CN112201359B

CN112201359B - Method and device for identifying severe inquiry data based on artificial intelligence

Info

Publication number: CN112201359B
Application number: CN202011065413.XA
Authority: CN
Inventors: 满晏松; 柳恭; 李响
Original assignee: Ping An Technology Shenzhen Co Ltd
Current assignee: Ping An Technology Shenzhen Co Ltd
Priority date: 2020-09-30
Filing date: 2020-09-30
Publication date: 2024-05-03
Anticipated expiration: 2040-09-30
Also published as: WO2022068160A1; CN112201359A

Abstract

The application relates to artificial intelligence and provides an artificial intelligence-based severe inquiry data identification method and device. The method for identifying the severe consultation data based on the artificial intelligence comprises the following steps: acquiring inquiry session data corresponding to a target user identifier; inputting the inquiry session data into a prediction model, and outputting a model identification result corresponding to the inquiry session data through the prediction model; determining expert recognition results corresponding to the consultation session data according to the labels hit by the consultation session data in the expert knowledge base; and combining the model identification result and the expert identification result to obtain a target identification result of whether the inquiry session data is critical inquiry data. The method can improve the identification accuracy of the severe inquiry data. In addition, the application also relates to a blockchain technology, and inquiry session data of a user can be stored in the blockchain.

Description

Method and device for identifying severe inquiry data based on artificial intelligence

Technical Field

The application relates to the technical field of artificial intelligence, in particular to an artificial intelligence-based severe inquiry data identification method and device.

Background

With the development of internet technology and medical technology, the application of internet technology in the medical industry is becoming more and more popular. For example, the user may self-address symptoms, consult conditions, learn about medications, seek out instruction for a visit, etc. through an online consultation application or an online consultation website. The doctor can judge whether the current inquiry data is the severe inquiry data or not through consulting the etiology, symptoms and other multi-round inquiries by manual consultation.

However, with the proliferation of users initiating on-line interrogation, the medical field began to adopt a mode of expert system for discrimination in order to alleviate the heavy work of manual authentication of doctors. Although expert systems may alleviate the workload of the relevant personnel, expert systems present a significant challenge to the accuracy of the identification of whether the current interrogation data is critical interrogation data due to the diversity of the on-line user interrogation data.

Disclosure of Invention

In view of the foregoing, it is desirable to provide an artificial intelligence-based severe consultation data identification method, apparatus, computer device and storage medium that can improve the accuracy of the artificial intelligence-based severe consultation data identification.

An artificial intelligence-based severe consultation data identification method, which is characterized by comprising the following steps:

Acquiring inquiry session data corresponding to a target user identifier;

Inputting the inquiry session data into a prediction model, and outputting a model identification result corresponding to the inquiry session data through the prediction model; the prediction model is obtained through training according to multi-dimensional feature data and training labels corresponding to the multi-dimensional feature data, the multi-dimensional feature data is extracted according to historical inquiry data, and the multi-dimensional feature data comprises entity feature data and entity relation feature data;

Determining expert recognition results corresponding to the consultation session data according to the labels hit by the consultation session data in an expert knowledge base;

And combining the model identification result and the expert identification result to obtain a target identification result of whether the inquiry session data is severe inquiry data.

In one embodiment, the inputting the interview session data into a prediction model, outputting a model identification result corresponding to the interview session data through the prediction model, includes:

Inputting the inquiry session data into a prediction model, and processing the inquiry session data through a plurality of neurons included in the prediction model to obtain a model identification result corresponding to the inquiry session data;

Outputting the model identification result through the prediction model;

The threshold value of each neuron and the weight of the connection relation between the neurons are determined when the prediction model is trained through multi-dimensional feature data and training labels corresponding to the multi-dimensional feature data, and the multi-dimensional feature data further comprises intention feature data.

In one embodiment, the multi-dimensional feature data further comprises intent feature data; the method further comprises the steps of:

Collecting historical inquiry data and training labels corresponding to the historical inquiry data; the training label corresponding to the historical inquiry data is used for indicating whether the historical inquiry data is severe inquiry data or not;

Extracting entity characteristic data, entity relation characteristic data and intention characteristic data from the historical inquiry data, and generating multidimensional characteristic data corresponding to the historical inquiry data;

The historical inquiry data and multidimensional characteristic data corresponding to the historical inquiry data are input into a prediction model to be trained together to obtain a prediction recognition result;

And training the prediction model based on the prediction recognition result of the prediction model and the training label.

In one embodiment, the extracting the entity feature data, the entity relationship feature data and the intention feature data from the historical inquiry data, and generating the multidimensional feature data corresponding to the historical inquiry data includes:

Extracting entity characteristic data corresponding to the historical inquiry data by adopting a rule engine drive;

Carrying out named entity identification on the historical inquiry data by adopting a sequence labeling model, and extracting entity relation characteristic data among the named entities based on the identified named entities;

extracting intention characteristic data in the historical inquiry data by adopting a semantic model;

And generating multidimensional feature data corresponding to the historical inquiry data according to the entity feature data, the entity relation feature data and the intention feature data.

In one embodiment, the extracting entity characteristic data corresponding to the historical query data using a rule engine driver includes:

Word segmentation is carried out on the historical inquiry data to obtain word segmentation results;

Based on the entity tag provided by the expert knowledge base, a rule engine is adopted to drive the entity tag to be extracted according to the word segmentation result, and entity characteristic data corresponding to the historical inquiry data is obtained;

the extracting the intention characteristic data in the historical inquiry data by adopting a semantic model comprises the following steps:

and extracting the intention labels corresponding to the historical inquiry data by adopting a semantic model based on the intention labels provided by the expert knowledge base to obtain intention characteristic data corresponding to the historical inquiry data.

In one embodiment, the method further comprises:

Screening historical severe inquiry data from the historical inquiry data;

counting occurrence frequency of entity keywords in the historical severe consultation data;

screening entity keywords with occurrence frequencies higher than a preset threshold value, and adding the entity keywords serving as entity tags into an expert knowledge base;

Determining the inquiry intention corresponding to the history severe inquiry data to obtain an intention label;

adding the intention labels to the expert knowledge base.

In one embodiment, the interview session data is stored in a blockchain; the method further comprises the steps of:

When the obtained inquiry session data is the target identification result of the severe inquiry data, accessing the inquiry session to which the inquiry session data belongs to a doctor terminal;

And when the inquiry session data is not the target identification result of the severe inquiry data, continuing to push the inquiry session to which the inquiry session data belongs.

