CN110874409A - Disease grading prediction system, method, electronic device and readable storage medium - Google Patents

Abstract

The present application discloses a disease classification prediction system, method, electronic device and storage medium based on electronic medical records. The system includes: storing electronic medical record information through a first storage module, reading the electronic medical record information through a first information filter and filtering the electronic medical record information, for structured and textual multimodal data, through deep learning-based The hierarchical predictor combined with the attention mechanism uses the hierarchical prediction model to process the structured data and text data to obtain text feature mining information, disease classification prediction results and visual analysis text. By using the classification prediction system of the present application, the electronic medical record information can be processed relatively quickly and the disease classification prediction results can be obtained. Improve grading speed and grading accuracy.

Description

Disease grading prediction system, method, electronic device and readable storage medium

technical field

The present invention relates to the technical field of information processing, and in particular, to a disease classification prediction system, method, electronic device and readable storage medium.

Background technique

Triage, as the first gateway for emergency patients to see a doctor, is the most important factor affecting emergency department congestion. According to the severity of the patient's condition, it is helpful to make full use of emergency resources, maintain the order of emergency patients, shorten the waiting time of critically ill patients, improve work efficiency, and prevent emergency resources from being exhausted in advance due to insufficient triage or excessive triage.

The advanced triage standards in foreign countries all adopt disease grading and are generated in a specific medical environment and social background. However, domestic and foreign medical insurance systems, emergency models and medical treatment methods are quite different, so it is impossible to copy or apply theirs. Therefore, it is necessary to establish a set of simple, effective and scientific emergency pre-examination and triage standards that are in line with international standards and in line with my country's national conditions.

At present, my country mainly adopts the empirical triage model, which has higher requirements on the doctor's level, and the triage speed is relatively slow. With the sharp increase in the number of emergency patients, it is easy to cause insufficient or excessive triage, resulting in emergency department. A situation where resources are exhausted prematurely.

SUMMARY OF THE INVENTION

In view of the above problems, embodiments of the present invention are proposed to provide a disease classification prediction system, method, electronic device, and readable storage medium that overcome the above problems or at least partially solve the above problems.

In a first aspect, an embodiment of the present application provides a disease classification prediction system based on an electronic medical record, the system comprising:

a first storage module for storing electronic medical record information, the electronic medical record information including objective indicators, current illness history and physical examination results;

a first information filter, connected to the first storage module, for reading the electronic medical record information from the first storage module and filtering the electronic medical record information to obtain filtered electronic medical record information, including structured data and textual data;

The classification predictor is used for inputting the structured data and the text data into the emergency disease classification prediction model respectively to obtain text feature mining information, disease classification prediction results and visual analysis text.

Optionally, the emergency condition classification prediction model includes a representation model and a fusion model, and the classification predictor includes:

a first representation module, configured to input the text data into the representation model to obtain the visual analysis text, the text feature mining information and the text feature vector;

a second representation module, for inputting the structured data into the representation model to obtain a structured feature vector;

The fusion module is used for inputting the structured feature vector and the text feature vector into the fusion model to obtain the disease classification prediction result.

Optionally, the first presentation module includes:

A mapping submodule for inputting the text data into the representation model, mapping each Chinese character of the text data to a vector, and extracting the context information embedded in the text data to obtain an embedded vector;

An attention submodule, for performing attention processing on the embedded vector to obtain the visual analysis text and text feature vector;

The extraction submodule is used to extract phrases from the visual analysis text to obtain the text feature mining information.

Optionally, the system further includes:

The second storage module is used to store a plurality of electronic medical record information samples;

The second information filter is connected to the second storage module, and is used for reading the multiple electronic medical record information samples from the second storage module and filtering the multiple electronic medical record information samples to obtain multiple electronic medical record information samples. Filtered medical record information samples, including multiple structured data samples and text data samples;

A model trainer is used to divide the multiple structured data samples and text data samples into a training set, a verification set and a test set, and perform multiple rounds of training on the preset model until the evaluation score of the trained preset model is in the If there is no increase in the preset number of rounds, the training is stopped, and the model corresponding to the highest evaluation score is determined as the emergency disease classification prediction model.

In a second aspect, the embodiments of the present application further provide a method for predicting disease classification based on electronic medical records, the method comprising:

storing electronic medical record information, the electronic medical record information including objective indicators, current illness history and physical examination results;

reading the electronic medical record information and filtering the electronic medical record information to obtain filtered electronic medical record information, including structured data and text data;

The structured data and the text data are respectively input into the emergency condition classification prediction model to obtain text feature mining information, disease classification prediction results and visual analysis text.

