Interactive triage terminal based on AI analysis

Abstract

The invention relates to the field of data processing, in particular to an interactive triage terminal based on AI analysis, and aims to obtain symptom entities, sign entities and identity information by performing target entity extraction on acquired symptom description information. And then confidence evaluation is carried out on the target entity based on a pre-constructed triage database, so that whether symptoms and sign descriptions which are not consistent with the identity information exist in description information input by a user or not, or mismatched or low-probability symptoms and sign combinations exist in the description information input by the user is determined. And if the condition exists, the confidence coefficient is relatively low. When the confidence coefficient is low, the guide text is constructed based on the target entity and the triage database, so that the user is guided to perform correct expression. And when a correct expression with high confidence is obtained, inputting the correct expression into the triage model to execute AI analysis, so that a triage result can be obtained. According to the method, the condition of triage errors caused by the fact that a user cannot correctly express can be effectively avoided.

Description

Technical Field

[0001] The present invention relates to the field of data processing, and in particular to an interactive triage terminal based on AI analysis. Background Art

[0002] Triage is a method used to assess and prioritize patients in healthcare settings, primarily in emergency departments, disaster response scenarios, and in resource-limited settings. Its primary purpose is to ensure that patients who need the most urgent care receive timely medical care while allocating limited medical resources.

[0003] In the existing technology, triage mainly relies on personnel or triage systems. The existing triage system relies on the user's description of symptom information, but in actual interaction, many users are actually unable to accurately describe their own symptom information. By extracting and identifying entities from the language organized by the user, no one can correct the symptom description like humans, which often leads to misclassification. Summary of the invention

[0004] In view of this, an object of the present invention is to provide an interactive triage terminal based on AI analysis to solve the problems in the background technology.

[0005] In order to achieve the above object, the present invention adopts the following technical solutions: An interactive triage terminal based on AI analysis of the present invention comprises: The acquisition module is used to obtain the user's current symptom description information; An entity extraction module, used to extract multiple target entities from the current symptom description information, wherein the multiple target entities include symptom entities, sign entities and identity information; An analysis module, configured to analyze the multiple target entities based on a pre-built triage database to obtain confidences of the multiple target entities, wherein the triage database includes entity combinations corresponding to multiple triage results; A result judgment module is used to input the multiple target entities into a pre-built triage model to obtain triage results when the confidence of the multiple target entities is greater than or equal to a preset confidence threshold; and to construct a guidance text based on the triage database and the multiple target entities when the confidence of the multiple target entities is less than the preset confidence threshold, and send the guidance text to the user; A loop module is used to, when receiving the user's symptom description information, use the symptom description information as the current symptom description information and return to the entity extraction module until there are multiple target entities whose confidence is greater than or equal to a preset confidence threshold, and then input the multiple target entities into a pre-built triage model to obtain a triage result, wherein the symptom description information is generated based on the guide text.

[0006] In one embodiment of the present application, the method for constructing the triage database includes: Obtain historical triage records; Extract the symptom entities, sign entities, identity information and triage results recorded during multiple triages from the historical triage records , and standardize the symptom entity and the sign entity to obtain a standard symptom entity and standard signs entities , wherein the standard symptom entity and the standard sign entity both include body parts and feature descriptions; The symptom entities, sign entities and triage results recorded during multiple triages are divided based on identity information to obtain data units; and the symptom entities and sign entities in each data unit are divided based on the triage results to obtain symptom data subunits and sign data subunits; and each data unit is directly divided based on the triage results to obtain symptom-sign data subunits; Calculate each symptom entity in the symptom data subunit The probability of occurrence , each physical sign entity in the physical sign data subunit The probability of occurrence ,in, is the entity serial number; Extract the symptom-sign entity combination of the same time from the symptom-sign data subunit , and calculate for each symptom-sign entity combination The probability of occurrence ; Symptom entities with a probability less than a preset probability threshold , physical entity and symptom-sign entity combinations Remove from the data unit and update the reference symptom entity in the symptom data subunit The probability of occurrence , the reference physical sign entity in the physical sign data subunit The probability of occurrence , and updating the reference symptom-sign entity combination in the symptom-sign data subunit The probability of occurrence , and obtain the triage database.

[0007] In one embodiment of the present application, the multiple target entities are analyzed based on a pre-built triage database to obtain the confidence levels of the multiple target entities, including: Determining a target data unit in the triage database based on the identity information of the user; Comparing the symptom entity in the current description information with the reference symptom entities in multiple symptom data subunits, comparing the sign entity in the current description information with the reference sign entities in multiple sign data subunits, and comparing the symptom-sign entity combination in the current description information with the reference symptom-sign entity combination in the symptom-sign data subunits; When there is no state entity consistent with the reference symptom entity in the current description information, or when there is no sign entity consistent with the reference sign entity, the confidences of the multiple target entities are set to zero, and it is determined that the current description information does not match the identity information of the user; There is a reference symptom entity in the current description information A state entity that is consistent with the reference sign entity When the physical sign entity is consistent, the probability of occurrence corresponding to the same triage result , and Perform weighted summation to obtain the confidence of multiple target entities .

