An electronic medical record analysis method and system based on natural language processing

Abstract

The present invention discloses an electronic medical record analysis method and system based on natural language processing, the method comprising: obtaining a user's electronic medical record, dividing the electronic medical record into multiple information blocks, each of the information blocks corresponding to a certain element in the electronic medical record, and extracting an embedded vector representation of each information block through a natural language model; setting a text mapping model for the electronic medical record, and obtaining a global representation of the electronic medical record based on the embedded vector representation of the information block; setting a temporal dynamic acquisition model for the electronic medical record, and obtaining an embedded vector representation of each time point of the information block based on the embedded vector representation of the information block; setting a disease and symptom prediction model, and predicting the user's disease type based on the global representation of the electronic medical record and the embedded vector representation of each time point; recommending drugs to the user based on the prediction results, and setting an efficacy evaluation model for the recommended drugs to evaluate the efficacy of the user after taking the drugs.

Description

Technical Field

[0001] The present invention belongs to the technical field of electronic medical record analysis, and more specifically, relates to an electronic medical record analysis method and system based on natural language processing. Background Art

[0002] An electronic medical record (EMR) is a system that digitally records, stores, and manages patient health information. Compared to traditional paper-based medical records, EMRs allow for more convenient storage, updating, querying, and sharing of patient medical information, improving the efficiency and quality of healthcare services.

[0003] Electronic medical records typically include the following:

[0004] 1. Basic information: patient's name, age, gender, contact information and other personal information.

[0005] 2. Medical history records: including the patient's family history, personal medical history, allergy history, etc.

[0006] 3. Diagnostic information: the doctor’s records of the patient’s disease diagnosis, clinical symptoms, physical signs, etc.

[0007] 4. Treatment records: including medication records, surgical records, treatment plans, etc.

[0008] 5. Examination and test results: such as blood tests, imaging tests, pathological examinations, etc.

[0009] 6. Doctor’s orders: Doctor’s suggestions on the patient’s treatment plan, medication, nursing care, rehabilitation, etc.

[0010] 7. Follow-up and re-examination records: follow-up tracking of the patient's illness, recovery status, re-examination recommendations, etc.

[0011] 8. Electronic medical records not only improve medical efficiency but also enable information sharing and facilitate collaboration across different medical institutions. By integrating data from different departments (such as laboratories and imaging departments), electronic medical records provide doctors with comprehensive information support, helping to improve the accuracy of diagnosis and treatment decisions.

[0012] However, the current existing technologies are insufficient in their ability to analyze and process information in electronic medical records, and lack the ability to mine electronic medical record information. Summary of the Invention

[0013] To solve the above technical problems, the present invention proposes an electronic medical record analysis method based on natural language processing, comprising:

[0014] Obtaining a user's electronic medical record, dividing the electronic medical record into multiple information blocks, each of which corresponds to an element in the electronic medical record, and extracting an embedding vector representation of each information block using a natural language model, where the element is: user information, symptoms, medication name, and recovery time;

[0015] Setting a text mapping model for the electronic medical record, and obtaining a global representation of the electronic medical record based on the embedded vector representation of the information block;

[0016] Setting a temporal dynamic acquisition model for electronic medical records, and acquiring an embedded vector representation of each time point of the information block according to the embedded vector representation of the information block;

[0017] Setting a disease and symptom prediction model, and predicting the user's disease type based on the global representation of the electronic medical record and the embedded vector representation of each time point;

[0018] Based on the prediction results, drugs are recommended to users, and an efficacy evaluation model for the recommended drugs is set up to evaluate the efficacy of the drugs after the users take them.

[0019] Furthermore, the text mapping model of electronic medical records includes:

[0020] ,

[0021] in, Electronic Medical Records The global representation of is the number of information blocks, For the The adjustment factor for each block of information, is the activation function, is the weight matrix of the text mapping model, For the The embedding vector representation of each information block is: is the bias term of the text mapping model.

[0022] Furthermore, the temporal dynamic acquisition model of electronic medical records includes:

[0023] ,

[0024] in, For the The embedding vector representation of each time point, Get the weight matrix of the model dynamically for time series, For the The embedding vector representation of each time point is Electronic Medical Records No. Time series data at a time point, Dynamically obtain the model bias for time series.