An artificial intelligence based intensive care data recognition apparatus, the apparatus comprising:

the acquisition module is used for acquiring inquiry session data corresponding to the target user identification;

The first recognition module is used for inputting the inquiry session data into a prediction model and outputting a model recognition result corresponding to the inquiry session data through the prediction model; the prediction model is obtained through training according to multi-dimensional feature data and training labels corresponding to the multi-dimensional feature data, the multi-dimensional feature data is extracted according to historical inquiry data, and the multi-dimensional feature data comprises entity feature data and entity relation feature data;

the second recognition module is used for determining expert recognition results corresponding to the consultation session data according to the labels hit by the consultation session data in the expert knowledge base;

And the decision module is used for combining the model identification result and the expert identification result to obtain a target identification result of whether the inquiry session data is severe inquiry data.

A computer device comprising a memory storing a computer program and a processor which when executing the computer program performs the steps of:

Acquiring inquiry session data corresponding to a target user identifier;

A computer storage medium having stored thereon a computer program which when executed by a processor performs the steps of:

Acquiring inquiry session data corresponding to a target user identifier;

According to the method, the device, the computer equipment and the storage medium for identifying the severe consultation data based on the artificial intelligence, after the consultation session data of the user are obtained, the consultation session data are input into the prediction model to obtain a model identification result, and the consultation session data are identified based on the expert knowledge base to obtain an expert identification result, and the model identification result and the expert identification result are combined to obtain a final identification result. Because the prediction model is obtained based on multi-dimensional feature data training, and the multi-dimensional feature data comprises entity feature data and entity relation feature data, the prediction model can learn information with different dimensions in the training process and better understand language logic by combining with a context semantic environment, so that the recognition capability of the prediction model on severe inquiry data is improved; therefore, the mode of combining model prediction and expert system is adopted to identify the severe consultation data, the defect that only the expert system is relied on can be overcome, and the accuracy of the identification of the severe consultation data is improved.

Drawings

FIG. 1 is an application scenario diagram of an artificial intelligence based method for identifying severe cases data in one embodiment;

FIG. 2 is a flow diagram of an artificial intelligence based method for identifying severe cases data in one embodiment;

FIG. 3 is a flow diagram of using a predictive model in one embodiment;

FIG. 4 is a block flow diagram of training a predictive model in one embodiment;

FIG. 5 is a block diagram of an artificial intelligence based severe consultation data identification apparatus according to an embodiment;

FIG. 6 is a block diagram of an artificial intelligence based severe consultation data identifying apparatus according to another embodiment;

FIG. 7 is an internal block diagram of a computer device in one embodiment;

Fig. 8 is an internal structural view of a computer device in another embodiment.

Detailed Description

The present application will be described in further detail with reference to the drawings and examples, in order to make the objects, technical solutions and advantages of the present application more apparent. It should be understood that the specific embodiments described herein are for purposes of illustration only and are not intended to limit the scope of the application.

The severe inquiry data identification method based on artificial intelligence provided by the application can be applied to an application environment shown in figure 1. Wherein the terminal 102 communicates with the server 104 via a network. Specifically, the terminal 102 may acquire the inquiry session data corresponding to the target user identifier, send the inquiry session data to the server 104, and the server 104 inputs the inquiry session data into the prediction model, and outputs a model identification result corresponding to the inquiry session data through the prediction model; the prediction model is obtained through training according to multi-dimensional feature data and training labels corresponding to the multi-dimensional feature data, the multi-dimensional feature data is obtained through extraction according to historical inquiry data, and the multi-dimensional feature data comprises entity feature data and entity relation feature data; the server 104 then determines expert recognition results corresponding to the consultation session data according to the tags hit in the expert knowledge base by the consultation session data; the server 104 combines the model recognition result and the expert recognition result to obtain a target recognition result of whether the inquiry session data is critical inquiry data. The terminal 102 may be, but not limited to, various personal computers, notebook computers, smartphones, tablet computers, and portable wearable devices, and the server 104 may be implemented by a stand-alone server or a server cluster composed of a plurality of servers.

In other embodiments, the terminal 102 or the server 104 may also be used separately to perform the artificial intelligence based critical inquiry data identification method, respectively. The application is not limited herein.

In one embodiment, as shown in fig. 2, an artificial intelligence based method for identifying severe consultation data is provided, and the method is applied to a computer device, which may be specifically a terminal or a server in fig. 1, for example. The method for identifying the severe consultation data based on the artificial intelligence specifically comprises the following steps:

Step 202, acquiring inquiry session data corresponding to a target user identifier.

Wherein the target user identification is used to uniquely identify a user. The target user identification such as an application account number or a hospital visit card number, etc. A interviewing session is a process of interviewing interactions between at least two users. The at least two users include a user corresponding to a patient role and a user corresponding to a doctor role. The inquiry session data is data generated during an inquiry interaction. The user corresponding to the doctor role may be the doctor himself or herself or an artificial intelligent robot.

Specifically, the terminal may be running an online consultation application or an online consultation website, which may provide a consultation portal. The user inputs inquiry session data based on an inquiry portal through the terminal to conduct an on-line inquiry.

In one embodiment, the interview session data may be voice data, text data, or image data, among others.

In one embodiment, the interview session data may include interview session data corresponding to a patient role and interview session data corresponding to a doctor role. Wherein, the inquiry session data corresponding to the patient role, such as user basic information, symptom description information, symptom photograph, medical examination report or past history information, etc. It will be appreciated that the querying user may or may not be the patient himself, such as in the case of a child or elderly query. The inquiry session data corresponding to the doctor's role, such as disease description information, symptom analysis information, cause analysis information, or reply information to user inquiry data, etc.

In one embodiment, the interview session data may be dialogue data of one or more rounds of interviews during a single interview. Therefore, the inquiry session data provided by the user for the first time can be identified as early as possible to carry out corresponding processing when the inquiry session data is critical inquiry data; when the information quantity is insufficient, the user is guided to provide more information so as to accurately identify the severe inquiry data by combining the information.