[0008] In one embodiment of the present application, constructing a guidance text based on the triage database and the multiple target entities includes: When the confidence of the multiple target entities is zero, constructing a guide text for indicating that the symptom or sign is inconsistent with the identity information; When the confidence of the multiple target entities is not zero, a confirmation guidance text is constructed based on the symptom entity and the sign entity in the target entity; and a confirmation guidance text is constructed based on the symptom entity or the sign entity for the reference symptom entity in the target data unit. Or reference sign entity Make a rating , and based on the top N reference symptom entities with the highest scores Or reference sign entity Construct reference guide text, where scoring The mathematical expression is:

[0009] In the formula, is the first weight, is the second weight, is the cosine similarity between the symptom entity and the reference symptom entity, is the cosine similarity between the physical sign entity and the reference physical sign entity.

[0010] In one embodiment of the present application, after sending the guide text to the user, the method further includes: When receiving the confirmation description from the user, the confidences of the multiple target entities are set to target values ​​that are greater than or equal to a preset confidence threshold.

[0011] In one embodiment of the present application, the reference symptom entity The probability of occurrence The mathematical expression of the reference sign entity The probability of occurrence , and the reference symptom-sign entity combination The probability of occurrence They are:

[0012]

[0013]

[0014] In the formula, is the total number of reference symptom entities in the symptom data subunit, is the total number of reference sign entities in the sign data subunit, is the total number of reference symptom-sign entity combinations in the symptom-sign data subunit.

[0015] In one embodiment of the present application, the method for constructing the triage model includes: Obtain triage records; Extracting a plurality of triage sample data from the triage record, wherein the triage sample data includes a standard symptom description entity, a physical sign conclusion entity, identity vector information and a triage conclusion; Using the triage conclusion as a label to construct a training data sample to obtain a training data set; The RNN convolutional neural network is trained based on the training data set to minimize the set multi-classification cross entropy loss function to obtain a triage model.

[0016] In one embodiment of the present application, a plurality of triage sample data are extracted from the triage record, including: Extract symptom information, sign information, and identity information from each triage record; The symptom information is standardized and vectorized to obtain a standard symptom description entity; the sign information is conclusively labeled and vectorized to obtain a sign conclusion entity; the triage conclusion is standardized and vectorized to obtain a triage conclusion; the identity information is vectorized to obtain identity vector information.

[0017] In one embodiment of the present application, the multi-classification cross entropy loss function The mathematical expression is:

[0018] In the formula, is the sample size, is the sample serial number, is the number of categories, is the category number, For labels, Predict samples for the model Belongs to category probability.

[0019] In an embodiment of the present application, obtaining the user's current symptom description information includes: When receiving a triage start command from the outside, output a guidance voice or an initial guidance text; When a voice reply or text reply is received from the outside, a preliminary entity check is performed on the voice reply or the text reply. When the voice reply or the text reply contains a predetermined entity, the voice reply or the text reply is used as symptom description information.

[0020] The beneficial effects of the present invention are as follows: an interactive triage terminal based on AI analysis of the present invention extracts target entities from the collected symptom description information, thereby obtaining symptom entities, sign entities and identity information. Then, the confidence of the target entity is evaluated based on the pre-built triage database, so as to determine whether there are symptom and sign descriptions that are inconsistent with the identity information, or whether there are mismatched or low-probability symptom and sign combinations in the description information input by the user. If the above situation exists, the confidence level will be low. When the confidence level is low, the present application will construct a guide text based on the target entity and the triage database to guide the user to express correctly. When a correct expression with a higher confidence level is obtained, it is input into the triage model to perform AI analysis to obtain the triage result. The present application can effectively avoid triage errors caused by the user's inability to express correctly. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] The present invention will be further described below in conjunction with the accompanying drawings and embodiments: Figure 1This is an application scenario diagram of an interactive triage terminal based on AI analysis shown in an embodiment of the present application; Figure 2 is a structural diagram of an interactive triage terminal based on AI analysis shown in an embodiment of the present application; Figure 3 The process of constructing a collaborative association database in an embodiment of the present application; Figure 4 A schematic diagram of a reaction time calculation process in an embodiment of the present application; Figure 5 This is a schematic diagram of the process of detecting a network based on collaborative verification in this application. DETAILED DESCRIPTION

[0022] The following describes the embodiments of the present invention by specific examples, and those skilled in the art can easily understand other advantages and effects of the present invention from the contents disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments, and the details in this specification can also be modified or changed in various ways based on different viewpoints and applications without departing from the spirit of the present invention. It should be noted that the following embodiments and features in the embodiments can be combined with each other without conflict.