[0025] Furthermore, the disease and symptom prediction model includes:

[0026] ,

[0027] in, Based on electronic medical records Determine if the user has a disease The probability of is the first weight matrix of the disease and symptom prediction model, is the second weight matrix of the disease and symptom prediction model, For the The embedding vector representation of each time point is is the bias term of the disease and symptom prediction model.

[0028] Furthermore, the efficacy evaluation model of the recommended drug includes:

[0029] ,

[0030] in, Based on electronic medical records The efficacy evaluation value of the recommended drug, For the types of recommended drugs, For the Adjustment factors for recommended drugs of this class, is the weight matrix of the efficacy evaluation model, For the Characteristic representation of recommended drugs of the class, is the bias term of the efficacy evaluation model.

[0031] Furthermore, a model for evaluating the efficacy of recommended drugs is set up to evaluate the efficacy of the drugs after the user takes the drugs, including comparing the efficacy evaluation value with a preset evaluation threshold. When the preset evaluation threshold is exceeded, the recommended drug is effective; otherwise, the drug is recommended again.

[0032] The present invention also proposes an electronic medical record analysis system based on natural language processing, comprising:

[0033] An information acquisition module is configured to obtain a user's electronic medical record, divide the electronic medical record into multiple information blocks, each of which corresponds to an element in the electronic medical record, and extract an embedded vector representation of each information block using a natural language model, wherein the element is: user information, symptoms, medication name, and recovery time;

[0034] a text mapping module, configured to set a text mapping model for the electronic medical record and obtain a global representation of the electronic medical record based on the embedded vector representation of the information block;

[0035] A time series dynamic acquisition module is used to set a time series dynamic acquisition model for electronic medical records, and obtain an embedded vector representation of each time point of the information block according to the embedded vector representation of the information block;

[0036] a symptom prediction module, configured to set up a disease and symptom prediction model and predict the user's disease type based on the global representation of the electronic medical record and the embedded vector representation of each time point;

[0037] The efficacy evaluation module is used to recommend drugs to users based on the prediction results, set up an efficacy evaluation model for the recommended drugs, and evaluate the efficacy of the drugs after the users take them.

[0038] Furthermore, the text mapping model of electronic medical records includes:

[0039] ,

[0040] in, Electronic Medical Records The global representation of is the number of information blocks, For the The adjustment factor for each block of information, is the activation function, is the weight matrix of the text mapping model, For the The embedding vector representation of each information block is: is the bias term of the text mapping model.

[0041] Furthermore, the temporal dynamic acquisition model of electronic medical records includes:

[0042] ,

[0043] in, For the The embedding vector representation of each time point, Get the weight matrix of the model dynamically for time series, For the The embedding vector representation of each time point is Electronic Medical Records No. Time series data at a time point, Dynamically obtain the model bias for time series.

[0044] Furthermore, the disease and symptom prediction model includes:

[0045] ,

[0046] in, Based on electronic medical records Determine if the user has a disease The probability of is the first weight matrix of the disease and symptom prediction model, is the second weight matrix of the disease and symptom prediction model, For the The embedding vector representation of each time point is is the bias term of the disease and symptom prediction model.

[0047] In general, the above technical solutions conceived by the present invention have the following beneficial effects compared with the prior art:

[0048] Through the above technical solutions, the present invention can extract key information from the user's electronic medical record, analyze the key information, and thus predict the user's possible diseases, recommend drugs to the user, and judge the effects of the recommended drugs, so as to make adjustments. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] Figure 1 is a flow chart of the method of embodiment 1 of the present invention;

[0050] Figure 2 This is a system structure diagram of Example 2 of the present invention. DETAILED DESCRIPTION

[0051] In order to better understand the above technical solution, the above technical solution will be described in detail below with reference to the accompanying drawings and specific implementation methods.

[0052] The method provided by the present invention can be implemented in the following terminal environment, wherein the terminal may include one or more of the following components: a processor, a storage medium, and a display screen. The storage medium stores at least one instruction, which is loaded and executed by the processor to implement the method described in the following embodiments.

[0053] A processor can include one or more processing cores. It connects various components within the terminal using various interfaces and circuits. It executes instructions, programs, code sets, or instruction sets stored in storage media, and accesses data stored in storage media to perform various terminal functions and process data.

[0054] The storage medium may include a random access memory (RAM) or a read-only memory (ROM). The storage medium may be used to store instructions, programs, codes, code sets, or instructions.

[0055] The display is used to show the user interface of each application.

