CN112035619A

CN112035619A - Medical questionnaire screening method, device, equipment and medium based on artificial intelligence

Info

Publication number: CN112035619A
Application number: CN202010899529.7A
Authority: CN
Inventors: 王明; 罗锐; 张云婵
Original assignee: Kangjian Information Technology Shenzhen Co Ltd
Current assignee: Kangjian Information Technology Shenzhen Co Ltd
Priority date: 2020-08-31
Filing date: 2020-08-31
Publication date: 2020-12-04

Abstract

The present application relates to artificial intelligence, and in particular, to a method, an apparatus, a device, and a medium for screening medical questionnaires based on artificial intelligence. The method comprises the following steps: acquiring a medical inquiry sheet to be marked; marking the main complaint content and the dialogue content in the medical inquiry list to be marked respectively to obtain a main complaint marking result and a dialogue marking result; determining a diagnosis condition according to a diagnosis result in the medical inquiry list to be annotated; and screening the chief complaint labeling results and the dialogue labeling results according to the diagnosis conditions to obtain a standard medical questionnaire. By adopting the method, the screening efficiency of the medical inquiry sheet can be improved. In addition, the invention also relates to a block chain technology, and privacy data such as preset keywords in a preset dictionary base can be stored in the block chain nodes.

Description

Medical questionnaire screening method, device, equipment and medium based on artificial intelligence

Technical Field

The application relates to the technical field of artificial intelligence, in particular to a medical questionnaire screening method, a medical questionnaire screening device, medical questionnaire screening equipment and a medical questionnaire screening medium based on artificial intelligence.

Background

With the rapid development of internet medical technology, intelligent medical treatment is the key development direction of internet medical treatment, and a standard test questionnaire is required for verifying the accuracy of an intelligent medical treatment algorithm and the test accuracy.

In the conventional technology, the standard test questionnaire is obtained by manually checking and verifying the questionnaire obtained on line, so that the checking efficiency of the standard questionnaire is low.

Disclosure of Invention

In view of the above, there is a need to provide a medical questionnaire screening method, apparatus, device and medium based on artificial intelligence, which can improve the efficiency of medical questionnaire screening.

A medical questionnaire screening method based on artificial intelligence comprises the following steps:

acquiring a medical inquiry sheet to be marked;

marking the main complaint content and the dialogue content in the medical inquiry list to be marked respectively to obtain a main complaint marking result and a dialogue marking result;

determining a diagnosis condition according to a diagnosis result in the medical inquiry list to be annotated;

and screening the chief complaint labeling results and the dialogue labeling results according to the diagnosis conditions to obtain a standard medical questionnaire.

In one embodiment, the annotation mode of the dialog annotation result comprises the following steps:

matching the conversation content in the medical questionnaire to be annotated with preset keywords in a preset dictionary library, and extracting the successfully matched preset keywords as conversation keywords; storing preset keywords in a preset dictionary library in a block chain;

extracting positive and negative directions of the conversation keywords from the conversation content, and configuring conversation labeling directions for the conversation keywords according to the positive and negative directions;

and obtaining a conversation labeling result according to the conversation keywords and the conversation labeling direction.

In one embodiment, the preset dictionary library comprises a plurality of preset keywords, and each preset keyword comprises a preset main keyword and a preset sub keyword; matching the conversation content in the medical questionnaire to be labeled with preset keywords in a preset dictionary library, and extracting the successfully matched preset keywords as conversation keywords, wherein the method comprises the following steps:

matching the conversation content in the medical questionnaire to be annotated with preset keywords in a preset dictionary library;

and when the conversation content is successfully matched with the preset sub-keywords in the preset keywords, acquiring preset main keywords corresponding to the successfully matched preset sub-keywords, and extracting the preset main keywords as the conversation keywords.

In one embodiment, matching the dialogue content in the medical questionnaire to be annotated with preset keywords in a preset dictionary library, and extracting the successfully matched preset keywords as the dialogue keywords comprises:

natural language processing is carried out on the conversation content to obtain keywords to be matched;

matching the keywords to be matched with preset keywords in a preset dictionary library;

and extracting the successfully matched preset keywords as the conversation keywords.

In one embodiment, the screening of the chief complaint labeling result and the dialogue labeling result according to the diagnosis condition to obtain a standard medical questionnaire comprises:

verifying the conversation keywords in the conversation labeling result, the conversation labeling directions corresponding to the conversation keywords, the chief complaint keywords in the chief complaint labeling result and the chief complaint labeling directions corresponding to the chief complaint keywords according to the diagnosis conditions;

and extracting the medical inquiry sheet to be marked which passes the verification as a standard medical inquiry sheet.

In one embodiment, after the standard medical questionnaire is obtained by screening the chief complaint annotation result and the dialogue annotation result according to the diagnosis condition, the method further comprises:

in the interactive interface, manually checking the fields to be checked in each standard medical questionnaire;

and deleting the standard questionnaire which fails in verification, extracting the field to be verified which fails in verification from the standard medical questionnaire which fails in verification, and updating the preset keywords in the preset dictionary base according to the field to be verified which fails in verification.

A precision testing method of medical inquiry algorithm based on artificial intelligence comprises the following steps:

acquiring a medical inquiry algorithm to be tested;

processing a standard medical inquiry sheet obtained according to an artificial intelligence-based medical inquiry sheet screening method by using a to-be-tested medical inquiry algorithm to obtain a test diagnosis result;

comparing the test diagnosis result with the real diagnosis result in the standard medical questionnaire;

and when the comparison is passed, judging that the precision of the medical inquiry algorithm to be tested meets the test requirement.

A medical questionnaire screening device based on artificial intelligence, the device includes:

the acquisition module is used for acquiring a medical inquiry sheet to be labeled;

the marking module is used for marking the main complaint content and the dialogue content in the medical inquiry list to be marked respectively to obtain a main complaint marking result and a dialogue marking result;

the determining module is used for determining diagnosis conditions according to the diagnosis result in the medical inquiry list to be annotated;

and the screening module is used for screening the chief complaint labeling result and the dialogue labeling result according to the diagnosis conditions to obtain a standard medical questionnaire.

A computer device comprising a memory storing a computer program and a processor implementing the steps of the method when the processor executes the computer program.

A computer-readable storage medium, on which a computer program is stored which, when being executed by a processor, carries out the steps of the above-mentioned method.