[0023] It should be noted that the illustrations provided in the following embodiments are only schematic illustrations of the basic concept of the present invention, and thus the drawings only show the layers related to the present invention rather than being drawn according to the number, shape and size of the layers in actual implementation. In actual implementation, the type, quantity and proportion of each layer may be changed arbitrarily, and the layer layout may also be more complicated.

[0024] In the following description, numerous details are discussed to provide a more thorough explanation of embodiments of the present invention; however, it is apparent to one skilled in the art that embodiments of the present invention may be practiced without these specific details.

[0025] Figure 1 is a hardware schematic diagram of an interactive triage terminal based on AI analysis shown in an embodiment of the present application, such as Figure 1 As shown, the interactive triage terminal in the present application is installed on the ground through a base. It is generally used in the waiting area outside the hospital clinic. The upper part of the terminal is provided with a touch screen 110 as an information input / output part. In addition, the terminal is provided with a microphone 120 and a speaker 130. The microphone 120 is used to collect the user's voice data, and the speaker 130 is used to interact with the user by voice. In addition, a computer is provided inside the terminal for executing the interactive logic in the present application.

[0026] Figure 2is a logical structure diagram of an interactive triage terminal based on AI analysis shown in an embodiment of the present application, such as Figure 2 As shown: An interactive triage terminal based on AI analysis in this embodiment includes: An acquisition module 210 is used to acquire the user's current symptom description information; The current symptom description information in this application can be text directly input by the user or audio data recorded by the user through a microphone. If it is audio data, Automatic Speech Recognition (ASR) technology is required to convert the voice data into text.

[0027] Specifically, this application adopts a guided process to obtain the user's current symptom description information, and the process includes: S211, upon receiving a triage start command from an external device, outputting a guidance voice or an initial guidance text; Among them, if the user touches the corresponding area on the terminal touch screen, a triage start command can be input. At startup, in order to avoid the user not knowing how to input the current symptom description information, the application presets an initial guide text as a template to guide the user to correctly input the description information.

[0028] For example, the initial guidance text may be: "Hello, please enter your triage information by speaking or typing on the keyboard, the reference example is XXX". At the same time, the terminal turns on the microphone and opens the virtual keyboard on the display screen. The user can perform input by speaking or typing on the virtual keyboard.

[0029] S212, when receiving a voice reply or a text reply from the outside, perform a preliminary entity verification on the voice reply or the text reply, and when the voice reply or the text reply contains a predetermined entity, use the voice reply or the text reply as symptom description information.

[0030] After receiving a reply, it is necessary to use the pre-built entity library to perform a preliminary check on the reply text. If the reply contains the predetermined entity, it can be considered that the user's input contains valid information and can be processed to the next step.

[0031] Specifically, when verifying, the reply text needs to be segmented, and the segmentation algorithm can be forward maximum matching, reverse maximum matching, bidirectional maximum matching, etc. The obtained words are vectorized and encoded, and quickly matched with the entity vector templates in the entity library. The entity templates in the entity library include common symptom or sign description entities such as "fever", "headache", and "high blood pressure".

[0032] An entity extraction module 220, configured to extract a plurality of target entities from the current symptom description information, wherein the plurality of target entities include symptom entities, sign entities and identity information; The process of target entity extraction is similar to the previous one, namely word segmentation and matching, which will not be repeated here.

[0033] An analysis module 230 is used to analyze the multiple target entities based on a pre-built triage database to obtain confidences of the multiple target entities, wherein the triage database includes entity combinations corresponding to multiple triage results; In this application, in order to verify whether the extracted target entity is reasonable, the triage database is used for verification. The big data of past triage is used to verify multiple target entities.

[0034] When a user makes a description error, it is usually because the more common symptoms do not match the identity information, or the more subtle symptom information does not match the physical sign information. For example, a female user describes that she has prostate pain, or a male user describes that he has fever symptoms, but the body temperature obtained through the preliminary physical examination before the initial triage is normal. At this time, it is necessary to verify this information through big data. Find entities that may have abnormalities and provide further guidance and confirmation for these abnormal entities.

[0035] The method for constructing the triage database in this application includes: (1) Obtain historical triage records; The triage record is an important part of medical documentation, which records in detail the key information of the patient's first contact with medical services. This record helps medical staff quickly assess the patient's urgency and health status and arrange the appropriate care path accordingly.

[0036] Generally, triage records contain the following information: Basic patient information, such as name, age or date of birth, gender, contact information, and consultation time; Chief Complaint: The main symptom or problem described by the patient, such as "chest pain" or "dyspnea" Symptoms and signs: Detailed description of symptoms, including but not limited to the nature, location, duration and intensity of pain; other significant signs such as fever and dyspnea.