Optionally, the emergency disease classification prediction model includes a representation model and a fusion model, and the structured data and the text data are respectively input into the emergency disease classification prediction model to obtain text feature mining information, disease classification prediction results and visual analysis. text, including:

Inputting the text data into the representation model to obtain the visual analysis text, the text feature mining information and the text feature vector;

Inputting the structured data into the representation model to obtain a structured feature vector;

Inputting the structured feature vector and the text feature vector into the fusion model to obtain the disease classification prediction result.

Optionally, inputting the text data into a representation model to obtain the visual analysis text, the text feature mining information and the text feature vector, including:

Inputting the text data into the representation model, mapping each Chinese character of the text data to a vector, and extracting the context information embedded in the text data to obtain an embedded vector;

Perform attention processing on the embedded vector to obtain the visual analysis text and text feature vector;

Phrase extraction is performed on the visual analysis text to obtain the text feature mining information.

Optionally, the method further includes:

Store multiple electronic medical record information samples;

reading the plurality of electronic medical record information samples and filtering the plurality of electronic medical record information samples to obtain a plurality of filtered electronic medical record information samples, including a plurality of structured data samples and text data samples;

The multiple structured data samples and text data samples are divided into training sets, verification sets and test sets, and the preset model is trained for multiple rounds until the evaluation score of the trained preset model is equal to the preset number of rounds. If it stops rising, the training is stopped, and the model corresponding to the highest evaluation score is determined as the emergency condition classification prediction model.

In a third aspect, an embodiment of the present application further provides an electronic device, including: a memory, a processor, and a computer program stored on the memory and executable on the processor, the computer program being processed by the processor When the device is executed, the steps of the method for predicting disease classification based on electronic medical records as described in the second aspect are implemented.

In a fourth aspect, an embodiment of the present application further provides a computer-readable storage medium, where a computer program is stored on the computer-readable storage medium, and when the computer program is executed by a processor, the computer program described in the second aspect is implemented. Steps of a disease classification prediction method based on electronic medical records.

The embodiments of the present invention include the following advantages:

In the embodiment of the present invention, the electronic medical record information is stored by the first storage module, and the electronic medical record information is read and filtered by the first information filter to obtain structured data and text data. The structured data and text data are processed by the hierarchical prediction model, and the text feature mining information, the disease classification prediction results and the visual analysis text are obtained. By using the classification prediction system of the present application, the electronic medical record information can be processed relatively quickly and the disease classification prediction result can be obtained. The classification prediction model used has a faster classification speed and higher accuracy, and the obtained disease classification prediction result , Text feature mining information and visual analysis text can be used as a reference for doctors to assist doctors in grading to improve grading speed and grading accuracy, so that limited emergency resources can be fully utilized.

Description of drawings

Fig. 1 is a structural block diagram of a disease classification prediction system based on electronic medical records of the present invention;

Fig. 2 is the structural block diagram of a kind of hierarchical predictor of the present invention;

Fig. 3 is the structural block diagram of another kind of disease grading prediction system based on electronic medical record of the present invention;

FIG. 4 is a flow chart of steps of a method for predicting disease classification based on electronic medical records of the present invention.

Detailed ways

In order to make the above objects, features and advantages of the present invention more clearly understood, the present invention will be described in further detail below with reference to the accompanying drawings and specific embodiments.

Please refer to FIG. 1. FIG. 1 shows a structural block diagram of an electronic medical record-based disease classification prediction system according to an embodiment of the present application. As shown in FIG. 1, the system includes the following structures:

The first storage module 101 is configured to store electronic medical record information, where the electronic medical record information includes objective indicators, current illness history and physical examination results.

In this embodiment, before the electronic medical record information is stored in the first storage module, the medical staff has obtained the electronic medical record information of the patient through various medical means, including objective indicators, current illness history and physical examination results. Specifically, the objective indicators can be Including respiratory rate, systolic blood pressure, diastolic blood pressure, pulse, body temperature, blood oxygen saturation, age and gender, history of present illness may include trauma, loss of consciousness, dysphagia, vomiting, limb swelling and pain, epilepsy, etc. Physical examination may include mental status , reflex function, limb tension, breath sounds, heart rhythm, tenderness, rebound tenderness, etc. The medical staff input objective indicators, current illness history and physical examination results into this electronic medical record-based disease classification prediction, and the first storage module stores them.