[0056] In addition, those skilled in the art will appreciate that the structure of the terminal described above does not limit the terminal. The terminal may include more or fewer components, or a combination of certain components, or a different arrangement of components. For example, the terminal may also include a radio frequency circuit, an input unit, a sensor, an audio circuit, a power supply, and other components, which will not be described in detail here.

[0057] Example 1

[0058] like Figure 1 As shown, an embodiment of the present invention provides an electronic medical record analysis method based on natural language processing, including:

[0059] Step 101: Obtain the user's electronic medical record and divide the electronic medical record into multiple information blocks. Each information block corresponds to an element in the electronic medical record. Then, use a natural language model (such as BERT or ChatGPT) to extract an embedding vector representation of each information block. The element may be: user information, symptoms, medication name, or recovery time.

[0060] Step 102: Setting a text mapping model for the electronic medical record, and obtaining a global representation of the electronic medical record based on the embedded vector representation of the information block;

[0061] Specifically, the text mapping model of electronic medical records includes:

[0062] ,

[0063] in, Electronic Medical Records The global representation of is the number of information blocks, For the The adjustment factor for each block of information, is the activation function (such as Sigmoid or tanh), is the weight matrix of the text mapping model, For the The embedding vector representation of each information block is: is the bias term of the text mapping model.

[0064] Step 103: Setting a temporal dynamic acquisition model for electronic medical records, and obtaining an embedded vector representation of each time point of the information block (e.g., the onset and change of symptoms, the use of medication, etc.) based on the embedded vector representation of the information block.

[0065] Specifically, the temporal dynamic acquisition model of electronic medical records includes:

[0066] ,

[0067] in, For the The embedding vector representation of each time point, Get the weight matrix of the model dynamically for time series, For the The embedding vector representation of each time point is Electronic Medical Records No. Time series data at a time point, Dynamically obtain the model bias for time series.

[0068] Step 104: Setting a disease and symptom prediction model and predicting the user's disease type based on the global representation of the electronic medical record and the embedded vector representation of each time point;

[0069] Specifically, the disease and symptom prediction models include:

[0070] ,

[0071] in, Based on electronic medical records Determine if the user has a disease The probability of is the first weight matrix of the disease and symptom prediction model, is the second weight matrix of the disease and symptom prediction model, For the The embedding vector representation of each time point is is the bias term of the disease and symptom prediction model.

[0072] Step 105 , based on the prediction results, recommend drugs to the user, and set up an efficacy evaluation model for the recommended drugs to evaluate the efficacy of the drugs after the user takes them.

[0073] Specifically, the recommended drug efficacy evaluation models include:

[0074] ,

[0075] in, Based on electronic medical records The efficacy evaluation value of the recommended drug, For the types of recommended drugs, For the Adjustment factors for recommended drugs of this class, is the weight matrix of the efficacy evaluation model, For the Feature representation of recommended drugs (can be constructed based on information such as the drug's chemical composition and dosage), is the bias term of the efficacy evaluation model.

[0076] Specifically, setting up an efficacy evaluation model for recommended drugs and evaluating the efficacy of the drugs after the user takes the drugs includes: comparing the efficacy evaluation value with a preset evaluation threshold; when the preset evaluation threshold is exceeded, the recommended drug is effective; otherwise, the drug is recommended again.

[0077] Example 2

[0078] like Figure 2 As shown, an embodiment of the present invention further provides an electronic medical record analysis system based on natural language processing, comprising:

[0079] An information acquisition module is used to obtain the user's electronic medical record, divide the electronic medical record into multiple information blocks, each of which corresponds to an element in the electronic medical record, and extract an embedded vector representation of each information block using a natural language model (such as BERT, ChatGPT, etc.), where the element is: user information, symptoms, medication name, and recovery time;

[0080] a text mapping module, configured to set a text mapping model for the electronic medical record and obtain a global representation of the electronic medical record based on the embedded vector representation of the information block;

[0081] Specifically, the text mapping model of electronic medical records includes:

[0082] ,

[0083] in, Electronic Medical Records The global representation of is the number of information blocks, For the The adjustment factor for each block of information, is the activation function, is the weight matrix of the text mapping model, For the The embedding vector representation of each information block is: is the bias term of the text mapping model.

[0084] A time series dynamic acquisition module is used to set a time series dynamic acquisition model for electronic medical records, and obtain an embedded vector representation of each time point of the information block according to the embedded vector representation of the information block;

[0085] Specifically, the temporal dynamic acquisition model of electronic medical records includes:

[0086] ,

[0087] in, For the The embedding vector representation of each time point, Get the weight matrix of the model dynamically for time series, For the The embedding vector representation of each time point is Electronic Medical Records No. Time series data at a time point, Dynamically obtain the model bias for time series.