[0037] Vital signs, measurement results of basic physiological indicators such as blood pressure, heart rate, respiratory rate, body temperature, and blood oxygen saturation.

[0038] Medical history, such as past medical history, such as hypertension, diabetes, etc.

[0039] Surgical history: past surgical experience.

[0040] The initial assessment is a rapid evaluation based on the patient's symptoms and signs and may include the use of a specific triage system (eg, CTAS, ESI) to determine the urgency of the condition.

[0041] The triage decision, based on all of the above information, determines whether the patient needs immediate attention or can wait, and which department or type of care they should be sent to.

[0042] In this application, user information, user-reported symptoms, and vital sign data generated by rapid testing before triage are mainly extracted.

[0043] (2) Extracting symptom entities, sign entities, identity information, and triage results recorded during multiple triages from the historical triage records , and standardize the symptom entity and the sign entity to obtain a standard symptom entity and standard signs entities , wherein the standard symptom entity and the standard sign entity both include body parts and feature descriptions; When building the database, in order to accurately match in the subsequent process. This application standardizes symptom entities and sign entities, and converts different types of text descriptions into standard formats. That is, body part + feature description; for example, "head + pain", "chest + dyspnea"; for sign data, there are some numbers of physical indicators. In order to simplify these numbers, they are uniformly converted into conclusive text, such as "body temperature + high", "blood pressure + medium-high".

[0044] In addition, triage results can also be standardized, for example, divided into department + urgency level.

[0045] (3) Based on the identity information, the symptom entities, sign entities and triage results recorded during multiple triages are divided to obtain data units; and based on the triage results, the symptom entities and sign entities in each data unit are divided to obtain symptom data sub-units and sign data sub-units; and each data unit is directly divided based on the triage results to obtain symptom-sign data sub-units; First, the above data is divided into multiple data units based on identity information. When dividing, since identity information is multidimensional, the resulting data units are also multidimensional. For example, three data units are obtained based on three age groups, and two data units are obtained based on gender. A total of five data units are obtained. The same is true for identity information of other dimensions, which will not be repeated here.

[0046] Then, based on the triage results, the data units are divided again to obtain symptom data subunits containing only symptom entities, sign data subunits containing only signs, and symptom-sign data subunits containing both sign entities and symptom entities.

[0047] When dividing the symptom-sign data into sub-units, it is necessary to extract the symptom entity and the sign entity in each triage record, and perform permutations and combinations to obtain a combination containing only one symptom entity and one sign entity.

[0048] (4) Calculate each symptom entity in the symptom data subunit The probability of occurrence , each physical sign entity in the physical sign data subunit The probability of occurrence ,in, is the entity serial number; (5) Extracting the symptom-sign entity combination of the same time from the symptom-sign data subunit , and calculate for each symptom-sign entity combination The probability of occurrence ; The probability here indicates the probability of different symptom entities or sign entities appearing in the triage results of the corresponding identity information. In addition, in order to further verify the probability of signs and symptoms appearing at the same time, each symptom-sign entity combination is also calculated. The probability of occurrence .

[0049] (6) Symptom entities with probabilities less than the preset probability threshold , physical entity and symptom-sign entity combinations Remove from the data unit and update the reference symptom entity in the symptom data subunit The probability of occurrence , the reference physical sign entity in the physical sign data subunit The probability of occurrence , and updating the reference symptom-sign entity combination in the symptom-sign data subunit The probability of occurrence , and obtain the triage database.

[0050] In this application, the entities or entity combinations that appear occasionally with a small probability are removed, and the remaining entities or entity combinations can reflect the rationality of multiple entities or entity combinations when different triage results appear in a certain user information group. The higher the probability, the higher the rationality of the appearance.

[0051] Wherein, the reference symptom entity The probability of occurrence The mathematical expression of the reference sign entity The probability of occurrence , and the reference symptom-sign entity combination The probability of occurrence They are:

[0052]

[0053]

[0054] In the formula, is the total number of reference symptom entities in the symptom data subunit, is the total number of reference sign entities in the sign data subunit, is the total number of reference symptom-sign entity combinations in the symptom-sign data subunit.

[0055] Finally, the above entities and entity combinations need to be converted into vectors. This application uses Word2Vec encoding to convert the above entities and entity combinations into vectors and store them in the database.

[0056] After constructing the above triage database, the process of using the above triage database to perform confidence analysis on multiple target entities is as follows: S231, determining a target data unit in the triage database based on the identity information of the user; The target data unit records all symptom entities, sign entities, symptom and sign combinations and their respective probabilities corresponding to all historical triage conclusions.