The severe consultation data are consultation session data related to the critical severe symptoms. "acute critical" is a medical term that generally indicates that the patient is suffering from an emergency or endangered condition, which should be treated as early as possible, or else may cause serious injury to the patient's body or cause death. Such as clinical manifestation symptoms of critical illness, etc. included in the session data. Clinical manifestations of critical illness such as "syncope", "dyspnea", etc.

Step 204, inputting the inquiry session data into a prediction model, and outputting a model identification result corresponding to the inquiry session data through the prediction model; the prediction model is obtained through training according to multi-dimensional feature data and training labels corresponding to the multi-dimensional feature data, the multi-dimensional feature data is obtained through extraction according to historical inquiry data, and the multi-dimensional feature data comprises entity feature data and entity relation feature data.

The prediction model is a machine learning model which is trained in advance and used for identifying whether the consultation session data is severe consultation data or not. The machine learning model may employ a neural network model, a support vector machine, a logistic regression model, or the like. Neural network models such as convolutional neural networks, back propagation neural networks, feedback neural networks, radial basis neural networks, or ad hoc neural networks, and the like.

Entity characteristics are data reflecting the characteristics of the entity itself. For example, the inquiry session data "drinking during supper, now abdominal pain is hard to endure" includes two entities, the first being "drinking" and the second being "abdominal pain". Entity relationship characteristics refer to data reflecting a relationship between at least two entities. For example, the inquiry session data "drinking during supper" is that the physical relationship between "drinking" and "abdominal pain" is now "causal", i.e. the cause of abdominal pain is drinking. Here, on one hand, consider that the entity in the inquiry session data is an important basis for identifying severe inquiry data, and on the other hand, consider that the relationship between different entities also affects the identification result, even the same entity has different semantics in different contexts, and thus affects the identification result. For example, the inquiry session data "big aunt, abdominal pain refractory" medium "big aunt" refers to menses, not to relative names. The computer device can fuse various information when designing the input data of the prediction model, for example, the information is fused in two characteristic dimensions, namely the entity dimension and the entity relationship dimension, of the data to be used as the input data of the prediction model, so that the prediction model can learn the effective information of the two characteristic dimensions in training, and the recognition capability of the model on severe inquiry data is improved.

Specifically, the computer device may input the inquiry session data into the prediction model, process the inquiry session data through a plurality of neurons included in the prediction model, obtain a model recognition result corresponding to the inquiry session data, and output the model recognition result through the prediction model. Where a neuron is the most basic structure in a neural network, most neurons are typically in a suppressed state, but when a neuron receives input information, causing its potential to exceed a threshold, the neuron is activated and in an "excited" state, which propagates the output information to other neurons. The connection lines connecting the neurons correspond to one weight (the value of which is called a weight value), and usually different connection lines correspond to different weights. The threshold value of each neuron and the weight of the connection relation between each neuron are determined when the prediction model is trained by the multidimensional feature data and the training labels corresponding to the multidimensional feature data. Neurons include input neurons, output neurons, and implicit neurons.

In one embodiment, the method for identifying severe cases based on artificial intelligence further comprises a training step of a predictive model, wherein the training step specifically comprises: collecting historical inquiry data and training labels corresponding to the historical inquiry data; the training label corresponding to the historical inquiry data is used for indicating whether the historical inquiry data is severe inquiry data or not; extracting entity characteristic data and entity relation characteristic data from the historical inquiry data, and generating multidimensional characteristic data corresponding to the historical inquiry data; the historical inquiry data and multidimensional feature data corresponding to the historical inquiry data are input into a prediction model to be trained together to obtain a prediction recognition result; and training the prediction model based on the prediction recognition result of the prediction model and the training label.

In one embodiment, the multi-dimensional feature data further includes intent feature data. The training step of the prediction model at this time specifically includes: collecting historical inquiry data and training labels corresponding to the historical inquiry data; the training label corresponding to the historical inquiry data is used for indicating whether the historical inquiry data is severe inquiry data or not; extracting entity characteristic data, entity relation characteristic data and intention characteristic data from the historical inquiry data, and generating multidimensional characteristic data corresponding to the historical inquiry data; the historical inquiry data and multidimensional feature data corresponding to the historical inquiry data are input into a prediction model to be trained together to obtain a prediction recognition result; and training the prediction model based on the prediction recognition result of the prediction model and the training label.

Wherein the intent feature is data reflecting the intent of the expression. The intention feature includes feature data reflecting doctor query or response intention, and user query or response intention. For example, the questioning session data "you are now not abdominal pain refractory" is intended to express a symptom confirmation intention; the inquiry session data "yes, i now have abdominal pain refractory" what is to be expressed is symptom confirmation intention, and so on.

Therefore, considering that the intentions of doctors and users for inquiring and answering can also be used as the basis for identifying the severe consultation data, the intention characteristic can be introduced when the input of the prediction model is involved, and the three characteristic dimensions of the entity dimension, the entity relation dimension and the intention dimension are fused, so that the prediction model can learn the effective information of the three characteristic dimensions in training, the identification capability of the model on the severe consultation data is improved, and the identification direction of the severe consultation data can be further enlarged.

For details of the predictive model training step reference may be made to the details in the following examples.

In one embodiment, the computer device may convert the inquiry session data into a data format that can be processed by the predictive model, and then input the converted data into the predictive model. The predictive model can handle data formats such as vector formats or matrix formats.

And step 206, determining expert recognition results corresponding to the consultation session data according to the labels hit by the consultation session data in the expert knowledge base.

The knowledge base refers to a rule set applied by expert system design, and comprises facts and data related by the rules, and the facts and the data form the knowledge base. The knowledge base is related to a specific expert system, and in the present application, the expert knowledge base is related to a consultation expert system in the medical field. The rules in the expert knowledge base are obtained by extracting data which frequently appear in the severe consultation data in the history consultation process, facts related to the rules comprise whether the consultation session data in the history consultation process are severe consultation data, and the data related to the rules comprise labels such as disease symptoms frequently appearing in the consultation session data in the history consultation process.

In particular, the computer device may employ a rules engine driven to determine tags for which the interview session data hits in the expert knowledge base based on a set of rules of the expert knowledge base. When the hit label is a label corresponding to severe consultation data, obtaining expert identification results corresponding to consultation session data, wherein the expert identification results are severe consultation data; and when the hit label is not the label corresponding to the severe consultation data, obtaining the expert recognition result corresponding to the consultation session data as the consultation session data is not the severe consultation data.