[0088] a symptom prediction module, configured to set up a disease and symptom prediction model and predict the user's disease type based on the global representation of the electronic medical record and the embedded vector representation of each time point;

[0089] Specifically, the disease and symptom prediction models include:

[0090] ,

[0091] in, Based on electronic medical records Determine if the user has a disease The probability of is the first weight matrix of the disease and symptom prediction model, is the second weight matrix of the disease and symptom prediction model, For the The embedding vector representation of each time point is is the bias term of the disease and symptom prediction model.

[0092] The efficacy evaluation module is used to recommend drugs to users based on the prediction results, set up an efficacy evaluation model for the recommended drugs, and evaluate the efficacy of the drugs after the users take them.

[0093] Specifically, the recommended drug efficacy evaluation models include:

[0094] ,

[0095] in, Based on electronic medical records The efficacy evaluation value of the recommended drug, For the types of recommended drugs, For the Adjustment factors for recommended drugs of this class, is the weight matrix of the efficacy evaluation model, For the Characteristic representation of recommended drugs of the class, is the bias term of the efficacy evaluation model.

[0096] Specifically, setting up an efficacy evaluation model for recommended drugs and evaluating the efficacy of the drugs after the user takes the drugs includes: comparing the efficacy evaluation value with a preset evaluation threshold; when the preset evaluation threshold is exceeded, the recommended drug is effective; otherwise, the drug is recommended again.

[0097] Example 3

[0098] An embodiment of the present invention further provides a storage medium storing a plurality of instructions, wherein the instructions are used to implement the electronic medical record analysis method based on natural language processing.

[0099] Optionally, in this embodiment, the storage medium may be located in any computer terminal in a computer terminal group in a computer network, or in any mobile terminal in a mobile terminal group.

[0100] Optionally, in this embodiment, the storage medium is configured to store program codes for executing the following steps: Step 101, obtaining a user's electronic medical record, dividing the electronic medical record into multiple information blocks, each of the information blocks corresponding to an element in the electronic medical record, and extracting an embedding vector representation of each information block through a natural language model (such as BERT, ChatGPT, etc.), wherein the element is: user information, symptoms, medication name, and time taken for recovery;

[0101] Step 102: Setting a text mapping model for the electronic medical record, and obtaining a global representation of the electronic medical record based on the embedded vector representation of the information block;

[0102] Specifically, the text mapping model of electronic medical records includes:

[0103] ,

[0104] in, Electronic Medical Records The global representation of is the number of information blocks, For the The adjustment factor for each block of information, is the activation function, is the weight matrix of the text mapping model, For the The embedding vector representation of each information block is: is the bias term of the text mapping model.

[0105] Step 103: Setting a temporal dynamic acquisition model for electronic medical records, and acquiring an embedded vector representation of each time point of the information block according to the embedded vector representation of the information block;

[0106] Specifically, the temporal dynamic acquisition model of electronic medical records includes:

[0107] ,

[0108] in, For the The embedding vector representation of each time point, Get the weight matrix of the model dynamically for time series, For the The embedding vector representation of each time point is Electronic Medical Records No. Time series data at a time point, Dynamically obtain the model bias for time series.

[0109] Step 104: Setting a disease and symptom prediction model and predicting the user's disease type based on the global representation of the electronic medical record and the embedded vector representation of each time point;

[0110] Specifically, the disease and symptom prediction models include:

[0111] ,

[0112] in, Based on electronic medical records Determine if the user has a disease The probability of is the first weight matrix of the disease and symptom prediction model, is the second weight matrix of the disease and symptom prediction model, For the The embedding vector representation of each time point is is the bias term of the disease and symptom prediction model.

[0113] Step 105 , based on the prediction results, recommend drugs to the user, and set up an efficacy evaluation model for the recommended drugs to evaluate the efficacy of the drugs after the user takes them.

[0114] Specifically, the recommended drug efficacy evaluation models include:

[0115] ,

[0116] in, Based on electronic medical records The efficacy evaluation value of the recommended drug, For the types of recommended drugs, For the Adjustment factors for recommended drugs of this class, is the weight matrix of the efficacy evaluation model, For the Characteristic representation of recommended drugs of the class, is the bias term of the efficacy evaluation model.