[0057] S232, comparing the symptom entity in the current description information with the reference symptom entities in multiple symptom data subunits, comparing the sign entity in the current description information with the reference sign entities in multiple sign data subunits, and comparing the symptom-sign entity combination in the current description information with the reference symptom-sign entity combination in the symptom-sign data subunit; Before making a comparison, it is necessary to construct multiple entity combinations based on the symptom entities and sign entities extracted in the previous text.

[0058] For example, from the current symptom description information "I feel stuffy nose, chills and headache" and the physical sign description "body temperature 38.2, heart rate 92", three symptom entities are extracted: nose + congestion, body + chills, head + pain; physical sign entities: high body temperature, moderate heart rate.

[0059] Six groups of entity combinations can be obtained, namely: Nose + blocked, high body temperature; Body + Chills, high body temperature; Head + pain, high temperature; nose + congestion, moderate heartbeat; Body + chills, moderate heartbeat; Head + pain, moderate heartbeat; By comparing these six entity combinations with the reference symptom-sign entity combinations in the database, the rationality of the occurrence of the combinations can be judged.

[0060] S233, when there is no state entity consistent with the reference symptom entity in the current description information, or when there is no sign entity consistent with the reference sign entity, the confidences of the multiple target entities are set to zero, and it is determined that the current description information does not match the identity information of the user; If the entity in the database does not exist in the current description information, it means that there is a high probability that the symptoms or signs described do not match the identity information. At this time, the confidence of multiple target entities is set to zero.

[0061] S234, there is a symptom entity corresponding to the reference symptom entity in the current description information A state entity that is consistent with the reference sign entity When the physical sign entity is consistent, the probability of occurrence corresponding to the same triage result , and Perform weighted summation to obtain the confidence of multiple target entities .

[0062] If the entity in the database exists in the current description information, the rationality of the current symptom description information is verified by calculating the confidence.

[0063] Since it is necessary to judge the rationality of multiple target entities appearing in the same information description, it is necessary to find the corresponding entities and the probability of entity combination from the symptom data subunit, sign data subunit, and symptom-sign combination data subunit corresponding to the same triage result, that is, , and Then perform weighted summation to get the confidence Confidence The mathematical expression is:

[0064] in, is the first confidence weight, is the second confidence weight, is the third confidence weight.

[0065] A result judgment module 240 is used to input the multiple target entities into a pre-built triage model to obtain a triage result when the confidence of the multiple target entities is greater than or equal to a preset confidence threshold; and to construct a guidance text based on the triage database and the multiple target entities when the confidence of the multiple target entities is less than the preset confidence threshold, and send the guidance text to the user; Confidence describes the reasonableness (comprehensive probability) of multiple target entities appearing in the same description information. Therefore, if the confidence is relatively high, it means that the user's description is in line with the historical triage rules. At this time, the extracted multiple target entities are input into the triage model, and the triage results can be obtained based on AI analysis.

[0066] If the confidence is low or even 0, it means that the user's description does not conform to the historical triage rules, or the probability of conforming to the rules is low. At this time, further confirmation and guidance are required, and the process is as follows: S241, when the confidence of the multiple target entities is zero, constructing a guidance text for prompting that the symptom or sign does not match the identity information; A confidence score of zero only occurs when the described symptoms or signs do not match the identity information. In this case, the main purpose is to prompt the user that the symptom description is incorrect, such as generating a guide text: "The symptom description may be incorrect, please re-enter or contact staff." S242, when the confidence of the multiple target entities is not zero, construct a confirmation guidance text based on the symptom entity and the sign entity in the target entity; and based on the symptom entity or the sign entity, confirm the reference symptom entity in the target data unit. Or reference sign entity Make a rating , and based on the top N reference symptom entities with the highest scores Or reference sign entity Construct reference guide text, where scoring The mathematical expression is:

[0067] In the formula, is the first weight, is the second weight, is the cosine similarity between the symptom entity and the reference symptom entity, is the cosine similarity between the physical sign entity and the reference physical sign entity.

[0068] If the confidence is not zero, but relatively small, then a confirmation guide text and a reference guide text are generated respectively.

[0069] For example, confirm the guide text: "Are you sure you have the following symptoms: (1) XXX, (2) XXX..." At this point, if the user is still unable to organize the correct language to describe, further guidance is needed.

[0070] This application uses similarity and probability to comprehensively evaluate reference entities. First, similar symptom entities or sign entities with the highest similarity to the target entity are matched from the target data unit. Then, the maximum probability is selected from the probabilities corresponding to the multiple triage results. The maximum probability and the similarity of similar symptom entities are weighted summed. The entity with the highest score may be the symptom entity or sign entity that the user wants to express. Use the highest-scoring entity to construct a reference guide text, for example: “Do you want to describe the symptoms or signs: (1) XXX, (2) XXX…”; The above-mentioned confirmation guide text and reference guide text can be used to effectively prompt the user. After the user consults the guide text, if the user confirms again, it is possible that the user has indeed experienced symptoms and signs with a low probability. Therefore, when receiving the confirmation description from the user, the confidence of the multiple target entities is set to a target value greater than or equal to the preset confidence threshold. Then the multiple target entities are input into the triage model to complete the triage.