Step 208, combining the model identification result and the expert identification result to obtain a target identification result of whether the inquiry session data is serious inquiry data.

Specifically, the model recognition results include two recognition results that the inquiry session data is severe inquiry data and that the inquiry session data is not severe inquiry data. The expert recognition results also include two recognition results that the inquiry session data is severe inquiry data and that the inquiry session data is not severe inquiry data. When the model recognition result and the expert recognition result are both the inquiry session data and the severe inquiry data, the target recognition result that the inquiry session data is the severe inquiry data can be obtained. At least one of the model recognition result and the expert recognition result is that the inquiry session data is not the severe inquiry data, then the target recognition result that the inquiry session data is not the severe inquiry data can be obtained.

In one embodiment, the expert recognition result may also include a case where whether the inquiry session data is the critical inquiry data is not recognized. At this time, the model recognition result may be used as the target recognition result.

According to the severe consultation data identification method based on the artificial intelligence, after the consultation session data of the user are obtained, the consultation session data are input into the prediction model to obtain the model identification result, the consultation session data are identified based on the expert knowledge base to obtain the expert identification result, and the model identification result and the expert identification result are combined to obtain the final identification result. Because the prediction model is obtained based on multi-dimensional feature data training, and the multi-dimensional feature data comprises entity feature data and entity relation feature data, the prediction model can learn information with different dimensions in the training process and better understand language logic by combining with a context semantic environment, so that the recognition capability of the prediction model on severe inquiry data is improved; therefore, the mode of combining model prediction and expert system is adopted to identify the severe consultation data, the defect that only the expert system is relied on can be overcome, and the accuracy of the identification of the severe consultation data is improved.

In one embodiment, after obtaining the target recognition result of whether the inquiry session data is the critical inquiry data, the computer device may perform an operation corresponding to the target recognition result according to the target recognition result.

In a specific embodiment, the method for identifying severe cases based on artificial intelligence further comprises: when the obtained inquiry session data is the target identification result of the severe inquiry data, accessing the inquiry session to which the inquiry session data belongs to a doctor terminal; and when the obtained inquiry session data is not the target identification result of the severe inquiry data, continuing to advance the inquiry session to which the inquiry session data belongs.

Specifically, FIG. 3 illustrates a flow diagram for using a predictive model in one embodiment. Referring to the figure, after acquiring the inquiry session data, the computer equipment can input the inquiry session data into a prediction model and an expert system in parallel, on one hand, whether the inquiry session data is severe inquiry data is identified through the prediction model to obtain a model identification result, and on the other hand, whether the inquiry session data is severe inquiry data is identified through the expert system to obtain an expert identification result. And then, combining the model identification result and the expert identification result through a decision maker to obtain a target identification result of whether the inquiry session data is critical inquiry data. When the obtained inquiry session data is the target identification result of the severe inquiry data, the inquiry session to which the inquiry session data belongs is accessed to a doctor terminal, and the doctor can check the final identification result of the severe inquiry data through manual intervention of the doctor terminal, and can further process in time when the identification is correct, such as giving out a consultation suggestion and the like. And when the inquiry session data is not the target identification result of the severe inquiry data, continuing to advance the inquiry session to which the inquiry session data belongs, for example, continuing to interact with the user through the artificial intelligent robot to conduct inquiry.

In this embodiment, when different target recognition results are obtained, a corresponding next operation is immediately performed, so that a user who needs help in emergency in case of emergency can be effectively responded when the inquiry session data is severe inquiry data, and on-line inquiry can be sequentially continued when the inquiry session data is not severe inquiry data.

In one embodiment, the interview session data is stored in a blockchain. It should be emphasized that, to further ensure the privacy and security of the interview session data, the interview session data may also be stored in a blockchain node.

For the specific content of the training step of the prediction model involved in the foregoing embodiment, reference may be made to the specific description in the following embodiment.

In one embodiment, the artificial intelligence based severe cases data identification method further comprises: collecting historical inquiry data and training labels corresponding to the historical inquiry data; the training label corresponding to the historical inquiry data is used for indicating whether the historical inquiry data is severe inquiry data or not; extracting entity characteristic data and entity relation characteristic data from the historical inquiry data, and generating multidimensional characteristic data corresponding to the historical inquiry data; the historical inquiry data and multidimensional feature data corresponding to the historical inquiry data are input into a prediction model to be trained together to obtain a prediction recognition result; and training the prediction model based on the prediction recognition result of the prediction model and the training label.

In one embodiment, the multi-dimensional feature data further includes intent feature data; the severe consultation data identification method based on artificial intelligence also comprises the following steps: collecting historical inquiry data and training labels corresponding to the historical inquiry data; the training label corresponding to the historical inquiry data is used for indicating whether the historical inquiry data is severe inquiry data or not; extracting entity characteristic data, entity relation characteristic data and intention characteristic data from the historical inquiry data, and generating multidimensional characteristic data corresponding to the historical inquiry data; the historical inquiry data and multidimensional feature data corresponding to the historical inquiry data are input into a prediction model to be trained together to obtain a prediction recognition result; and training the prediction model based on the prediction recognition result of the prediction model and the training label.

Specifically, the computer device may collect historical inquiry data and training tags corresponding to the historical inquiry data. The training label may be a manually labeled result, indicating whether the historical data is critical data. The computer device may then extract the entity characteristic data and the entity relationship characteristic data from the historical interview data, generate multidimensional characteristic data corresponding to the historical interview data, the multidimensional characteristic data including at least two characteristic dimensions, and use the multidimensional characteristic data and the historical interview data together as input data for a predictive model to be trained. The computer device may also extract entity feature data, entity relationship feature data, and intent feature data from the historical interview data, generate multidimensional feature data corresponding to the historical interview data, the multidimensional feature data including at least three feature dimensions, and use the multidimensional feature data and the historical interview data together as input data for a predictive model to be trained. Wherein generating the multi-dimensional feature data based on the plurality of feature data may be stitching or fusing the plurality of feature data.

And finally, the computer equipment can obtain a prediction recognition result output by the prediction model to be trained, then construct a training loss function according to the difference between the prediction recognition result and the training label, optimize the parameters of the prediction model by adopting a back propagation algorithm according to the direction of minimizing the training loss function, and obtain weights and thresholds of all the implicit neurons and the input/output neurons after training, thereby obtaining a model parameter file of the trained prediction model. The computer device may store the model parameter file in an expert knowledge base.