According to the precision testing method of the medical inquiry algorithm based on artificial intelligence, a medical inquiry list to be annotated is obtained, the main complaint content of a patient in the medical inquiry list to be annotated is automatically annotated in a server to obtain a main complaint annotation result, and the dialogue content of the patient in the medical inquiry list to be annotated is automatically annotated to obtain a dialogue annotation result. And then, determining diagnosis conditions according to the diagnosis results in the medical questionnaire to be annotated, and automatically checking the chief complaint annotation result and the dialogue annotation result according to the diagnosis conditions so as to extract the medical questionnaire to be annotated passing the checking as a standard medical questionnaire. The medical questionnaire is automatically marked and screened through a computer in the whole process of screening the medical questionnaire, manual screening is not needed, and the screening efficiency of the medical questionnaire to be marked is greatly improved.

Drawings

Fig. 1 is an application scenario diagram of a precision testing method of a medical inquiry algorithm based on artificial intelligence in an embodiment;

FIG. 2 is a schematic flow chart illustrating a method for testing the accuracy of an artificial intelligence-based medical interrogation algorithm in one embodiment;

FIG. 3 is an interface diagram for labeling an questionnaire to be labeled according to an embodiment;

FIG. 4 is an interface diagram of an interview order verification system provided in one embodiment;

FIG. 5 is a flow chart illustrating a method for testing the accuracy of a medical interrogation algorithm in accordance with one embodiment;

FIG. 6 is a block diagram of an embodiment of an artificial intelligence based medical questionnaire screening apparatus;

FIG. 7 is a block diagram of an apparatus for testing the accuracy of a medical interrogation algorithm based on artificial intelligence in one embodiment;

FIG. 8 is a diagram illustrating an internal structure of a computer device according to an embodiment.

Detailed Description

In order to make the objects, technical solutions and advantages of the present application more apparent, the present application is described in further detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the present application and are not intended to limit the present application.

The medical questionnaire screening method based on artificial intelligence can be applied to the application environment shown in fig. 1. Wherein the terminal 110 communicates with the server 120 through a network. The terminal 110 may upload the medical questionnaire to be annotated to the server 120; after the server 120 obtains the medical questionnaire to be annotated, the chief complaint content and the dialogue content in the medical questionnaire to be annotated are respectively annotated to obtain a chief complaint annotation result and a dialogue annotation result; determining a diagnosis condition according to a diagnosis result in the medical inquiry list to be annotated; and screening the chief complaint labeling results and the dialogue labeling results according to the diagnosis conditions to obtain a standard medical questionnaire. The terminal 110 may be, but not limited to, various personal computers, notebook computers, smart phones, tablet computers, and portable wearable devices, and the server 120 may be implemented by an independent server or a server cluster formed by a plurality of servers.

In one embodiment, as shown in fig. 2, a flow chart of a medical questionnaire screening method based on artificial intelligence is provided, which is illustrated by taking the method as an example of being applied to the server 120 in fig. 1, and in other embodiments, the method can also be applied to a terminal, and the method includes the following steps:

step 202, acquiring a medical inquiry sheet to be annotated.

The medical inquiry sheet to be labeled is real medical inquiry data acquired from an online manner. If the patient performs medical inquiry in the online medical inquiry platform, the doctor analyzes the medical inquiry of the patient to obtain a diagnosis result, and then a medical inquiry sheet to be annotated is generated in the server according to the medical inquiry of the patient, the inquiry and diagnosis process of the doctor and the diagnosis result. It is understood that the medical questionnaire to be annotated may include the patient's chief complaints, the patient's dialogue with the doctor, the doctor's diagnosis results, and the like. Further, the medical questionnaire to be annotated may also include personal information of the patient, personal information of the doctor, and the like.

Specifically, the medical questionnaire to be annotated may be an initial medical questionnaire acquired by the server on line, or may be obtained by the server by screening the initial medical questionnaire acquired on line. And if the medical questionnaire to be annotated is the screened medical questionnaire, the precision of the medical questionnaire to be annotated is higher than that of the initial medical questionnaire, for example, the medical questionnaire to be annotated may be the questionnaire obtained by removing the initial medical questionnaire with wrong diagnosis result or less information in the server.

In one embodiment, the step of obtaining the medical questionnaire to be annotated further comprises: and acquiring an initial medical inquiry sheet corresponding to the disease type, and screening the initial medical inquiry sheet by utilizing at least one preset rule to obtain the medical inquiry sheet to be labeled.

Wherein, the initial medical questionnaire is the most original questionnaire acquired by the server from the online. And an initial medical questionnaire comprising a plurality of disease types online. In one embodiment, the server may retrieve initial medical questionnaires corresponding to different online disease types from a database. In another embodiment, the server may further obtain initial medical inquiry data of departments corresponding to different departments on the line from the database, and further, may further extract an initial medical inquiry sheet corresponding to each type of disease from the initial medical inquiry data of departments according to different disease types in the departments. For example, the server may count the number of initial medical questionnaires corresponding to each disease type in the acquired department initial medical questionnaire, determine the priority of different disease types according to the counted number, and extract the initial medical questionnaire corresponding to the disease type from the department initial medical questionnaire according to the priority order.

Specifically, a large number of initial medical questionnaires acquired on line are screened in the server, the initial medical questionnaires which do not meet the standard are deleted, and the medical questionnaires to be annotated are determined according to the initial medical questionnaires which meet the standard. Specifically, the screening process for the initial medical questionnaire includes, but is not limited to, at least one of the following steps:

step one, whether the patient points out the disease name in the initial medical questionnaire is detected in the server, if the patient is judged to point out the disease name, the initial medical questionnaire with the disease name pointed out by the patient is removed, and the rest of the initial medical questionnaire is used as the medical questionnaire to be annotated. It is understood that the server may detect from the patient's complaint whether the patient has actively referred to the disease name, or the server may detect from the patient's dialog with the doctor whether the patient has actively referred to the disease name. The detection method may be a keyword matching algorithm, and the like, which is not limited herein.

And step two, the server detects whether the number of conversation rounds of doctors and patients in the initial medical questionnaire is larger than a preset number of rounds. And if the number of the dialogue rounds of the initial medical questionnaire is smaller than the preset number of rounds, removing the initial medical questionnaire, and taking the rest initial medical questionnaire as the questionnaire to be annotated. Each initial medical questionnaire corresponds to one stored data in the database, the number of conversation rounds is stored in the stored data, the conversation discussion refers to that the user asks for the information once, the doctor assistant answers the information once, and the number of conversation rounds stored in advance can be directly read from the background because the number of conversation rounds is corresponding to the count +1 in the background.