[0071] The loop module 250 is used to, when receiving the user's symptom description information, use the symptom description information as the current symptom description information and return to the entity extraction module until there are multiple target entities whose confidence is greater than or equal to a preset confidence threshold, and then input the multiple target entities into a pre-built triage model to obtain a triage result, wherein the symptom description information is generated based on the guide text.

[0072] If the user does not make further confirmation but inputs new symptom description information, the entity extraction module is returned to perform a round of entity extraction and analysis. If the user describes again according to the reference guide text, then the second result will most likely show multiple target entities with a confidence greater than or equal to the preset confidence threshold. Otherwise, the cycle continues, and after multiple target entities with a confidence greater than or equal to the preset confidence threshold appear, the multiple target entities are input into the pre-built triage model to obtain the triage result.

[0073] By performing triage through the above process, it is possible to promptly discover unreasonable descriptions of users and guide users to make correct descriptions, thereby improving the triage accuracy of the triage model.

[0074] In this application, AI analysis is used for triage, and the basis for triage is mainly user symptoms and signs. If the user is an elderly person or a child, age also needs to be considered. Therefore, triage can be regarded as a multi-classification problem that classifies samples containing user symptoms, signs, and users. This multi-classification scenario is very suitable for execution using RNN convolutional neural networks. Therefore, the construction method of the triage model based on RNN convolutional neural networks includes: (1) Obtain triage records; (2) extracting a plurality of triage sample data from the triage record, wherein the triage sample data includes a standard symptom description entity, a physical sign conclusion entity, identity vector information, and a triage conclusion; The triage records are in text format, so they need to be standardized to improve training efficiency and model performance. The annotation processing includes: extracting symptom information, sign information and identity information from each triage record; standardizing and vectorizing the symptom information to obtain a standard symptom description entity; conclusively annotating and vectorizing the sign information to obtain a sign conclusion entity; standardizing and vectorizing the triage conclusion to obtain a triage conclusion; vectorizing the identity information to obtain identity vector information.

[0075] The vectorized encoding method used in this application all adopts Word2Vec.

[0076] (3) constructing training data samples using the triage conclusions as labels to obtain a training data set; (4) Training the RNN convolutional neural network based on the training data set to minimize the set multi-classification cross entropy loss function and obtain a triage model.

[0077] In one embodiment of the present application, the multi-classification cross entropy loss function The mathematical expression is:

[0078] In the formula, is the sample size, is the sample serial number, is the number of categories, is the category number, For labels, Predict samples for the model Belongs to category probability.

[0079] The triage model obtained through the above training process can effectively use multiple target entities verified in the previous article to perform triage.

[0080] This application can not only perform automatic triage, but also replace human verification and error correction of user's expression problems to a certain extent. In actual application scenarios, it has strong practicality.

[0081] An interactive triage terminal based on AI analysis of the present invention extracts target entities from the collected symptom description information, thereby obtaining symptom entities, sign entities and identity information. Then, the confidence of the target entity is evaluated based on the pre-built triage database, so as to determine whether there are symptom and sign descriptions that are inconsistent with the identity information, or whether there are mismatched or low-probability symptom and sign combinations in the description information input by the user. If the above situation exists, the confidence level will be low. When the confidence level is low, the present application will construct a guide text based on the target entity and the triage database to guide the user to express correctly. When a correct expression with a higher confidence level is obtained, it is input into the triage model to perform AI analysis, and the triage result can be obtained. The present application can effectively avoid triage errors caused by the user's inability to express correctly.

[0082] The present application also provides an interactive triage method based on AI analysis, comprising the steps of: S310, obtaining the user's current symptom description information; S320, extracting multiple target entities from the current symptom description information, wherein the multiple target entities include symptom entities, sign entities, and identity information; S330, analyzing the multiple target entities based on a pre-built triage database to obtain confidences of the multiple target entities, wherein the triage database includes entity combinations corresponding to multiple triage results; S340, when the confidence of the multiple target entities is greater than or equal to a preset confidence threshold, the multiple target entities are input into a pre-built triage model to obtain a triage result; when the confidence of the multiple target entities is less than the preset confidence threshold, a guidance text is constructed based on the triage database and the multiple target entities, and the guidance text is sent to the user; S350, when receiving the user's symptom description information, use the symptom description information as the current symptom description information, and return to the entity extraction module, until there are multiple target entities whose confidence is greater than or equal to the preset confidence threshold, input the multiple target entities into the pre-built triage model to obtain the triage result, and the symptom description information is generated based on the guide text.