[0117] Specifically, setting up an efficacy evaluation model for recommended drugs and evaluating the efficacy of the drugs after the user takes the drugs includes: comparing the efficacy evaluation value with a preset evaluation threshold; when the preset evaluation threshold is exceeded, the recommended drug is effective; otherwise, the drug is recommended again.

[0118] Example 4

[0119] An embodiment of the present invention also proposes an electronic device, including a processor and a storage medium connected to the processor, wherein the storage medium stores multiple instructions, which can be loaded and executed by the processor to enable the processor to execute the electronic medical record analysis method based on natural language processing.

[0120] Specifically, the electronic device of this embodiment may be a computer terminal, which may include: one or more processors, and a storage medium.

[0121] Among them, the storage medium can be used to store software programs and modules, such as the electronic medical record analysis method based on natural language processing in the embodiment of the present invention, and the corresponding program instructions / modules. The processor executes various functional applications and data processing by running the software programs and modules stored in the storage medium, that is, realizing the above-mentioned electronic medical record analysis method based on natural language processing. The storage medium may include high-speed random access storage media, and may also include non-volatile storage media, such as one or more magnetic storage systems, flash memory, or other non-volatile solid-state storage media. In some instances, the storage medium may further include storage media remotely located relative to the processor, and these remote storage media may be connected to the terminal via a network. Examples of the above-mentioned network include but are not limited to the Internet, corporate intranet, local area network, mobile communication network, and combinations thereof.

[0122] The processor can call the information and application stored in the storage medium through the transmission system to perform the following steps: Step 101: obtain the user's electronic medical record, divide the electronic medical record into multiple information blocks, each of the information blocks corresponds to an element in the electronic medical record, and extract the embedding vector representation of each information block through a natural language model (such as BERT, ChatGPT, etc.), where the element is: user information, symptoms, medication name, and recovery time;

[0123] Step 102: Setting a text mapping model for the electronic medical record, and obtaining a global representation of the electronic medical record based on the embedded vector representation of the information block;

[0124] Specifically, the text mapping model of electronic medical records includes:

[0125] ,

[0126] in, Electronic Medical Records The global representation of is the number of information blocks, For the The adjustment factor for each block of information, is the activation function, is the weight matrix of the text mapping model, For the The embedding vector representation of each information block is: is the bias term of the text mapping model.

[0127] Step 103: Setting a temporal dynamic acquisition model for electronic medical records, and acquiring an embedded vector representation of each time point of the information block according to the embedded vector representation of the information block;

[0128] Specifically, the temporal dynamic acquisition model of electronic medical records includes:

[0129] ,

[0130] in, For the The embedding vector representation of each time point, Get the weight matrix of the model dynamically for time series, For the The embedding vector representation of each time point is Electronic Medical Records No. Time series data at a time point, Dynamically obtain the model bias for time series.

[0131] Step 104: Setting a disease and symptom prediction model and predicting the user's disease type based on the global representation of the electronic medical record and the embedded vector representation of each time point;

[0132] Specifically, the disease and symptom prediction models include:

[0133] ,

[0134] in, Based on electronic medical records Determine if the user has a disease The probability of is the first weight matrix of the disease and symptom prediction model, is the second weight matrix of the disease and symptom prediction model, For the The embedding vector representation of each time point is is the bias term of the disease and symptom prediction model.

[0135] Step 105 , based on the prediction results, recommend drugs to the user, and set up an efficacy evaluation model for the recommended drugs to evaluate the efficacy of the drugs after the user takes them.

[0136] Specifically, the recommended drug efficacy evaluation models include:

[0137] ,

[0138] in, Based on electronic medical records The efficacy evaluation value of the recommended drug, For the types of recommended drugs, For the Adjustment factors for recommended drugs of this class, is the weight matrix of the efficacy evaluation model, For the Characteristic representation of recommended drugs of the class, is the bias term of the efficacy evaluation model.

[0139] Specifically, setting up an efficacy evaluation model for recommended drugs and evaluating the efficacy of the drugs after the user takes the drugs includes: comparing the efficacy evaluation value with a preset evaluation threshold; when the preset evaluation threshold is exceeded, the recommended drug is effective; otherwise, the drug is recommended again.

[0140] The serial numbers of the above embodiments of the present invention are for description only and do not represent the advantages or disadvantages of the embodiments.