And step three, detecting whether the initial medical questionnaire comprises preset keywords in the server according to a predefined test script and a dictionary model, wherein the preset keywords can be keywords stored in DKD dictionary data. And if the server judges that the initial medical questionnaire does not comprise the DKD dictionary data, removing the initial medical questionnaire which does not comprise the DKD dictionary data, and taking the rest initial medical questionnaire as the questionnaire to be annotated. Wherein, the DKD dictionary comprises keywords corresponding to disease types, such as 40 types of diseases, symptoms, body parts, body systems, colors, disease course correlation, etiology correlation, organs, examination items, nerves, operations, physical signs and the like.

It should be noted that, the above three methods for screening the initial medical questionnaire in the server may be used to screen the initial medical questionnaire according to any one of the above three methods. The three screening methods can be combined according to a certain screening priority to generate a screening order, and the initial medical questionnaire is screened in the server according to the certain screening order to obtain the medical questionnaire to be annotated. For example, the server may first delete the data indicating the name of the disease actively by the patient, then the server may remove the questionnaire with the number of rounds of dialogue less than the preset number of rounds from the remaining initial medical questionnaire, and then the server may delete the questionnaire without the DKD dictionary data from the questionnaire with the number of rounds of dialogue less than the preset number of rounds of dialogue, so as to obtain the medical questionnaire to be annotated according to at least three screening steps. That is, the order of screening medical questionnaires using different screening conditions is not limited in the present application, and specifically, screening may be performed according to individual screening conditions, or screening may be performed by combining different screening conditions, and in the case of combined screening, the order of combined screening is not limited.

In a specific embodiment, it is considered that the diagnosis result in the real medical inquiry data acquired by the server from the online may have inaccurate diagnosis, that is, the real medical inquiry data acquired by the server from the online is not 100% reliable and accurate, such as the real inquiry list on the online may have wrong diagnosis. Therefore, the real medical data acquired on the line needs to be screened to obtain medical data with higher accuracy, and then the medical data with higher accuracy is used to check the accuracy of the medical diagnosis algorithm. The method for screening the acquired on-line real medical inquiry data in the application comprises the following steps:

first, the server obtains a large amount of on-line real medical inquiry data. Specifically, the real medical inquiry data acquired by the server on line is acquired by taking the department as latitude, and deeper, the data can be acquired by taking the disease in the department as latitude. Specifically, the server screens the on-line inquiry sheets (with the order of tens of millions) in the last month to obtain initial medical inquiry sheets with diagnosis results of corresponding diseases, and then screens the dialogue contents in the initial medical inquiry sheets to screen the initial medical inquiry sheets with the number of inquiry rounds being more than 5. And the initial medical questionnaire with less than 5 sessions was removed.

Further, a sub-standard medical questionnaire is screened in the server. Specifically, after the initial medical questionnaire is selected, the selected patient's chief complaint content and the doctor-patient dialogue content in the initial medical questionnaire are further screened, the initial medical questionnaire in which the patient's chief complaint directly says the disease name is filtered, the initial medical questionnaire in which the patient mentions the disease name in the dialogue process is filtered, and the secondary standard medical questionnaire is obtained through the two-layer screening. It should be noted that, if the precision of the medical diagnosis algorithm is tested by using the medical questionnaire in which the patient refers to the disease name during the inquiry process, the diagnostic precision of the algorithm is interfered, so that the algorithm is over-fitted, and the diagnostic accuracy of the algorithm is reduced. Thus, it is preferable that the patient be automatically informed of the initial questionnaire removal of the disease during the course of the questionnaire.

Then, in the server, algorithmic filtering is performed again on the sub-standard questionnaire using the test script and the dictionary model. Specifically, an algorithm model such as a keyword matching algorithm is used for extracting information of IE (information extraction) of main complaint content and conversation content in the secondary standard questionnaire, IE extraction corresponds to a DKD dictionary, and the secondary standard medical questionnaire without the DKD dictionary is filtered in the server. The DKD dictionary is a self-defined dictionary, keyword data in the DKD dictionary can be added or deleted or modified, the DKD dictionary comprises various symptom information corresponding to each disease, the DKD dictionary also comprises information such as synonyms of each keyword, the information is used for different expression conditions of the same semantic, and for different semantic expression conditions, the output is standardized by taking one keyword as a standard.

As table one, table one provides a DKD dictionary. Specifically, the DKD dictionary includes at least four parts of DKD names, and a system ID, a DKD code, and a DKD _ Tag (DKD Tag) corresponding to each DKD name in the system. Further, the DKD name can be 40 types such as disease name, disease symptom, body part, body system, color, disease course related, etiology related, organ, examination item, nerve, operation, sign, etc., which are only partially illustrated in table one and not limited herein.

table-DKD dictionary

System ID	DKD coding	DKD name	DKD_TAG
				568	1	Hand part	BodyPart
840	2	Foot part	BodyPart
				620	4	Buttocks part	BodyPart
20417	8	Worry about more suspicion	Symptom
				20465	9	One case has more than 2 convulsions	Course
322	10	Convulsion	Symptom
				19191	15	Prostatitis	Symptom
442	17	Routine of urine	LabTest
				993	18	Epigastric pain	Symptom
550	37	Physical examination of nervous system	Physical

In the above embodiment, the most complex four steps in the standard medical questionnaire screening steps, such as algorithm filtering, manual labeling verification, doctor verification and manual countermarking, are systematically integrated, so that the standard questionnaire screening is simpler and more orderly.

And step 204, marking the main complaint content and the dialogue content in the medical inquiry list to be marked respectively to obtain a main complaint marking result and a dialogue marking result.

Specifically, the server marks the main complaint content of the patient in the medical questionnaire to be marked to obtain a main complaint marking result. The complaint is the patient's own symptoms and/or signs, nature, and duration. Specifically, the server extracts the main complaint keywords according to the main complaint content in the questionnaire to be labeled, and generates a main complaint labeling result according to the main complaint keywords. And the server marks the conversation contents of the doctors and the patients in the medical questionnaire to be marked to obtain a conversation marking result. Specifically, the server extracts the conversation keywords according to the conversation content in the questionnaire to be labeled, and generates a conversation labeling result according to the conversation keywords.

As shown in fig. 3, fig. 3 is an interface diagram for labeling an inquiry sheet to be labeled according to an embodiment. In fig. 3, it can be seen that the case details include diagnosis conditions, chief complaint contents, basic information of inquiry, template summary, human conversation, and the like. The diagnosis condition is a condition formula consisting of a plurality of information according to a certain rule, and the basic information of the inquiry comprises inquiry ID, inquiry time, inquiry department and other information. The annotation work can be performed in the annotation details according to the content in the case details. And the type, Code, slot position, qualitative, remark and deletion operation of the labeling keyword can be selected from the labeling details. And further, actions such as adding operation, submitting slot position, main complaint reference and template reference can be executed.