[0083] The present invention discloses an interactive triage method based on AI analysis, which extracts target entities from the collected symptom description information to obtain symptom entities, sign entities and identity information. Then, the confidence of the target entity is evaluated based on the pre-built triage database to determine whether there are symptom and sign descriptions that are inconsistent with the identity information, or whether there are mismatched or low-probability symptom and sign combinations in the description information input by the user. If the above situation exists, the confidence level will be low. When the confidence level is low, the present application will construct a guide text based on the target entity and the triage database to guide the user to express correctly. When a correct expression with a higher confidence level is obtained, it is input into the triage model to perform AI analysis to obtain the triage result. The present application can effectively avoid triage errors caused by the user's inability to express correctly.

[0084] This embodiment further provides a computer-readable storage medium on which a computer program is stored. When the program is executed by a processor, any one of the methods in this embodiment is implemented, wherein the method is the execution logic of this system.

[0085] This embodiment also provides an electronic terminal, including: a processor and a memory; The memory is used to store a computer program, and the processor is used to execute the computer program stored in the memory, so that the terminal executes any one of the methods in this embodiment.

[0086] The computer-readable storage medium in this embodiment can be understood by ordinary technicians in this field: all or part of the steps of implementing the above-mentioned method embodiments can be completed by hardware related to the computer program. The aforementioned computer program can be stored in a computer-readable storage medium. When the program is executed, the steps of the above-mentioned method embodiments are executed; and the aforementioned storage medium includes: ROM, RAM, magnetic disk or optical disk and other media that can store program codes.

[0087] The electronic terminal provided in this embodiment includes a processor, a memory, a transceiver and a communication interface. The memory and the communication interface are connected to the processor and the transceiver and complete communication with each other. The memory is used to store computer programs, the communication interface is used to communicate, and the processor and the transceiver are used to run computer programs so that the electronic terminal executes each step of the above method.

[0088] In this embodiment, the memory may include a random access memory (RAM), and may also include a non-volatile memory (non-volatile memory), such as at least one disk memory.

[0089] The above-mentioned processor can be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it can also be a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field programmable gate array (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components.

[0090] In the above-mentioned embodiments, although the present invention has been described in conjunction with the specific embodiments of the present invention, many replacements, modifications and variations of these embodiments will be apparent to those of ordinary skill in the art based on the foregoing description. The embodiments of the present invention are intended to cover all such replacements, modifications and variations falling within the broad scope of the appended claims.

[0091] The above embodiments are merely illustrative of the principles and effects of the present invention, and are not intended to limit the present invention. Anyone familiar with the art may modify or alter the above embodiments without departing from the spirit and scope of the present invention. Therefore, all equivalent modifications or alterations made by a person of ordinary skill in the art without departing from the spirit and technical concept disclosed by the present invention shall still be covered by the claims of the present invention.

Claims

1. An interactive triage terminal based on AI analysis, characterized in that: include: The acquisition module is used to obtain the user's current symptom description information; An entity extraction module, used to extract multiple target entities from the current symptom description information, wherein the multiple target entities include symptom entities, sign entities and identity information; An analysis module, configured to analyze the multiple target entities based on a pre-built triage database to obtain confidences of the multiple target entities, wherein the triage database includes entity combinations corresponding to multiple triage results; A result judgment module is used to input the multiple target entities into a pre-built triage model to obtain triage results when the confidence of the multiple target entities is greater than or equal to a preset confidence threshold; and to construct a guidance text based on the triage database and the multiple target entities when the confidence of the multiple target entities is less than the preset confidence threshold, and send the guidance text to the user; A loop module is used to, when receiving the user's symptom description information, use the symptom description information as the current symptom description information and return to the entity extraction module until there are multiple target entities whose confidence is greater than or equal to a preset confidence threshold, and then input the multiple target entities into a pre-built triage model to obtain a triage result, wherein the symptom description information is generated based on the guide text.

2. The interactive triage terminal based on AI analysis according to claim 1, characterized in that: The method for constructing the triage database comprises: Obtain historical triage records; Extract the symptom entities, sign entities, identity information and triage results recorded during multiple triages from the historical triage records , and standardize the symptom entity and the sign entity to obtain a standard symptom entity and standard signs entities , wherein the standard symptom entity and the standard sign entity both include body parts and feature descriptions; The symptom entities, sign entities and triage results recorded during multiple triages are divided based on identity information to obtain data units; and the symptom entities and sign entities in each data unit are divided based on the triage results to obtain symptom data subunits and sign data subunits; and each data unit is directly divided based on the triage results to obtain symptom-sign data subunits; Calculate each symptom entity in the symptom data subunit The probability of occurrence , each physical sign entity in the physical sign data subunit The probability of occurrence ,in, is the entity serial number; Extract the symptom-sign entity combination of the same time from the symptom-sign data subunit , and calculate for each symptom-sign entity combination The probability of occurrence ; Symptom entities with a probability less than a preset probability threshold , physical entity and symptom-sign entity combinations Remove from the data unit and update the reference symptom entity in the symptom data subunit The probability of occurrence , the reference physical sign entity in the physical sign data subunit The probability of occurrence , and updating the reference symptom-sign entity combination in the symptom-sign data subunit The probability of occurrence , and obtain the triage database.