In this embodiment, when the computer device designs the input data of the prediction model, multiple aspects of information are fused, for example, two feature dimensions, namely, the entity dimension and the entity relationship dimension of the data are fused to be used as the input data of the prediction model, so that the prediction model can learn at least the effective information of the two feature dimensions in training, and the recognition capability of the model on the severe consultation data is improved.

In addition, the intentions of doctors and users for inquiring and answering can be taken into consideration as the basis for identifying the severe consultation data, so that the intension features can be introduced when the input data of the prediction model is designed, and fusion is carried out on three feature dimensions, namely the entity dimension, the entity relation dimension and the intension dimension, so that the prediction model can learn at least the effective information of the three feature dimensions in training, the identification capability of the model on the severe consultation data is improved, and the identification direction of the severe consultation data can be enlarged.

It will be appreciated that in actual use, the feature extraction is performed in a number of ways, mainly from the following aspects:

In one embodiment, extracting entity feature data, entity relationship feature data, and intention feature data from historical query data, generating multidimensional feature data corresponding to the historical query data includes: adopting a rule engine to drive and extract entity characteristic data corresponding to the historical inquiry data; carrying out named entity identification on the historical inquiry data by adopting a sequence labeling model, and extracting entity relation characteristic data among named entities based on the identified named entities; extracting intention characteristic data in the historical inquiry data by adopting a semantic model; and generating multidimensional feature data corresponding to the historical inquiry data according to the entity feature data, the entity relation feature data and the intention feature data.

(1) In an aspect of extracting entity characteristic data, in a specific embodiment, the extracting entity characteristic data corresponding to the historical query data by using a rule engine driver includes: word segmentation is carried out on the historical inquiry data to obtain word segmentation results; based on the entity tag provided by the expert knowledge base, a rule engine is adopted to drive the entity tag to be extracted according to the word segmentation result, and entity characteristic data corresponding to the historical inquiry data is obtained.

Specifically, the computer device may segment the historical inquiry data to obtain a segmentation result, and extract the entity tag from the segmentation result by using a rule engine according to rules in a rule set of the expert knowledge base, according to the entity tag provided from the expert knowledge base, to obtain entity feature data. Such as: entity tag: physical characterization data → patient: infants, symptoms: vomiting, weight loss, etc. The rule engine is a component embedded in the application program, accepts data input, interprets rules, and makes decisions based on the rules.

Therefore, when the entity characteristic data is extracted, the effective data in the expert knowledge base is effectively utilized, the characteristic extraction efficiency and effectiveness are improved, and unnecessary training time for long tail words is avoided.

(2) In the aspect of extracting entity relationship feature data, in a specific embodiment, the computer device may group the historical inquiry data according to inquiry sessions, generate an inquiry data sequence according to response sequence from the historical inquiry data of one inquiry session, then identify named entities of the inquiry data sequence by adopting a sequence labeling model, and extract entity relationship feature data between the named entities based on the identified named entities. Such as named entities (alcohol, causes, abdominal pain), and the extraction yields entity relationships: the cause of abdominal pain is drinking. Therefore, when named entity identification and entity relation extraction are carried out, the sequence-based processing mode can be effectively combined with the context information, more accurate relation extraction between entities can be carried out, and the method has wide practical significance.

(3) In an intention feature data extraction aspect, in a specific embodiment, extracting intention feature data in historical query data using a semantic model includes: and extracting an intention label corresponding to the historical inquiry data by adopting a semantic model based on the intention label provided by the expert knowledge base to obtain intention characteristic data corresponding to the historical inquiry data.

In particular, the computer device may also employ the semantic model in the third aspect to extract the intent features of the historical query data during the stage of training the predictive model. The computer device can obtain the historical inquiry data as a sample, manually label the intention type label (the intention label provided by the expert knowledge base), supervise and train the semantic model, and then extract the intention characteristics of the historical inquiry data by using the trained semantic model.

In one embodiment, the artificial intelligence based severe cases data identification method further comprises: screening historical severe inquiry data from the historical inquiry data; counting occurrence frequency of entity keywords in historical severe consultation data; screening entity keywords with occurrence frequencies higher than a preset threshold value, and adding the entity keywords serving as entity tags into an expert knowledge base; determining inquiry intentions corresponding to the history severe inquiry data to obtain intent labels; the intent labels are added to the expert knowledge base.

In particular, the computer device may also build an expert knowledge base. The computer equipment can firstly acquire the severe inquiry data identified in the historical inquiry data to obtain a severe sample data set; and then counting the occurrence frequency of entity keywords such as causes, objects, symptoms and the like included in the severe sample data set, screening entity keywords with occurrence frequency higher than a preset threshold value, and adding the entity keywords serving as entity labels into an expert knowledge base. In addition, the computer equipment can also determine the inquiry intention corresponding to the history severe inquiry data to obtain an intention label; the intent labels are added to the expert knowledge base.

Therefore, when the multidimensional features of the training samples are extracted during training of the prediction model, the multidimensional features of the training samples can be extracted by accurately inducing the summarized labels by using an expert system, so that unnecessary training time for long tail words is avoided, and the training time for training of the machine learning model is greatly prolonged.

By way of example, FIG. 4 illustrates a block flow diagram for training a predictive model in one embodiment. With reference to the figure, it can be seen that the computer device may perform preparation first, that is, collect historical interview data and training tags corresponding to the historical interview data, and then extract entity feature data from the historical interview data based on various data provided by the expert knowledge base on the one hand, extract entity relationship feature data from the historical interview data based on semantic understanding on the other hand, and extract intention feature data from the historical interview data in combination with various data and semantic understanding provided by the expert knowledge base. And the computer equipment fuses the characteristic data extracted from the three aspects, combines the historical inquiry data as input data, constructs a neural network structure, and trains to obtain a prediction model for identifying severe inquiry data.

It should be understood that, although the steps in the flowcharts of the above embodiments are shown in order as indicated by the arrows, these steps are not necessarily performed in order as indicated by the arrows. The steps are not strictly limited to the order of execution unless explicitly recited herein, and the steps may be executed in other orders. Moreover, at least a portion of the steps in the flowcharts of the above embodiments may include a plurality of steps or a plurality of stages, which are not necessarily performed at the same time, but may be performed at different times, and the order of the execution of the steps or stages is not necessarily sequential, but may be performed in turn or alternately with at least a portion of the steps or stages in other steps or others.

