input continuity are negatively correlated. Finally, by comparing the second quantity with a preset value, it can be determined whether the patient is a target patient.

[0070] In addition, there is a negative correlation between the speech input continuity and the obtained third quantity and total duration. Moreover, the present invention does not specifically limit the conversion formula or comparison table of all the above negative correlation relationships.

[0071] In some embodiments, generating and outputting a set of inquiry questions to the target patient according to each of the first response semantic contents includes:

[0072] Determining the disease type corresponding to each of the first response semantic contents according to the control relationship data between the disease type and the disease symptoms, generating a set of the inquiry questions based on other disease symptoms corresponding to the disease type, and outputting them to the target patient.

[0073] In this embodiment, at least one disease symptom is included in the first response semantic content parsed from the first response semantic content. By comparing this disease symptom with the control relationship number between the disease type and the disease symptoms, the disease type (which may be multiple) to which it belongs can be determined. Then, based on other symptoms of this disease type included in the control relationship number between the disease symptoms and the disease type and the disease symptoms, a set of inquiry questions is generated, and the patient is questioned accordingly.

[0074] For example, the first response semantic content contains the content of "stomachache". Based on the matching of the number of control relationships between disease symptoms, disease types, and disease symptoms, multiple diseases such as gastritis, appendicitis, and gallstones are obtained. Then, other symptoms corresponding to gastritis, appendicitis, and gallstones are respectively determined from the number of control relationships between disease symptoms, disease types, and disease symptoms. Accordingly, the "vomiting symptom", "flatulence symptom", etc. of gastritis, the "metastatic right lower abdominal pain", "low fever", etc., and the "fever / chill", "biliary colic", etc. are generated and asked of the patient in sequence. By asking, several diseases associated with the symptom of "stomachache" can be excluded one by one, and finally appendicitis is confirmed.

[0075] In some embodiments, the screening of the second number of expert doctors based on each group of the second response semantic content includes:

[0076] Performing semantic similarity calculation on each group of the second response semantic content and the medical ability description information corresponding to each expert doctor, and screening out the second number of expert doctors with a semantic similarity higher than the semantic similarity threshold.

[0077] In this embodiment, medical ability description information is established in advance for each expert doctor, which includes the department, diseases good at treating, rank, etc. Performing semantic similarity calculation on each group of the second response semantic content obtained from the second response voice content and the medical ability description information corresponding to each expert doctor, and determining the expert doctor with a semantic similarity higher than the semantic similarity threshold as the expert doctor suitable for the target patient and screening it out. Among them, the second number of the selected suitable expert doctors can be preset or determined by the patient independently.

[0078] As Figure 3 shown, an agent-based expert doctor recommendation system is also disclosed in an embodiment of the present invention, which is applied to an agent. The system includes a prompt module, a first analysis module, an inquiry module, and a second analysis module;

[0079] The prompt module is used to output a set of prompt questions to the patient and receive the first response voice input by the patient for the set of prompt questions;

[0080] The first analysis module is used to extract several groups of first response semantic content from the first response voice using a semantic analysis plugin. If the first number of the first response semantic content is higher than a preset value, it is determined that the patient is a target patient; among them, the semantic similarity between each group of the first response semantic content is lower than the similarity threshold;

[0081] The query module is configured to generate a set of query questions based on each of the first response semantic contents and output the set of query questions to the target patient, and receive second response speech input by the target patient in response to the set of query questions.

[0082] The second analysis module uses a semantic analysis plugin to extract several sets of second response semantic contents from the second response speech, screens out a second quantity of expert doctors based on each set of the second response semantic contents, and outputs the expert doctors to the target patient.

[0083] An embodiment of the present invention also discloses an electronic device, including: at least one processor, a memory, and a computer program stored in the memory and executable on the at least one processor, where the processor executes the computer program to implement the method as described in the foregoing embodiment.

[0084] An embodiment of the present invention also discloses a computer storage medium storing a computer program, where the computer program is executed by a processor to implement the method as described in the foregoing embodiment.

[0085] An embodiment of the present invention also discloses a computer program product, which implements the method as described in the foregoing embodiment when running on a terminal.

[0086] Each component embodiment of the present invention may be implemented in hardware, or in software modules running on one or more processors, or in a combination thereof. Those skilled in the art should understand that a microprocessor or a digital signal processor (DSP) can be used in practice to implement some or all of the functions of some or all of the components in the device according to the embodiments of the present invention. The present invention can also be implemented as a device or apparatus program (such as a computer program and a computer program product) for executing part or all of the methods described herein. Such a program for implementing the present invention can be stored on a computer-readable medium, or can be in the form of one or more signals. Such signals can be downloaded from an Internet website, or provided on a carrier signal, or provided in any other form.

[0087] As used herein, the terms "one embodiment", "an embodiment", or "one or more embodiments" mean that a particular feature, structure, or characteristic described in connection with the embodiment is included in at least one embodiment of the present invention. In addition, note that the examples of the phrase "in one embodiment" herein do not necessarily all refer to the same embodiment.

[0088] In the description provided herein, numerous specific details are set forth. However, it will be understood that embodiments of the invention may be practiced without these specific details. In some instances, well-known methods, structures and techniques have not been shown in detail so as not to obscure an understanding of this description.