3. The interactive triage terminal based on AI analysis according to claim 2, characterized in that: Analyzing the multiple target entities based on a pre-built triage database to obtain confidences of the multiple target entities includes: Determining a target data unit in the triage database based on the identity information of the user; Comparing the symptom entity in the current description information with the reference symptom entities in multiple symptom data subunits, comparing the sign entity in the current description information with the reference sign entities in multiple sign data subunits, and comparing the symptom-sign entity combination in the current description information with the reference symptom-sign entity combination in the symptom-sign data subunits; When there is no state entity consistent with the reference symptom entity in the current description information, or when there is no sign entity consistent with the reference sign entity, the confidences of the multiple target entities are set to zero, and it is determined that the current description information does not match the identity information of the user; There is a reference symptom entity in the current description information A state entity that is consistent with the reference sign entity When the physical sign entity is consistent, the probability of occurrence corresponding to the same triage result , and Perform weighted summation to obtain the confidence of multiple target entities .

4. The interactive triage terminal based on AI analysis according to claim 3, characterized in that: Constructing a guidance text based on the triage database and the multiple target entities includes: When the confidence of the multiple target entities is zero, constructing a guide text for indicating that the symptom or sign is inconsistent with the identity information; When the confidence of the multiple target entities is not zero, a confirmation guidance text is constructed based on the symptom entity and the sign entity in the target entity; and a confirmation guidance text is constructed based on the symptom entity or the sign entity for the reference symptom entity in the target data unit. Or reference sign entity Make a rating , and based on the top N reference symptom entities with the highest scores Or reference sign entity Construct reference guide text, where scoring The mathematical expression is: In the formula , is the first weight, is the second weight, is the cosine similarity between the symptom entity and the reference symptom entity, is the cosine similarity between the physical sign entity and the reference physical sign entity.

5. The interactive triage terminal based on AI analysis according to claim 1, characterized in that: After sending the guide text to the user, the method further includes: When receiving the confirmation description from the user, the confidences of the multiple target entities are set to target values ​​that are greater than or equal to a preset confidence threshold.

6. The interactive triage terminal based on AI analysis according to claim 2, characterized in that: The reference symptom entity The probability of occurrence The mathematical expression of the reference sign entity The probability of occurrence , and the reference symptom-sign entity combination The probability of occurrence They are: In the formula, is the total number of reference symptom entities in the symptom data subunit, is the total number of reference sign entities in the sign data subunit, is the total number of reference symptom-sign entity combinations in the symptom-sign data subunit.

7. The interactive triage terminal based on AI analysis according to claim 1, characterized in that: The method for constructing the triage model includes: Obtain triage records; Extracting a plurality of triage sample data from the triage record, wherein the triage sample data includes a standard symptom description entity, a physical sign conclusion entity, identity vector information and a triage conclusion; Using the triage conclusion as a label to construct a training data sample to obtain a training data set; The RNN convolutional neural network is trained based on the training data set to minimize the set multi-classification cross entropy loss function to obtain a triage model.

8. The interactive triage terminal based on AI analysis according to claim 7, characterized in that: Extracting multiple triage sample data from the triage record, including: Extract symptom information, sign information, and identity information from each triage record; The symptom information is standardized and vectorized to obtain a standard symptom description entity; the sign information is conclusively labeled and vectorized to obtain a sign conclusion entity; the triage conclusion is standardized and vectorized to obtain a triage conclusion; the identity information is vectorized to obtain identity vector information.

9. The interactive triage terminal based on AI analysis according to claim 7, characterized in that: The multi-classification cross entropy loss function The mathematical expression is: In the formula, is the sample size, is the sample serial number, is the number of categories, is the category number, For labels, Predict samples for the model Belongs to category probability.

10. The interactive triage terminal based on AI analysis according to claim 1, characterized in that: Get the user's current symptom description information, including: When receiving a triage start command from the outside, output a guidance voice or an initial guidance text; When a voice reply or text reply is received from the outside, a preliminary entity check is performed on the voice reply or the text reply. When the voice reply or the text reply contains a predetermined entity, the voice reply or the text reply is used as symptom description information.

Images