In one embodiment, as shown in fig. 5, there is provided an artificial intelligence based intensive care data recognition apparatus including: an acquisition module 501, a first identification module 502, a second identification module 503, and a decision module 504, wherein:

an obtaining module 501, configured to obtain inquiry session data corresponding to a target user identifier;

the first recognition module 502 is configured to input the inquiry session data into a prediction model, and output a model recognition result corresponding to the inquiry session data through the prediction model; the prediction model is obtained through training according to multi-dimensional feature data and training labels corresponding to the multi-dimensional feature data, the multi-dimensional feature data is extracted according to historical inquiry data, and the multi-dimensional feature data comprises entity feature data and entity relation feature data;

a second identifying module 503, configured to determine an expert identifying result corresponding to the consultation session data according to a label hit by the consultation session data in an expert knowledge base;

And the decision module 504 is configured to combine the model identification result and the expert identification result to obtain a target identification result of whether the inquiry session data is critical inquiry data.

In one embodiment, the first identifying module 502 is further configured to input the interview session data into a prediction model, and process the interview session data through a plurality of neurons included in the prediction model to obtain a model identifying result corresponding to the interview session data; outputting a model identification result through a prediction model; the threshold value of each neuron and the weight of the connection relation between the neurons are determined when the prediction model is trained through the multidimensional feature data and the training labels corresponding to the multidimensional feature data, and the multidimensional feature data further comprises intention feature data.

As shown in fig. 6, in one embodiment, the multi-dimensional feature data further includes intent feature data; the severe consultation data identification device based on artificial intelligence further comprises: the training module 505 is configured to collect historical inquiry data and training labels corresponding to the historical inquiry data; the training label corresponding to the historical inquiry data is used for indicating whether the historical inquiry data is severe inquiry data or not; extracting entity characteristic data, entity relation characteristic data and intention characteristic data from the historical inquiry data, and generating multidimensional characteristic data corresponding to the historical inquiry data; the historical inquiry data and multidimensional feature data corresponding to the historical inquiry data are input into a prediction model to be trained together to obtain a prediction recognition result; and training the prediction model based on the prediction recognition result of the prediction model and the training label.

In one embodiment, the training module 505 is further configured to extract entity feature data corresponding to the historical query data using a rule engine driver; carrying out named entity identification on the historical inquiry data by adopting a sequence labeling model, and extracting entity relation characteristic data among named entities based on the identified named entities; extracting intention characteristic data in the historical inquiry data by adopting a semantic model; and generating multidimensional feature data corresponding to the historical inquiry data according to the entity feature data, the entity relation feature data and the intention feature data.

In one embodiment, the training module 505 is further configured to perform word segmentation on the historical query data to obtain a word segmentation result; based on the entity tag provided by the expert knowledge base, a rule engine is adopted to drive the entity tag to be extracted according to the word segmentation result, and entity characteristic data corresponding to the historical inquiry data is obtained; and extracting an intention label corresponding to the historical inquiry data by adopting a semantic model based on the intention label provided by the expert knowledge base to obtain intention characteristic data corresponding to the historical inquiry data.

In one embodiment, the training module 505 is further configured to screen historical severe consultation data from the historical consultation data; counting occurrence frequency of entity keywords in historical severe consultation data; screening entity keywords with occurrence frequencies higher than a preset threshold value, and adding the entity keywords serving as entity tags into an expert knowledge base; determining inquiry intentions corresponding to the history severe inquiry data to obtain intent labels; the intent labels are added to the expert knowledge base.

In one embodiment, the interview session data is stored in a blockchain; the decision module 504 is further configured to access a consultation session to which the consultation session data belongs to the doctor terminal when the consultation session data is obtained as a target identification result of the severe consultation data; and when the obtained inquiry session data is not the target identification result of the severe inquiry data, continuing to advance the inquiry session to which the inquiry session data belongs.

According to the severe consultation data identification device based on the artificial intelligence, after the consultation session data of the user are obtained, the consultation session data are input into the prediction model to obtain the model identification result, the consultation session data are identified based on the expert knowledge base to obtain the expert identification result, and the model identification result and the expert identification result are combined to obtain the final identification result. Because the prediction model is obtained based on multi-dimensional feature data training, and the multi-dimensional feature data comprises entity feature data and entity relation feature data, the prediction model can learn information with different dimensions in the training process and better understand language logic by combining with a context semantic environment, so that the recognition capability of the prediction model on severe consultation data based on artificial intelligence is improved; the method adopts a mode of combining model prediction and expert system to identify the severe consultation data, can make up for the defect of only relying on the expert system, and improves the accuracy of the severe consultation data identification based on artificial intelligence.

For specific limitations on the artificial intelligence-based severe consultation data identification means, reference may be made to the above limitations on the artificial intelligence-based severe consultation data identification method, and the detailed description thereof will be omitted. The above-mentioned critical issue inquiry data identification apparatus based on artificial intelligence may be implemented in whole or in part by software, hardware, and a combination thereof. The above modules may be embedded in hardware or may be independent of a processor in the computer device, or may be stored in software in a memory in the computer device, so that the processor may call and execute operations corresponding to the above modules.

In one embodiment, a computer device is provided, which may be a server, the internal structure of which may be as shown in fig. 7. 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 includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, computer programs, and a database. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage media. The database of the computer device is used for storing the critical inquiry data identification data based on artificial intelligence. The network interface of the computer device is used for communicating with an external terminal through a network connection. The computer program, when executed by a processor, implements an artificial intelligence based method for identifying severe cases data.

In one embodiment, a computer device is provided, which may be a terminal, and the internal structure thereof may be as shown in fig. 8. The computer device includes a processor, a memory, a network interface, a display screen, and an input device 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 includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage media. The network interface of the computer device is used for communicating with an external terminal through a network connection. The computer program, when executed by a processor, implements an artificial intelligence based method for identifying severe cases data. The display screen of the computer equipment can be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer equipment can be a touch layer covered on the display screen, can also be keys, a track ball or a touch pad arranged on the shell of the computer equipment, and can also be an external keyboard, a touch pad or a mouse and the like.