Step 206, determining the diagnosis condition according to the diagnosis result in the medical inquiry list to be annotated.

The diagnosis condition is a condition statement which is obtained by combining more than one keyword according to a certain rule. It can be understood that a plurality of keywords can be combined according to a preset rule by using boolean operations to obtain a conditional statement. For example, the diagnosis condition corresponding to the menostaxis disease is "[ 8y-65y ] and female menstrual period >7 days and (hypogastric pain or anemia or lumbosacral pain or dizziness or pregnancy) and! Menstrual cycle disorders. And then, the dialogue labeling result is verified in the server (or manually) by using the diagnosis condition so as to check whether the dialogue labeling result meets the corresponding diagnosis condition.

And step 208, screening the chief complaint labeling results and the dialogue labeling results according to the diagnosis conditions to obtain a standard medical questionnaire.

Specifically, the server verifies the chief complaint labeling result and the dialogue labeling result according to the diagnosis conditions, and extracts the medical questionnaire to be labeled passing the verification as a standard medical questionnaire. Specifically, if the server judges that the section main complaint labeling result and the dialogue labeling result meet the diagnosis condition, the corresponding medical questionnaire to be labeled is judged to be a qualified questionnaire, and if not, the corresponding medical questionnaire to be labeled is judged to be an unqualified questionnaire. In other words, when the medical questionnaire to be annotated is determined to be an unqualified questionnaire, it is described that the diagnosis result obtained according to the dialogue content in the medical questionnaire to be annotated has an error, and thus the medical questionnaire to be annotated is an unqualified questionnaire. It should be noted that, in this embodiment, the medical questionnaire to be annotated may be verified manually, and at this time, in order to accelerate the annotation efficiency, an automatic association function is implemented in the DKD dictionary slot annotation, so that the annotation is simpler.

In the above embodiment, a large amount of real medical data is acquired in the server, and then the medical questionnaire acquired online is screened by using at least one screening algorithm to obtain the standard medical questionnaire, so that the acquisition efficiency of the standard medical questionnaire is improved. Specifically, the dialogue labeling result and the chief complaint labeling result are extracted from the medical questionnaire to be labeled, and the medical questionnaire to be labeled is verified according to the diagnosis conditions, the dialogue labeling result and the chief complaint labeling result, so that the automatic verification of the diagnosis result in the medical questionnaire to be labeled is realized, and the screening precision of the questionnaire is improved.

In one embodiment, the annotation mode of the dialog annotation result comprises the following steps: matching the conversation content in the medical questionnaire to be annotated with preset keywords in a preset dictionary library, and extracting the successfully matched preset keywords as conversation keywords; storing preset keywords in a preset dictionary library in a block chain; extracting positive and negative directions of the conversation keywords from the conversation content, and configuring conversation labeling directions for the conversation keywords according to the positive and negative directions; and obtaining a conversation labeling result according to the conversation keywords and the conversation labeling direction.

The dialog annotation direction is used to indicate whether the dialog keyword exists, for example, when the dialog annotation direction is positive (+) the patient has symptoms corresponding to the dialog keyword, and when the dialog annotation direction is negative (-), the patient does not have symptoms corresponding to the dialog keyword. Furthermore, the method also comprises the steps of acquiring the label (such as ID or code) of each conversation keyword in the DKD dictionary, and then obtaining the conversation labeling result according to the conversation keyword, the conversation labeling direction corresponding to the conversation keyword and the label of the conversation keyword in the DKD dictionary.

It should be emphasized that, in order to further ensure the privacy and security of the preset keywords in the preset dictionary repository, the preset keywords in the preset dictionary repository may also be stored in a node of a block chain.

In the above embodiment, the technical effect of extracting effective content from the medical questionnaire to be annotated is achieved by extracting the dialogue keywords from the questionnaire to be annotated and the dialogue annotation direction of each dialogue keyword, and technical support is provided for subsequently checking the dialogue content. And the screening efficiency of the subsequent medical questionnaire to be annotated is improved.

In one embodiment, the method for labeling the main complaint content in the medical questionnaire to be labeled to obtain a main complaint labeling result includes: matching the main complaint content of the patient in the medical questionnaire to be annotated with preset keywords in a preset dictionary library, and extracting the successfully matched preset keywords as the main complaint keywords; extracting positive and negative directions of the main complaint keywords from the main complaint content, and configuring main complaint labeling directions for the main complaint keywords according to the positive and negative directions; and obtaining a main complaint labeling result according to the main complaint labeling keyword and the main complaint labeling direction.

The direction of the chief complaint label is used to indicate whether the chief complaint keyword exists, and if the direction of the chief complaint label is positive (+) the patient has the symptom corresponding to the chief complaint keyword, and if the direction of the chief complaint label is negative (-) the patient does not have the symptom corresponding to the chief complaint keyword. Furthermore, the method also comprises the steps of obtaining the label of each main complaint keyword in the DKD dictionary, and then obtaining a main complaint labeling result according to the main complaint keyword, the main complaint labeling direction corresponding to the main complaint keyword and the label of the main complaint keyword in the DKD dictionary.

In one embodiment, the preset dictionary library comprises a plurality of preset keywords, and each preset keyword comprises a preset main keyword and a preset sub keyword; matching the conversation content in the medical questionnaire to be labeled with preset keywords in a preset dictionary library, and extracting the successfully matched preset keywords as conversation keywords, wherein the method comprises the following steps: matching the conversation content in the medical questionnaire to be annotated with preset keywords in a preset dictionary library; and when the conversation content is successfully matched with the preset sub-keywords in the preset keywords, acquiring preset main keywords corresponding to the successfully matched preset sub-keywords, and extracting the preset main keywords as the conversation keywords.

Wherein the predetermined dictionary database is a pre-established dictionary database, such as a DKD dictionary database. The DKD dictionary database comprises keywords corresponding to disease types and similar words of each keyword, and the keywords are used as main keywords and the similar words are used as related keywords; and if the server detects that the medical questionnaire to be annotated comprises the main keywords or the related keywords in the DKD dictionary database, judging that the questionnaire to be annotated comprises the DKD data, and generating the conversation keywords according to the matched main keywords.

The similar words may be in different expression forms of the same keyword, such as different chinese expression forms or different language type expression forms (chinese, english, etc.) of the same keyword, which is not limited herein.