A first information filter 102, connected to the first storage module, for reading the electronic medical record information from the first storage module and filtering the electronic medical record information to obtain filtered electronic medical record information , including structured data and text data.

In this embodiment, the first information filter is connected to the first storage module, and then reads the electronic medical record information from the first storage module, and filters the electronic medical record information, specifically, deletes missing data and erroneous record data , for example, to delete the following data:

(1) Temperature<30℃, (2)SBP>400mmHg, (3)DBP<5mmHg.

Further, the filtered electronic medical record information includes structured data and text data, wherein the structured data includes objective indicators, and the text data includes current illness history and physical examination results, so as to improve the classification accuracy.

The classification predictor 103 is configured to input the structured data and the text data into the emergency disease classification prediction model respectively, and obtain text feature mining information, disease classification prediction results and visual analysis text.

In this embodiment, the classification predictor inputs structured data and text data into the trained emergency disease classification prediction model, processes the structured data and text data respectively through the emergency disease classification prediction model, and uses LSTM (Long Short Term Memory Network) to process the structured data and text data respectively. ) combined with the attention mechanism to obtain text feature mining information, disease grading prediction results, and visual analysis text, which can be used as a reference for doctors' grading, thereby improving the grading rate and grading accuracy.

In the embodiment of the present invention, the electronic medical record information is stored by the first storage module, and the electronic medical record information is read and filtered by the first information filter to obtain structured data and text data. The structured data and text data are processed by the hierarchical prediction model, and the text feature mining information, the disease classification prediction results and the visual analysis text are obtained. By using the classification prediction system of the present application, the electronic medical record information can be processed relatively quickly and the disease classification prediction result can be obtained. The classification prediction model used has a faster classification speed and higher accuracy, and the obtained disease classification prediction result , Text feature mining information and visual analysis text can be used as a reference for doctors to assist doctors in grading to improve grading speed and grading accuracy, so that limited emergency resources can be fully utilized.

Please refer to FIG. 2. FIG. 2 shows a structural block diagram of a classification predictor according to an embodiment of the present application. As shown in FIG. 2, in a feasible implementation, the emergency disease classification prediction model includes a representation model and a Fusion model, the hierarchical predictor includes the following 201-203 three modules:

The first representation module 201 is configured to input the text data into the representation model to obtain the visual analysis text, the text feature mining information and the text feature vector.

In this embodiment, the first representation module can input the text data into the representation model and process it to obtain the visual analysis text, text feature mining information and text feature vector. Among them, the visual analysis text and text feature mining information are directly output, and the text feature vector is used for subsequent processing.

In a feasible implementation manner, the first presentation module includes:

A mapping submodule for inputting the text data into the representation model, mapping each Chinese character of the text data to a vector, and extracting the context information embedded in the text data to obtain an embedded vector;

An attention submodule, for performing attention processing on the embedded vector to obtain the visual analysis text and text feature vector;

The extraction submodule is used to extract phrases from the visual analysis text to obtain the text feature mining information.

In this embodiment, the mapping sub-module inputs the text data into the representation model, maps each Chinese character of the text data to a vector through the Embedding layer, and then extracts the context information embedded in the text data through the BiLSTM layer to obtain the embedding vector, and The embedded vector is subjected to attention processing by the attention sub-module, and the visual analysis text and text feature vector are obtained. The visual analysis text is further processed by the extraction sub-module, and the text features at the phrase level are extracted, and the text feature mining information is obtained, wherein , the visual analysis text and text feature mining information can be directly output, and the text feature vector is used for subsequent further processing. Extracting text features at the phrase level is more meaningful and can improve the classification accuracy. For example, if "no fever" is divided into "no fever" and "fever", "no fever" cannot be correctly mined. Visual analysis text and text feature mining information can satisfy clinical interpretability. In order to facilitate doctors to quickly obtain key information, in specific implementations, the key information in the visual analysis text (ie, text feature mining information) can be displayed in different colors. , use darker colors for more important information.

The second representation module 202 is configured to input the structured data into the representation model to obtain a structured feature vector.

In this embodiment, the second representation module may input the structured data into the representation model, and perform feature extraction on the structured data through two fully connected layers to obtain a structured feature vector.

In the above manner, different structural layers are used to process structured data and text data respectively, so that more accurate data can be obtained.