It will be appreciated by those skilled in the art that the structures shown in fig. 7 or 8 are merely block diagrams of portions of structures associated with aspects of the application and are not intended to limit the computer device to which aspects of the application may be applied, and that a particular computer device may include more or fewer components than those shown, or may combine certain components, or may have a different arrangement of components.

In one embodiment, a computer device is provided comprising a memory storing a computer program and a processor that when executing the computer program performs the steps of: acquiring inquiry session data corresponding to a target user identifier; inputting the inquiry session data into a prediction model, and outputting a model identification result corresponding to the inquiry session data through the prediction model; the prediction model is obtained through training according to multi-dimensional feature data and training labels corresponding to the multi-dimensional feature data, the multi-dimensional feature data is obtained through extraction according to historical inquiry data, and the multi-dimensional feature data comprises entity feature data and entity relation feature data; determining expert recognition results corresponding to the consultation session data according to the labels hit by the consultation session data in the expert knowledge base; and combining the model identification result and the expert identification result to obtain a target identification result of whether the inquiry session data is critical inquiry data.

In one embodiment, inputting the interview session data into the predictive model, outputting a model identification result corresponding to the interview session data through the predictive model, comprising: inputting the inquiry session data into a prediction model, and processing the inquiry session data through a plurality of neurons included in the prediction model to obtain a model identification result corresponding to the inquiry session data; outputting a model identification result through a prediction model; the threshold value of each neuron and the weight of the connection relation between the neurons are determined when the prediction model is trained through the multidimensional feature data and the training labels corresponding to the multidimensional feature data, and the multidimensional feature data further comprises intention feature data.

In one embodiment, the multi-dimensional feature data further includes intent feature data. The processor when executing the computer program also implements the steps of: collecting historical inquiry data and training labels corresponding to the historical inquiry data; the training label corresponding to the historical inquiry data is used for indicating whether the historical inquiry data is severe inquiry data or not; extracting entity characteristic data, entity relation characteristic data and intention characteristic data from the historical inquiry data, and generating multidimensional characteristic data corresponding to the historical inquiry data; the historical inquiry data and multidimensional feature data corresponding to the historical inquiry data are input into a prediction model to be trained together to obtain a prediction recognition result; and training the prediction model based on the prediction recognition result of the prediction model and the training label.

In one embodiment, the method for extracting entity characteristic data corresponding to the historical query data by using a rule engine driver comprises the following steps: word segmentation is carried out on the historical inquiry data to obtain word segmentation results; based on the entity tag provided by the expert knowledge base, a rule engine is adopted to drive the entity tag to be extracted according to the word segmentation result, and entity characteristic data corresponding to the historical inquiry data is obtained. Extracting intention characteristic data in the historical inquiry data by adopting a semantic model comprises the following steps: and extracting an intention label corresponding to the historical inquiry data by adopting a semantic model based on the intention label provided by the expert knowledge base to obtain intention characteristic data corresponding to the historical inquiry data.

In one embodiment, the processor when executing the computer program further performs the steps of: screening historical severe inquiry data from the historical inquiry data; counting occurrence frequency of entity keywords in historical severe consultation data; screening entity keywords with occurrence frequencies higher than a preset threshold value, and adding the entity keywords serving as entity tags into an expert knowledge base; determining inquiry intentions corresponding to the history severe inquiry data to obtain intent labels; the intent labels are added to the expert knowledge base.

In one embodiment, the interview session data is stored in a blockchain; the processor when executing the computer program also implements the steps of: when the obtained inquiry session data is the target identification result of the severe inquiry data, accessing the inquiry session to which the inquiry session data belongs to a doctor terminal; and when the obtained inquiry session data is not the target identification result of the severe inquiry data, continuing to advance the inquiry session to which the inquiry session data belongs.

In one embodiment, a computer storage medium is provided having a computer program stored thereon, which when executed by a processor, performs the steps of: acquiring inquiry session data corresponding to a target user identifier; inputting the inquiry session data into a prediction model, and outputting a model identification result corresponding to the inquiry session data through the prediction model; the prediction model is obtained through training according to multi-dimensional feature data and training labels corresponding to the multi-dimensional feature data, the multi-dimensional feature data is obtained through extraction according to historical inquiry data, and the multi-dimensional feature data comprises entity feature data and entity relation feature data; determining expert recognition results corresponding to the consultation session data according to the labels hit by the consultation session data in the expert knowledge base; and combining the model identification result and the expert identification result to obtain a target identification result of whether the inquiry session data is critical inquiry data.

In one embodiment, the multi-dimensional feature data further includes intent feature data; the computer program when executed by the processor also performs the steps of: collecting historical inquiry data and training labels corresponding to the historical inquiry data; the training label corresponding to the historical inquiry data is used for indicating whether the historical inquiry data is severe inquiry data or not; extracting entity characteristic data, entity relation characteristic data and intention characteristic data from the historical inquiry data, and generating multidimensional characteristic data corresponding to the historical inquiry data; the historical inquiry data and multidimensional feature data corresponding to the historical inquiry data are input into a prediction model to be trained together to obtain a prediction recognition result; and training the prediction model based on the prediction recognition result of the prediction model and the training label.

In one embodiment, the computer program when executed by the processor further performs the steps of: screening historical severe inquiry data from the historical inquiry data; counting occurrence frequency of entity keywords in historical severe consultation data; screening entity keywords with occurrence frequencies higher than a preset threshold value, and adding the entity keywords serving as entity tags into an expert knowledge base; determining inquiry intentions corresponding to the history severe inquiry data to obtain intent labels; the intent labels are added to the expert knowledge base.

In one embodiment, the interview session data is stored in a blockchain; the computer program when executed by the processor also performs the steps of: when the obtained inquiry session data is the target identification result of the severe inquiry data, accessing the inquiry session to which the inquiry session data belongs to a doctor terminal; and when the obtained inquiry session data is not the target identification result of the severe inquiry data, continuing to advance the inquiry session to which the inquiry session data belongs.

The blockchain is a novel application mode of computer technologies such as distributed data storage, point-to-point transmission, consensus mechanism, encryption algorithm and the like. The blockchain (Blockchain), essentially a de-centralized database, is a string of data blocks that are generated in association using cryptographic methods, each of which contains information from a batch of network transactions for verifying the validity (anti-counterfeit) of its information and generating the next block. The blockchain may include a blockchain underlying platform, a platform product services layer, an application services layer, and the like.