In the above embodiment, the main keywords and the related keywords are pre-stored in the DKD database, so that the method of the present disclosure can adapt to medical scenes of more users without considering medical expressions of the users, and the extraction process of the keywords is not affected by different medical expressions of the users. In addition, in the embodiment, the main keywords are used as the dialogue keywords, so that the standardized processing of the dialogue keywords is realized, and technical conditions are provided for the subsequent screening of medical diagnosis lists.

In one embodiment, matching the dialogue content in the medical questionnaire to be annotated with preset keywords in a preset dictionary library, and extracting the successfully matched preset keywords as the dialogue keywords comprises: natural language processing is carried out on the conversation content to obtain keywords to be matched; matching the keywords to be matched with preset keywords in a preset dictionary library; and extracting the successfully matched preset keywords as the conversation keywords.

Specifically, natural language processing is carried out on conversation content between a patient and a doctor in the server so as to realize semantic analysis, text classification, emotion analysis, intention identification and the like on the conversation content, and then the undetermined keywords are obtained according to the result of the natural language processing. And matching the pending keywords with the content in the DKD dictionary to extract the successfully matched keywords as the conversation keywords.

Because the conversation contents of the patient and the doctor are flexible, if the information in the conversation contents is extracted by simply using a keyword matching technology without considering the relevance between the contexts in the real conversation contents, the problem of low keyword extraction accuracy is easily caused. Therefore, in the embodiment, the natural language algorithm is used for analyzing and processing the conversation content to obtain the keywords, so that the accuracy of analyzing the conversation content is improved, and the accuracy of the labeling result of the conversation content is improved. It is to be appreciated that the natural language algorithm can include, but is not limited to, an artificial intelligence algorithm, a deep learning algorithm, and the like.

In one embodiment, the screening of the chief complaint labeling result and the dialogue labeling result according to the diagnosis condition to obtain a standard medical questionnaire comprises: verifying the conversation keywords in the conversation labeling result, the conversation labeling directions corresponding to the conversation keywords, the chief complaint keywords in the chief complaint labeling result and the chief complaint labeling directions corresponding to the chief complaint keywords according to the diagnosis conditions; and extracting the medical inquiry sheet to be marked which passes the verification as a standard medical inquiry sheet.

Specifically, the server determines whether the dialog annotation result satisfies the diagnosis condition according to the dialog keyword in the dialog annotation result and the dialog annotation direction corresponding to the dialog keyword.

In the above embodiment, the server automatically extracts the key information in the medical questionnaire to be annotated, generates the chief complaint annotation result and the dialogue annotation result, and automatically verifies the chief complaint annotation result and the dialogue annotation result according to the diagnosis conditions, thereby greatly improving the screening efficiency of the medical questionnaire.

In one embodiment, after the standard medical questionnaire is obtained by screening the chief complaint annotation result and the dialogue annotation result according to the diagnosis condition, the method further comprises: in the interactive interface, manually checking the fields to be checked in each standard medical questionnaire; and deleting the standard questionnaire which fails in verification, extracting the field to be verified which fails in verification from the standard medical questionnaire which fails in verification, and updating the preset keywords in the preset dictionary base according to the field to be verified which fails in verification.

Specifically, the standard questionnaire is checked again in the server according to professional knowledge (or professionals), and whether the dialogue marking result in the standard questionnaire meets the corresponding diagnosis condition is checked. And if the standard medical questionnaire is met, the standard medical questionnaire is confirmed to be the final standard medical questionnaire, otherwise, the standard medical questionnaire is removed.

In one embodiment, referring to fig. 4, fig. 4 is an interface diagram of an interview order verification system provided in one embodiment. Inclusion of the chief complaints, diagnostic conditions, chief complaint labels, and dialogue labels can be seen in FIG. 4. And the case situation is also included in fig. 4, so that the server records the screening result data of the medical questionnaire to be annotated according to the chief complaint annotation and the dialogue annotation.

Specifically, a professional doctor team may be invited to perform the final check of the standard questionnaire, and in order to facilitate the checking efficiency of the professional doctors on the standard questionnaire, the present proposal further provides a questionnaire checking system, which is an interface diagram of the questionnaire checking system provided in an embodiment as shown in fig. 4. In the questionnaire examination system, the specialist doctor can check all the information of the standard questionnaire to be examined in the questionnaire examination system. Specific viewable information includes, but is not limited to, patient's chief complaints, doctor-patient's dialogue (human dialogue), doctor-patient's inquiry details, and annotated DKD dictionary summaries, diagnostic conditions for intelligent diagnosis, etc. And the professional doctor can check the standard questionnaire to be checked, and judge the medical record condition, quality inspection diagnosis, whether the standard medical record is available, whether the template is complete, whether the prescription is accurate and the like of the current standard questionnaire to be checked. In the embodiment, the systematic checking mode is executed in the inquiry list checking system, so that the checking efficiency and the checking accuracy of the professional doctors are greatly improved. And after the professional doctor checks, the professional doctor can supplement part of the standard questionnaire, points out the places with missing labeling and places with wrong labeling, performs labeling inspection on the standard questionnaire, and supplements and modifies the corresponding DKD dictionary, so that the standard medical questionnaire is screened based on the updated DKD dictionary subsequently, and the screening accuracy of the standard medical questionnaire is improved.

In this embodiment, the professional doctor or the professional knowledge data is used to check the selected labeled questionnaire again, and the standard questionnaire that passes the check is used as the test questionnaire, while the standard questionnaire that fails the check is removed, so as to further improve the accuracy of the test questionnaire. In addition, in this embodiment, in order to improve the efficiency of the professional doctor in checking the standard questionnaire, the standard questionnaire submitted to the professional doctor is already the questionnaire screened by the layer-by-layer standard, so that the number of the questionnaires sent to the professional doctor nodes is reduced, and the checking efficiency of the professional doctor is improved.

In the above embodiment, a large amount of real medical data is acquired in the server, and then the medical questionnaire acquired on-line is screened by using at least one screening algorithm to obtain a standard medical questionnaire, so that the usability of the questionnaire is improved. And the dialogue labeling result and the chief complaint labeling result are extracted from the inquiry list so as to verify the dialogue labeling result and the chief complaint labeling result according to the diagnosis conditions, so that the verification of the diagnosis result in the inquiry list is realized, and the screening precision of the inquiry list is improved. And when the annotation information is extracted from the inquiry list, the accuracy of extracting the annotation information can be further improved by combining a natural language algorithm. The method also comprises the step of screening the inquiry sheet through manual work (professional knowledge), and the part with the judgment error in the previous step can be reversely corrected, so that the subsequent marking accuracy is improved.