The fusion module 203 is configured to input the structured feature vector and the text feature vector into the fusion model to obtain the disease classification prediction result.

In this embodiment, the fusion module can fuse the structured feature vector and the text feature vector for processing, so as to obtain the disease classification prediction result. After the disease classification prediction result is output, it can be used as the doctor classification together with the text feature mining information and the visual analysis text. Reference, thereby improving the doctor's grading rate and grading accuracy.

Please refer to FIG. 3 . FIG. 3 shows a structural block diagram of another electronic medical record-based disease classification prediction system according to an embodiment of the present application. As shown in FIG. 3 , in a feasible implementation manner, the system further includes: :

The second storage module 301 is configured to store a plurality of electronic medical record information samples.

Before classifying the electronic medical record information, it is also necessary to train to obtain the emergency disease classification prediction model. Therefore, the second storage module needs to store multiple electronic medical record information samples. It is used as a training sample, and each electronic medical record information sample is manually marked with a classification result, and then input into the second storage module.

The second information filter 302, connected to the second storage module, is configured to read the plurality of electronic medical record information samples from the second storage module and filter the plurality of electronic medical record information samples to obtain Multiple filtered medical record information samples, including multiple structured data samples and text data samples;

The filtering operation of the second information filter is the same as the filtering operation of the first information filter, and the second information classifier is similar to the first information classifier. The operation explanation will not be repeated here.

The model trainer 303 is used to divide the multiple structured data samples and text data samples into a training set, a verification set and a test set, and perform multiple rounds of training on the preset model until the evaluation score of the preset model after training If there is no increase in the preset number of rounds, the training is stopped, and the model corresponding to the highest evaluation score is determined as the emergency disease classification prediction model.

In this embodiment, multiple structured data samples and text data samples are divided into training sets, validation sets and test sets. For example, multiple structured data samples and text data samples can be divided into Set 10% of the test set and 20% of the test set, perform multiple rounds of training on the preset model, specifically, use the training set to train the preset model, use the validation set to verify the trained preset model, and calculate the error value, through The back-propagation algorithm updates the preset model, and uses the test set to evaluate the trained preset model, and repeats the above steps to perform multiple rounds of training operations on the updated preset model until the evaluation of the trained preset model If the score no longer increases within the preset number of rounds, the training is stopped, and the model corresponding to the highest evaluation score is determined as the emergency disease classification prediction model. The model trainer is connected with the classification predictor, and the trained emergency disease classification prediction model can be input into the classification predictor.

In this embodiment, the emergency disease classification prediction model trained in the above manner can quickly process structured data and text data, and the obtained classification result has a high accuracy rate.

Based on the same inventive concept, an embodiment of the present application provides a method for predicting disease classification based on electronic medical records. Referring to FIG. 4 , FIG. 4 is a flowchart of steps of a method for predicting disease classification based on electronic medical records according to an embodiment of the application, as shown in FIG. 4, the method includes:

Step S401: Store electronic medical record information, where the electronic medical record information includes objective indicators, current illness history and physical examination results;

Step S402: reading the electronic medical record information and filtering the electronic medical record information to obtain filtered electronic medical record information, including structured data and text data;

Step S403: Input the structured data and the text data into the emergency disease classification prediction model respectively, and obtain text feature mining information, disease classification prediction results and visual analysis text.

Optionally, the emergency disease classification prediction model includes a representation model and a fusion model, and the structured data and the text data are respectively input into the emergency disease classification prediction model to obtain text feature mining information, disease classification prediction results and visual analysis. text, including:

Inputting the text data into the representation model to obtain the visual analysis text, the text feature mining information and the text feature vector;

Inputting the structured data into the representation model to obtain a structured feature vector;

Inputting the structured feature vector and the text feature vector into the fusion model to obtain the disease classification prediction result.

Optionally, inputting the text data into a representation model to obtain the visual analysis text, the text feature mining information and the text feature vector, including:

Inputting the text data into the representation model, and mapping each Chinese character of the text data to a vector by the representation model to obtain an embedded vector;

Perform attention processing on the embedded vector to obtain the visual analysis text and text feature vector;

Phrase extraction is performed on the visual analysis text to obtain the text feature mining information.