[0089] It should be noted that the above embodiments are illustrative of the invention and not restrictive thereof, and that alternative embodiments may be devised by those skilled in the art without departing from the scope of the appended claims. In the claims, any reference signs placed between parentheses shall not be construed as limiting the claim. The word "comprising" does not exclude the presence of elements or steps not listed in a claim. The word "a" or "an" preceding an element does not exclude the presence of a plurality of such elements. The invention may be implemented by means of hardware comprising several distinct elements, and by means of a suitably programmed computer. In a unitary claim listing several means, several of these means may be embodied by one and the same item of hardware. The use of the words first, second, and third, etc. does not denote any order. These words may be interpreted as names.

[0090] Furthermore, it should be noted that the language used in this specification has been principally selected for readability and instructional purposes, and not for the purpose of interpreting or limiting the subject matter of the invention. Accordingly, many modifications and variations will be apparent to those of ordinary skill in the art without departing from the scope and spirit of the appended claims. For the purposes of the present invention, the disclosure herein is illustrative, not restrictive, and the scope of the invention is defined by the appended claims.

Claims

1. An agent-based expert doctor recommendation method, applied to an agent, characterized in that, The method includes the following steps: Output a set of prompt questions to the patient and receive the first response voice input by the patient for this set of prompt questions; Use a semantic analysis plugin to extract several sets of first response semantic contents from the first response voice. If the first quantity of the first response semantic contents is higher than a preset value, determine that the patient is a target patient who cannot accurately describe their condition; wherein, the semantic similarity between each set of the first response semantic contents is lower than a similarity threshold; Generate and output a set of inquiry questions to the target patient according to each of the first response semantic contents. The inquiry question is that the intelligent agent directly asks the patient whether there are certain symptoms, and receive the second response voice input by the target patient for this set of inquiry questions; Use a semantic analysis plugin to extract several sets of second response semantic contents from the second response voice, screen out a second quantity of expert doctors based on each set of the second response semantic contents, and output them to the target patient; The determination that the patient is a target patient who cannot accurately describe their condition includes: Extract multiple voice interruption node segments from the first response voice content, calculate the third quantity and total duration of each voice interruption node segment, and evaluate the voice input continuity according to the third quantity and the total duration; the voice input continuity is negatively correlated with both the third quantity and the total duration; Determine an adjustment coefficient value according to the voice input continuity; the adjustment coefficient value is negatively correlated with the voice input continuity; Calculate the first quantity of the first response semantic contents output by the semantic aggregation model, and multiply the first quantity by the adjustment coefficient value to obtain an adjusted second quantity; If the second quantity of the patient is higher than the preset value, determine that the patient is a target patient; The output of a set of prompt questions to the patient and the reception of the first response voice input by the patient for this set of prompt questions includes: Judge whether the operation selection information of the patient is received. If not, execute the following step S0; If so, judge whether the operation selection information is the prompt mode. If so, execute the following step S0; if not, do not execute the following step S0; S0: Output a set of prompt questions to the patient and receive the first response voice input by the patient for this set of prompt questions.

2. The method for recommending expert doctors based on an agent according to claim 1, wherein: Using a semantic analysis plugin to extract several sets of first response semantic contents from the first response voice includes: Use a semantic analysis plugin to extract several response statements from the first response voice, and input each response statement into a semantic aggregation model based on a large model. The semantic aggregation model outputs multiple sets of the first response semantic contents; wherein, each set of the first response semantic contents contains multiple of the response statements; Among them, the disease type and disease symptom comparison relationship data is embedded in the semantic aggregation model based on the large model.

3. The method for recommending expert doctors based on an agent according to claim 1, wherein: Generating and outputting a set of inquiry questions to the target patient according to each of the first response semantic contents includes: Determine the disease type corresponding to each of the first response semantic contents according to the data of the correspondence between disease types and disease symptoms, generate a set of the inquiry questions based on other disease symptoms corresponding to the disease type, and output them to the target patient.

4. The method for recommending expert doctors based on an agent according to claim 3, characterized in that: Screen out a second quantity of expert doctors based on each group of the second response semantic contents, including: Perform semantic similarity calculation on each group of the second response semantic contents and the medical ability description information corresponding to each expert doctor, and screen out a second quantity of expert doctors with a semantic similarity higher than the semantic similarity threshold.

5. An agent-based expert doctor recommendation system applied to an agent, the system being based on the method according to any one of claims 1-4, characterized in that: Including a prompt module, a first analysis module, an inquiry module, and a second analysis module; The prompt module is configured to output a set of prompt questions to the patient and receive the first response voice input by the patient for the set of prompt questions; The first analysis module is configured to use a semantic analysis plugin to extract several groups of first response semantic contents from the first response voice. If the first quantity of the first response semantic contents is higher than a preset value, determine that the patient is the target patient; wherein, the semantic similarity between each group of the first response semantic contents is lower than the similarity threshold; The inquiry module is configured to generate and output a set of inquiry questions to the target patient according to each of the first response semantic contents and receive the second response voice input by the target patient for the set of inquiry questions; The second analysis module uses a semantic analysis plugin to extract several groups of second response semantic contents from the second response voice, screens out a second quantity of expert doctors based on each group of the second response semantic contents, and outputs them to the target patient.

6. An electronic device, comprising: At least one processor, a memory, and a computer program stored in the memory and executable on the at least one processor, wherein: the processor executes the computer program to implement the method according to any one of claims 1-4.

7. A computer storage medium storing a computer program, characterized in that: The computer program is executed by the processor to implement the method according to any one of claims 1-4.

8. A computer program product, characterized in that: When the computer program product runs on the terminal, the method according to any one of claims 1-4 is implemented.

Images