Those skilled in the art will appreciate that implementing all or part of the above described methods may be accomplished by way of a computer program stored on a non-transitory computer readable storage medium, which when executed, may comprise the steps of the embodiments of the methods described above. Any reference to memory, storage, database, or other medium used in embodiments provided herein may include non-volatile and/or volatile memory. The nonvolatile 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 (SYNCHLINK) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), among others.

The technical features of the above embodiments may be arbitrarily combined, and all possible combinations of the technical features in the above embodiments are not described for brevity of description, however, as long as there is no contradiction between the combinations of the technical features, they should be considered as the scope of the description.

The above examples illustrate only a few embodiments of the application, which are described in detail and are not to be construed as limiting the scope of the application. It should be noted that it will be apparent to those skilled in the art that several variations and modifications can be made without departing from the spirit of the application, which are all within the scope of the application. Accordingly, the scope of protection of the present application is to be determined by the appended claims.

Claims

1. An artificial intelligence-based severe consultation data identification method, which is characterized by comprising the following steps:

Acquiring inquiry session data corresponding to a target user identifier; the inquiry session data includes at least one of voice data, text data, and image data;

Inputting the inquiry session data into a prediction model, and outputting a model identification result corresponding to the inquiry session data through the prediction model; the prediction model is obtained through training according to multi-dimensional feature data and training labels corresponding to the multi-dimensional feature data, the multi-dimensional feature data is obtained through extraction according to historical inquiry data, and the multi-dimensional feature data comprises entity feature data and entity relation feature data;

combining the model identification result and the expert identification result to obtain a target identification result of whether the inquiry session data is severe inquiry data or not;

The method further comprises the steps of:

Screening historical severe inquiry data from the historical inquiry data;

adding the intention labels to the expert knowledge base;

The multi-dimensional feature data further includes intent feature data; the method further comprises the steps of:

extracting intention labels corresponding to the historical inquiry data by adopting a semantic model based on the intention labels provided by the expert knowledge base to obtain intention characteristic data corresponding to the historical inquiry data;

generating multidimensional feature data corresponding to the historical inquiry data according to the entity feature data, the entity relation feature data and the intention feature data;

2. The artificial intelligence-based severe consultation data identification method according to claim 1, characterized in that inputting the consultation session data into a prediction model, outputting a model identification result corresponding to the consultation session data through the prediction model, includes:

Outputting the model identification result through the prediction model;

The threshold value of each neuron and the weight of the connection relation between the neurons are determined when the prediction model is trained through the multidimensional feature data and the training labels corresponding to the multidimensional feature data.

3. The artificial intelligence based severe consultation data identification method of claim 1 or 2 wherein the consultation session data is stored in a blockchain;

The method further comprises the steps of:

4. An artificial intelligence based intensive care data recognition apparatus, the apparatus comprising:

The acquisition module is used for acquiring inquiry session data corresponding to the target user identification; the inquiry session data includes at least one of voice data, text data, and image data;

The first recognition module is used for inputting the inquiry session data into a prediction model and outputting a model recognition result corresponding to the inquiry session data through the prediction model; the prediction model is obtained through training according to multi-dimensional feature data and training labels corresponding to the multi-dimensional feature data, the multi-dimensional feature data is obtained through extraction according to historical inquiry data, and the multi-dimensional feature data comprises entity feature data and entity relation feature data;

the decision module is used for combining the model identification result and the expert identification result to obtain a target identification result of whether the inquiry session data is severe inquiry data or not;

The training module is used for screening historical severe consultation data from the historical consultation data; counting occurrence frequency of entity keywords in the historical severe consultation data; screening entity keywords with occurrence frequencies higher than a preset threshold value, and adding the entity keywords serving as entity tags into an expert knowledge base; determining the inquiry intention corresponding to the history severe inquiry data to obtain an intention label; adding the intention labels to the expert knowledge base;

The multi-dimensional feature data further comprises intention feature data, and the training module is further used for collecting historical inquiry data and training labels corresponding to the historical inquiry data; the training label corresponding to the historical inquiry data is used for indicating whether the historical inquiry data is severe inquiry data or not; word segmentation is carried out on the historical inquiry data to obtain word segmentation results; based on the entity tag provided by the expert knowledge base, a rule engine is adopted to drive the entity tag to be extracted according to the word segmentation result, and entity characteristic data corresponding to the historical inquiry data is obtained; carrying out named entity identification on the historical inquiry data by adopting a sequence labeling model, and extracting entity relation characteristic data among the named entities based on the identified named entities; extracting intention labels corresponding to the historical inquiry data by adopting a semantic model based on the intention labels provided by the expert knowledge base to obtain intention characteristic data corresponding to the historical inquiry data; generating multidimensional feature data corresponding to the historical inquiry data according to the entity feature data, the entity relation feature data and the intention feature data; the historical inquiry data and multidimensional characteristic data corresponding to the historical inquiry data are input into a prediction model to be trained together to obtain a prediction recognition result; and training the prediction model based on the prediction recognition result of the prediction model and the training label.

5. The artificial intelligence based severe cases data identifying apparatus according to claim 4, wherein,

The first recognition module is used for inputting the inquiry session data into a prediction model, and processing the inquiry session data through a plurality of neurons included in the prediction model to obtain a model recognition result corresponding to the inquiry session data; outputting the model identification result through the prediction model; the threshold value of each neuron and the weight of the connection relation between the neurons are determined when the prediction model is trained through the multidimensional feature data and the training labels corresponding to the multidimensional feature data.

6. The artificial intelligence based severe consultation data identification apparatus of claim 4 or 5 wherein the consultation session data is stored in a blockchain;

the decision module is used for accessing the inquiry session to which the inquiry session data belongs to a doctor terminal when the inquiry session data is obtained as a target identification result of the severe inquiry data; and when the inquiry session data is not the target identification result of the severe inquiry data, continuing to push the inquiry session to which the inquiry session data belongs.

7. A computer device comprising a memory and a processor, the memory storing a computer program, characterized in that the processor implements the steps of the artificial intelligence based critical inquiry data recognition method of any of claims 1 to 3 when the computer program is executed.

8. A computer storage medium having stored thereon a computer program, which when executed by a processor, implements the steps of the artificial intelligence based critical inquiry data recognition method of any of claims 1 to 3.