Optionally, the method further includes:

Store multiple electronic medical record information samples;

reading the plurality of electronic medical record information samples and filtering the plurality of electronic medical record information samples to obtain a plurality of filtered electronic medical record information samples;

Classifying the plurality of filtered electronic medical record information to obtain a plurality of structured data samples and text data samples;

The multiple structured data samples and text data samples are divided into training sets, verification sets and test sets, and the preset model is trained for multiple rounds until the evaluation score of the trained preset model is equal to the preset number of rounds. If it stops rising, the training is stopped, and the model corresponding to the highest evaluation score is determined as the emergency condition classification prediction model.

Based on the same inventive concept, another embodiment of the present application provides an electronic device, including: a memory, a processor, and a computer program stored on the memory and executable on the processor, the computer program being When executed by the processor, the steps in the method described in any of the foregoing embodiments are implemented.

Based on the same inventive concept, another embodiment of the present application provides a computer-readable storage medium, where a computer program is stored on the computer-readable storage medium, and when the computer program is executed by a processor, any one of the above-mentioned applications in the present application is implemented. Steps in the methods described in the Examples.

As for the method embodiment, since it is basically similar to the system embodiment, the description is relatively simple, and reference may be made to the partial description of the system embodiment for related parts.

The various embodiments in this specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments, and the same and similar parts between the various embodiments may be referred to each other.

It should be understood by those skilled in the art that the embodiments of the embodiments of the present invention may be provided as a method, an apparatus, or a computer program product. Accordingly, embodiments of the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, embodiments of the present invention may take the form of a computer program product implemented on one or more computer-usable storage media having computer-usable program code embodied therein, including but not limited to disk storage, CD-ROM, optical storage, and the like.

Embodiments of the present invention are described with reference to flowcharts and/or block diagrams of methods, terminal devices (systems), and computer program products according to embodiments of the present invention. It will be understood that each flow and/or block in the flowchart illustrations and/or block diagrams, and combinations of flows and/or blocks in the flowchart illustrations and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to the processor of a general purpose computer, special purpose computer, embedded processor or other programmable data processing terminal equipment to produce a machine that causes the instructions to be executed by the processor of the computer or other programmable data processing terminal equipment Means are created for implementing the functions specified in the flow or flows of the flowcharts and/or the blocks or blocks of the block diagrams.

These computer program instructions may also be stored in a computer readable memory capable of directing a computer or other programmable data processing terminal equipment to operate in a particular manner, such that the instructions stored in the computer readable memory result in an article of manufacture comprising instruction means, the The instruction means implement the functions specified in the flow or flow of the flowcharts and/or the block or blocks of the block diagrams.

These computer program instructions can also be loaded on a computer or other programmable data processing terminal equipment, so that a series of operational steps are performed on the computer or other programmable terminal equipment to produce a computer-implemented process, thereby executing on the computer or other programmable terminal equipment The instructions executed on the above provide steps for implementing the functions specified in the flowchart or blocks and/or the block or blocks of the block diagrams.

Although preferred embodiments of the embodiments of the present invention have been described, additional changes and modifications to these embodiments may be made by those skilled in the art once the basic inventive concepts are known. Therefore, the appended claims are intended to be construed to include the preferred embodiments as well as all changes and modifications that fall within the scope of the embodiments of the present invention.

Finally, it should also be noted that in this document, relational terms such as first and second are used only to distinguish one entity or operation from another, and do not necessarily require or imply these entities or that there is any such actual relationship or sequence between operations. Moreover, the terms "comprising", "comprising" or any other variation thereof are intended to encompass non-exclusive inclusion, such that a process, method, article or terminal device comprising a list of elements includes not only those elements, but also a non-exclusive list of elements. other elements, or also include elements inherent to such a process, method, article or terminal equipment. Without further limitation, an element defined by the phrase "comprises a..." does not preclude the presence of additional identical elements in the process, method, article or terminal device comprising said element.

A disease classification prediction system based on electronic medical records, a disease classification prediction method based on electronic medical records, an electronic device and a computer-readable storage medium provided by the present invention have been described in detail above. The principles and implementations of the present invention are described with a single example, and the descriptions of the above embodiments are only used to help understand the method and the core idea of the present invention; There will be changes in the implementation manner and the application scope. To sum up, the contents of this specification should not be construed as limiting the present invention.