As shown in fig. 5, fig. 5 provides a flow chart of a method for testing the accuracy of a medical inquiry algorithm, and the method includes:

step 502, a medical algorithm to be tested is obtained.

The medical algorithm to be tested is an algorithm for automatic medical diagnosis, and an algorithm for further verifying the accuracy of diagnosis thereof is required.

Step 504, the standard medical questionnaire obtained by the screening of the artificial intelligence-based medical questionnaire screening method provided in any one of the above embodiments is processed to obtain a test diagnosis result.

In the above embodiment, the standard medical questionnaire obtained by screening with the medical questionnaire screening method based on artificial intelligence has higher accuracy, so that the standard medical questionnaire screened in this step is used as input data of the medical algorithm to be tested to check the testing accuracy of the medical algorithm to be tested. Specifically, a medical algorithm to be tested is used for carrying out data analysis on standard medical data to obtain a test diagnosis result.

Step 506, comparing the test diagnosis result with the real diagnosis result in the standard medical questionnaire.

Then, the test diagnosis result obtained by the medical algorithm to be tested is compared with the real diagnosis result in the standard medical questionnaire.

And step 508, judging that the precision of the medical inquiry algorithm to be tested meets the test requirement when the comparison is passed.

Since the standard medical questionnaire is the screened questionnaire with high precision, it can be said that the real diagnosis result in the standard medical questionnaire is the result verified by the professional doctor, so that the test diagnosis result obtained by the medical algorithm to be tested is compared with the real diagnosis result in the standard medical questionnaire, and the precision of the automatic inquiry of the medical algorithm to be tested can be evaluated. Only when the comparison is passed, the diagnosis precision of the medical algorithm to be tested is satisfied, that is, the medical algorithm to be tested is the algorithm with the precision satisfying the automatic inquiry requirement. Otherwise, the accuracy of the medical algorithm to be tested does not meet the requirement of automatic diagnosis, and the algorithm is an algorithm which needs to be optimized again.

With the rapid development of internet medical technology, the development of intelligent medical technology is accelerated. In this embodiment, before the medical diagnosis is automatically performed by using the medical inquiry algorithm, the diagnosis accuracy of the medical inquiry algorithm is verified. In addition, in order to improve the accuracy of the accuracy verification of the medical inquiry algorithm, a large amount of medical test data is firstly obtained on line, a standard medical inquiry list is obtained through the combination of multiple algorithms (a medical inquiry list screening method based on artificial intelligence) screening, and finally the standard medical inquiry list is used as test data for testing the accuracy of the medical inquiry algorithm, so that the accuracy of the accuracy verification of the medical inquiry algorithm is greatly improved.

It should be understood that although the steps in the flowcharts of fig. 2 and 5 are shown in order as indicated by the arrows, the steps are not necessarily performed in order as indicated by the arrows. The steps are not performed in the exact order shown and described, and may be performed in other orders, unless explicitly stated otherwise. Moreover, at least some of the steps in fig. 2 and 5 may include multiple sub-steps or multiple stages, which are not necessarily performed at the same time, but may be performed at different times, and the order of performing the sub-steps or stages is not necessarily sequential, but may be performed alternately or alternately with other steps or at least some of the sub-steps or stages of other steps.

In one embodiment, as shown in fig. 6, there is provided an artificial intelligence based medical questionnaire screening apparatus 600, comprising:

the obtaining module 601 is configured to obtain a medical questionnaire to be annotated.

And the labeling module 602 is configured to label the chief complaint content and the conversation content in the medical questionnaire to be labeled respectively, so as to obtain a chief complaint labeling result and a conversation labeling result.

The determining module 603 is configured to determine a diagnosis condition according to a diagnosis result in the medical questionnaire to be annotated.

And the screening module 604 is configured to screen the chief complaint labeling result and the dialogue labeling result according to the diagnosis condition to obtain a standard medical questionnaire.

In one embodiment, the medical questionnaire screening apparatus 600 further includes a dialogue labeling module 605, where the dialogue labeling module 605 is configured to match the dialogue content in the medical questionnaire to be labeled with the preset keywords in the preset dictionary library, and extract the successfully matched preset keywords as the dialogue keywords; storing preset keywords in a preset dictionary library in a block chain; extracting positive and negative directions of the conversation keywords from the conversation content, and configuring conversation labeling directions for the conversation keywords according to the positive and negative directions; and obtaining a conversation labeling result according to the conversation keywords and the conversation labeling direction.

In one embodiment, the dialog labeling module 605 is further configured to match the dialog contents in the medical questionnaire to be labeled with the preset keywords in the preset dictionary library; and when the conversation content is successfully matched with the preset sub-keywords in the preset keywords, acquiring preset main keywords corresponding to the successfully matched preset sub-keywords, and extracting the preset main keywords as the conversation keywords.

In one embodiment, the dialog annotation module 605 is further configured to perform natural language processing on the dialog content to obtain a keyword to be matched; matching the keywords to be matched with preset keywords in a preset dictionary library; and extracting the successfully matched preset keywords as the conversation keywords.

In one embodiment, the screening module 604 is further configured to verify the conversation keyword in the conversation annotation result, the conversation annotation direction corresponding to each conversation keyword, and the chief complaint keyword in the chief complaint annotation result, and the chief complaint annotation direction corresponding to each chief complaint keyword according to the diagnosis condition; and extracting the medical inquiry sheet to be marked which passes the verification as a standard medical inquiry sheet.

In one embodiment, the medical questionnaire screening apparatus 600 further includes a verification module 606, and the verification module 606 is configured to perform manual verification on the fields to be verified in each standard medical questionnaire in the interactive interface; and deleting the standard questionnaire which fails in verification, extracting the field to be verified which fails in verification from the standard medical questionnaire which fails in verification, and updating the preset keywords in the preset dictionary base according to the field to be verified which fails in verification.

In one embodiment, as shown in fig. 7, an apparatus 700 for testing the accuracy of an artificial intelligence-based medical interrogation algorithm includes:

an algorithm obtaining module 701, configured to obtain a medical inquiry algorithm to be tested.

The test result obtaining module 702 is configured to process the standard medical questionnaire obtained according to the artificial intelligence-based medical questionnaire screening method by using the to-be-tested medical questionnaire algorithm, so as to obtain a test diagnosis result.