Claims (10)

1. A system for staging and predicting a medical condition based on an electronic medical record, the system comprising:

the first storage module is used for storing electronic medical record information, and the electronic medical record information comprises objective indexes, current medical history and physical examination results;

the first information filter is connected with the first storage module and used for reading the electronic medical record information from the first storage module and filtering the electronic medical record information to obtain the filtered electronic medical record information which comprises structured data and text data;

and the hierarchical predictor is used for respectively inputting the structured data and the text data into an emergency disease hierarchical prediction model to obtain text characteristic mining information, a disease hierarchical prediction result and a visual analysis text.

2. The system of claim 1, wherein the emergency condition staging prediction model comprises a representation model and a fusion model, and wherein the staging predictor comprises:

the first representation module is used for inputting the text data into the representation model to obtain the visual analysis text, the text feature mining information and a text feature vector;

the second representation module is used for inputting the structured data into the representation model to obtain a structured feature vector;

and the fusion module is used for inputting the structural feature vector and the text feature vector into the fusion model to obtain the disease condition grading prediction result.

3. The system of claim 2, wherein the first representation module comprises:

the mapping submodule is used for inputting the text data into the representation model, mapping each Chinese character of the text data to a vector, and extracting context information embedded in the text data to obtain an embedded vector;

the attention submodule is used for carrying out attention processing on the embedded vector to obtain the visual analysis text and the text feature vector;

and the extraction submodule is used for carrying out phrase extraction on the visual analysis text to obtain the text feature mining information.

4. The system of claim 1, further comprising:

the second storage module is used for storing a plurality of electronic medical record information samples;

the second information filter is connected with the second storage module and used for reading the plurality of electronic medical record information samples from the second storage module and filtering the plurality of electronic medical record information samples to obtain a plurality of filtered medical record information samples which comprise a plurality of structured data samples and text data samples;

and the model trainer is used for dividing the multiple structured data samples and the text data samples into a training set, a verification set and a test set, performing multiple rounds of training on the preset model until the evaluation score of the trained preset model does not rise within the preset number of rounds, stopping training, and determining the model corresponding to the highest evaluation score as the emergency disease condition graded prediction model.

5. A disease grading prediction method based on an electronic medical record is characterized by comprising the following steps:

storing electronic medical record information, wherein the electronic medical record information comprises objective indexes, current medical history and physical examination results;

reading the electronic medical record information and filtering the electronic medical record information to obtain filtered electronic medical record information, wherein the filtered electronic medical record information comprises

Structured data and text data;

and respectively inputting the structured data and the text data into an emergency disease condition grading prediction model to obtain text characteristic mining information, a disease condition grading prediction result and a visual analysis text.

6. The method of claim 5, wherein the emergency disease grading prediction model comprises a representation model and a fusion model, and the structured data and the text data are respectively input into the emergency disease grading prediction model to obtain text feature mining information, a disease grading prediction result and a visual analysis text, and the method comprises the following steps:

inputting the text data into the representation model to obtain the visual analysis text, the text feature mining information and a text feature vector;

inputting the structured data into the representation model to obtain a structured feature vector;

and inputting the structural feature vector and the text feature vector into the fusion model to obtain the disease condition grading prediction result.

7. The method of claim 6, wherein inputting the text data into a representation model, resulting in the visual analysis text, the text feature mining information, and a text feature vector, comprises:

inputting the text data into the representation model, mapping each Chinese character of the text data to a vector, and extracting context information embedded in the text data to obtain an embedded vector;

performing attention processing on the embedded vector to obtain the visual analysis text and a text feature vector;

and performing phrase extraction on the visual analysis text to obtain text feature mining information.

8. The method of claim 5, further comprising:

storing a plurality of electronic medical record information samples;

reading the plurality of electronic medical record information samples and filtering the plurality of electronic medical record information samples to obtain a plurality of filtered electronic medical record information samples, wherein the plurality of filtered electronic medical record information samples comprise a plurality of structured data samples and text data samples;

and dividing the multiple structured data samples and the text data samples into a training set, a verification set and a test set, performing multiple rounds of training on the preset model until the evaluation score of the trained preset model does not rise within the preset rounds, stopping training, and determining the model corresponding to the highest evaluation score as the emergency disease condition grading prediction model.

9. An electronic device, comprising: a memory, a processor and a computer program stored on the memory and executable on the processor, the computer program when executed by the processor implementing the steps of the method for predicting a medical condition grade based on an electronic medical record according to any one of claims 5 to 8.

10. A computer-readable storage medium, having a computer program stored thereon, which, when being executed by a processor, implements the steps of the method for medical condition stratification prediction based on electronic medical record according to any of claims 5 to 8.

Images