A comparison module 703, configured to compare the test diagnosis result with the real diagnosis result in the standard medical questionnaire.

And the judging module 704 is used for judging that the precision of the medical inquiry algorithm to be tested meets the testing requirement when the comparison is passed.

For the specific limitations of the above medical inquiry sheet screening device based on artificial intelligence and the precision testing device based on the medical inquiry algorithm based on artificial intelligence, reference may be made to the limitations of the above medical inquiry sheet screening method based on artificial intelligence and the precision testing method based on the medical inquiry algorithm based on artificial intelligence, which are not described herein again. All modules in the medical inquiry list screening device based on artificial intelligence and the precision testing device based on the medical inquiry algorithm based on artificial intelligence can be completely or partially realized by software, hardware and a combination thereof. The modules can be embedded in a hardware form or independent from a processor in the computer device, and can also be stored in a memory in the computer device in a software form, so that the processor can call and execute operations corresponding to the modules.

In one embodiment, a computer device is provided, which may be a server, and its internal structure diagram may be as shown in fig. 8. The computer device includes a processor, a memory, a network interface, and a database connected by a system bus. Wherein the processor of the computer device is configured to provide computing and control capabilities. The memory of the computer device comprises a nonvolatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program, and a database. The internal memory provides an environment for the operation of an operating system and computer programs in the non-volatile storage medium. The database of the computer device is used for storing data related to medical interrogation. The network interface of the computer device is used for communicating with an external terminal through a network connection. The computer program is executed by a processor to implement an artificial intelligence-based medical questionnaire screening method capable of improving the efficiency of medical questionnaire screening.

Those skilled in the art will appreciate that the architecture shown in fig. 8 is merely a block diagram of some of the structures associated with the disclosed aspects and is not intended to limit the computing devices to which the disclosed aspects apply, as particular computing devices may include more or less components than those shown, or may combine certain components, or have a different arrangement of components.

In one embodiment, a computer device is provided, comprising a memory storing a computer program and a processor implementing the following steps when the processor executes the computer program: acquiring a medical inquiry sheet to be marked; marking the main complaint content and the dialogue content in the medical inquiry list to be marked respectively to obtain a main complaint marking result and a dialogue marking result; determining a diagnosis condition according to a diagnosis result in the medical inquiry list to be annotated; and screening the chief complaint labeling results and the dialogue labeling results according to the diagnosis conditions to obtain a standard medical questionnaire.

In one embodiment, the computer program when executed by the processor is further operable to: matching the conversation content in the medical questionnaire to be annotated with preset keywords in a preset dictionary library, and extracting the successfully matched preset keywords as conversation keywords; storing preset keywords in a preset dictionary library in a block chain; extracting positive and negative directions of the conversation keywords from the conversation content, and configuring conversation labeling directions for the conversation keywords according to the positive and negative directions; and obtaining a conversation labeling result according to the conversation keywords and the conversation labeling direction.

In one embodiment, the processor, when executing the computer program, performs the steps of matching the dialogue content in the medical questionnaire to be annotated with the preset keywords in the preset dictionary library, and extracting the successfully matched preset keywords as the dialogue keywords, further: matching the conversation content in the medical questionnaire to be annotated with preset keywords in a preset dictionary library; and when the conversation content is successfully matched with the preset sub-keywords in the preset keywords, acquiring preset main keywords corresponding to the successfully matched preset sub-keywords, and extracting the preset main keywords as the conversation keywords.

In one embodiment, the processor, when executing the computer program, performs the steps of matching the dialogue content in the medical questionnaire to be annotated with the preset keywords in the preset dictionary library, and extracting the successfully matched preset keywords as the dialogue keywords, further: natural language processing is carried out on the conversation content to obtain keywords to be matched; matching the keywords to be matched with preset keywords in a preset dictionary library; and extracting the successfully matched preset keywords as the conversation keywords.

In one embodiment, the processor, when executing the computer program, further performs the step of screening the chief complaint labeling result and the dialogue labeling result according to the diagnosis condition to obtain a standard medical questionnaire: verifying the conversation keywords in the conversation labeling result, the conversation labeling directions corresponding to the conversation keywords, the chief complaint keywords in the chief complaint labeling result and the chief complaint labeling directions corresponding to the chief complaint keywords according to the diagnosis conditions; and extracting the medical inquiry sheet to be marked which passes the verification as a standard medical inquiry sheet.

In one embodiment, the computer program when executed by the processor is further operable to: in the interactive interface, manually checking the fields to be checked in each standard medical questionnaire; and deleting the standard questionnaire which fails in verification, extracting the field to be verified which fails in verification from the standard medical questionnaire which fails in verification, and updating the preset keywords in the preset dictionary base according to the field to be verified which fails in verification.

In one embodiment, a computer device is provided, comprising a memory storing a computer program and a processor implementing the following steps when the processor executes the computer program: acquiring a medical inquiry algorithm to be tested; processing a standard medical inquiry sheet obtained according to an artificial intelligence-based medical inquiry sheet screening method by using a to-be-tested medical inquiry algorithm to obtain a test diagnosis result; comparing the test diagnosis result with the real diagnosis result in the standard medical questionnaire; and when the comparison is passed, judging that the precision of the medical inquiry algorithm to be tested meets the test requirement.

In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored, which when executed by a processor performs the steps of: acquiring a medical inquiry sheet to be marked; marking the main complaint content and the dialogue content in the medical inquiry list to be marked respectively to obtain a main complaint marking result and a dialogue marking result; determining a diagnosis condition according to a diagnosis result in the medical inquiry list to be annotated; and screening the chief complaint labeling results and the dialogue labeling results according to the diagnosis conditions to obtain a standard medical questionnaire.

In one embodiment, the computer program when executed by the processor performs the steps of matching the dialogue content in the medical questionnaire to be annotated with preset keywords in a preset dictionary library, and extracting the successfully matched preset keywords as the dialogue keywords, further: matching the conversation content in the medical questionnaire to be annotated with preset keywords in a preset dictionary library; and when the conversation content is successfully matched with the preset sub-keywords in the preset keywords, acquiring preset main keywords corresponding to the successfully matched preset sub-keywords, and extracting the preset main keywords as the conversation keywords.

In one embodiment, the computer program when executed by the processor performs the steps of matching the dialogue content in the medical questionnaire to be annotated with preset keywords in a preset dictionary library, and extracting the successfully matched preset keywords as the dialogue keywords, further: natural language processing is carried out on the conversation content to obtain keywords to be matched; matching the keywords to be matched with preset keywords in a preset dictionary library; and extracting the successfully matched preset keywords as the conversation keywords.

In one embodiment, the computer program when executed by the processor performs the step of screening the chief complaint annotation result and the dialogue annotation result to obtain a standard medical questionnaire according to the diagnosis condition is further configured to: verifying the conversation keywords in the conversation labeling result, the conversation labeling directions corresponding to the conversation keywords, the chief complaint keywords in the chief complaint labeling result and the chief complaint labeling directions corresponding to the chief complaint keywords according to the diagnosis conditions; and extracting the medical inquiry sheet to be marked which passes the verification as a standard medical inquiry sheet.

In one embodiment, a computer device is provided, comprising a memory and a processor, the memory storing a computer program that when executed by the processor performs the steps of: acquiring a medical inquiry algorithm to be tested; processing a standard medical inquiry sheet obtained according to an artificial intelligence-based medical inquiry sheet screening method by using a to-be-tested medical inquiry algorithm to obtain a test diagnosis result; comparing the test diagnosis result with the real diagnosis result in the standard medical questionnaire; and when the comparison is passed, judging that the precision of the medical inquiry algorithm to be tested meets the test requirement.

The block chain is a novel application mode of computer technologies such as distributed data storage, point-to-point transmission, a consensus mechanism, an encryption algorithm and the like. A block chain (Blockchain), which is essentially a decentralized database, is a series of data blocks associated by using a cryptographic method, and each data block contains information of a batch of network transactions, so as to verify the validity (anti-counterfeiting) of the information and generate a next block. The blockchain may include a blockchain underlying platform, a platform product service layer, an application service layer, and the like.

It will be understood by those skilled in the art that all or part of the processes of the methods of the embodiments described above can be implemented by hardware related to instructions of a computer program, which can be stored in a non-volatile computer-readable storage medium, and when executed, can include the processes of the embodiments of the methods described above. Any reference to memory, storage, database, or other medium used in the embodiments provided herein may include non-volatile and/or volatile memory, among others. Non-volatile memory can include read-only memory (ROM), Programmable ROM (PROM), Electrically Programmable ROM (EPROM), Electrically Erasable Programmable ROM (EEPROM), or flash memory. Volatile memory can include Random Access Memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in a variety of forms such as Static RAM (SRAM), Dynamic RAM (DRAM), Synchronous DRAM (SDRAM), Double Data Rate SDRAM (DDRSDRAM), Enhanced SDRAM (ESDRAM), Synchronous Link DRAM (SLDRAM), Rambus Direct RAM (RDRAM), direct bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).

The technical features of the above embodiments can be arbitrarily combined, and for the sake of brevity, all possible combinations of the technical features in the above embodiments are not described, but should be considered as the scope of the present specification as long as there is no contradiction between the combinations of the technical features.

The above examples only express several embodiments of the present application, and the description thereof is more specific and detailed, but not construed as limiting the scope of the invention. It should be noted that, for a person skilled in the art, several variations and modifications can be made without departing from the concept of the present application, which falls within the scope of protection of the present application. Therefore, the protection scope of the present patent shall be subject to the appended claims.

Claims

1. A medical questionnaire screening method based on artificial intelligence is characterized by comprising the following steps:

acquiring a medical inquiry sheet to be marked;

2. The method of claim 1, wherein the manner of labeling the dialog labeling result comprises:

matching the conversation content in the medical questionnaire to be annotated with preset keywords in a preset dictionary library, and extracting the successfully matched preset keywords as conversation keywords; storing preset keywords in the preset dictionary database in a block chain;

extracting the positive and negative directions of the conversation keywords from the conversation content, and configuring conversation labeling directions for the conversation keywords according to the positive and negative directions;

and obtaining a conversation labeling result according to the conversation keyword and the conversation labeling direction.

3. The method according to claim 2, wherein the predetermined dictionary base comprises a plurality of predetermined keywords, and each of the predetermined keywords comprises a predetermined main keyword and a predetermined sub keyword; the step of matching the conversation content in the medical questionnaire to be annotated with the preset keywords in the preset dictionary library and extracting the successfully matched preset keywords as the conversation keywords comprises the following steps:

and when the conversation content is successfully matched with preset sub-keywords in the preset keywords, acquiring preset main keywords corresponding to the successfully matched preset sub-keywords, and extracting the preset main keywords as the conversation keywords.

4. The method according to claim 2, wherein the matching of the dialogue content in the medical questionnaire to be annotated with the preset keywords in the preset dictionary library and the extracting of the successfully matched preset keywords as the dialogue keywords comprise:

carrying out natural language processing on the conversation content to obtain keywords to be matched;

5. The method of claim 1, wherein the screening the chief complaint annotation result and the dialogue annotation result according to the diagnosis condition to obtain a standard medical questionnaire comprises:

6. The method according to any one of claims 1 to 5, wherein after the screening of the chief complaint labeling result and the dialogue labeling result according to the diagnosis condition to obtain a standard medical questionnaire, the method further comprises:

in an interactive interface, manually checking the fields to be checked in each standard medical questionnaire;

deleting the standard questionnaire which is not verified, extracting the to-be-verified field which is failed to be verified from the standard medical questionnaire which is not verified, and updating the preset keywords in the preset dictionary base according to the to-be-verified field which is failed to be verified.

7. A precision testing method of a medical inquiry algorithm based on artificial intelligence is characterized by comprising the following steps:

acquiring a medical inquiry algorithm to be tested;

processing the standard medical questionnaire obtained by the artificial intelligence-based medical questionnaire screening method according to any one of claims 1 to 6 by using a to-be-tested medical questionnaire algorithm to obtain a test diagnosis result;

comparing the test diagnosis result with a real diagnosis result in the standard medical questionnaire;

8. An artificial intelligence based medical questionnaire screening apparatus, the apparatus comprising:

the marking module is used for marking the main complaint content and the dialogue content in the medical questionnaire to be marked respectively to obtain a main complaint marking result and a dialogue marking result;

the determining module is used for determining diagnosis conditions according to the diagnosis result in the medical questionnaire to be annotated;

and the screening module is used for screening the chief complaint labeling result and the dialogue labeling result according to the diagnosis condition to obtain a standard medical questionnaire.

9. A computer device comprising a memory and a processor, the memory storing a computer program, characterized in that the processor, when executing the computer program, implements the steps of the method as claimed in any one of claims 1 to 5 or 6 or 7.

10. A computer-readable storage medium, on which a computer program is stored, which, when being executed by a processor, carries out the steps of the method as claimed in any one of claims 1 to 5 or 6 or 7.