[0141] In the above embodiments of the present invention, the description of each embodiment has its own focus. For parts that are not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0142] In the several embodiments provided by the present invention, it should be understood that the disclosed technical content can be implemented in other ways. Among them, the system embodiments described above are merely illustrative. For example, the division of the units is only a logical function division. In actual implementation, there may be other division methods. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of units or modules, and can be electrical or other forms.

[0143] The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of these units may be selected to achieve the purpose of this embodiment according to actual needs.

[0144] In addition, the functional units in the various embodiments of the present invention may be integrated into a single processing unit, each unit may exist physically separately, or two or more units may be integrated into a single unit. The aforementioned integrated units may be implemented in the form of hardware or software functional units.

[0145] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions for enabling a computer device (which can be a personal computer, server or network device, etc.) to execute all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, read-only storage medium (ROM, Read-Only Memory), random access storage medium (RAM, Random Access Memory), mobile hard disk, magnetic disk or optical disk, and other media that can store program code.

[0146] Obviously, the above embodiments are merely examples for clarity of explanation and are not intended to limit the implementation methods. Those skilled in the art will readily appreciate that other variations or modifications based on the above descriptions are possible. It is not necessary and impossible to enumerate all implementation methods here. Obvious variations or modifications arising therefrom remain within the scope of protection of the present invention.

Claims

1. A method for analyzing electronic medical records based on natural language processing, characterized in that: include: Obtain the user's electronic medical record, divide the electronic medical record into multiple information blocks, each of which corresponds to an element in the electronic medical record, and extract an embedded vector representation of each information block using a natural language model, where the element is: user information, symptoms, medication name, and recovery time. The natural language model is a BERT model or a ChatGPT model. Setting a text mapping model for the electronic medical record, and obtaining a global representation of the electronic medical record based on the embedded vector representation of the information block; Setting a temporal dynamic acquisition model for electronic medical records, and acquiring an embedded vector representation of each time point of the information block according to the embedded vector representation of the information block; The time series dynamic acquisition model includes: ， in, For the The embedding vector representation of each time point, Get the weight matrix of the model dynamically for time series, For the The embedding vector representation of each time point is Electronic Medical Records No. Time series data at a time point, Dynamically obtain the model bias for time series; Setting a disease and symptom prediction model, and predicting the user's disease type based on the global representation of the electronic medical record and the embedded vector representation of each time point; Based on the prediction results, drugs are recommended to users, and an efficacy evaluation model for the recommended drugs is set up to evaluate the efficacy of the drugs after the users take them.

2. The electronic medical record analysis method based on natural language processing according to claim 1, characterized in that: The text mapping model for electronic medical records includes: ， in, Electronic Medical Records The global representation of is the number of information blocks, For the The adjustment factor for each block of information, is the activation function, is the weight matrix of the text mapping model, For the The embedding vector representation of each information block is: is the bias term of the text mapping model.

3. The electronic medical record analysis method based on natural language processing according to claim 1, characterized in that: Disease and symptom prediction models include: ， in, Based on electronic medical records Determine if the user has a disease The probability of is the first weight matrix of the disease and symptom prediction model, is the second weight matrix of the disease and symptom prediction model, For the The embedding vector representation of each time point is is the bias term of the disease and symptom prediction model.

4. The electronic medical record analysis method based on natural language processing according to claim 3, characterized in that: The recommended drug efficacy evaluation models include: ， in, Based on electronic medical records The efficacy evaluation value of the recommended drug, For the types of recommended drugs, For the Adjustment factors for recommended drugs of this class, is the weight matrix of the efficacy evaluation model, For the Characteristic representation of recommended drugs of the class, is the bias term of the efficacy evaluation model.

5. The electronic medical record analysis method based on natural language processing according to claim 4, characterized in that: Setting up an efficacy evaluation model for recommended drugs and evaluating the efficacy of the drugs after the user takes the drugs includes: comparing the efficacy evaluation value with a preset evaluation threshold; when the preset evaluation threshold is exceeded, the recommended drug is effective; otherwise, the drug is recommended again.

6. An electronic medical record analysis system based on natural language processing, characterized in that: include: An information acquisition module is used to obtain the user's electronic medical record, divide the electronic medical record into multiple information blocks, each of which corresponds to an element in the electronic medical record, and extract an embedded vector representation of each information block through a natural language model, where the element is: user information, symptoms, medication name, and recovery time. The natural language model is a BERT model or a ChatGPT model; a text mapping module, configured to set a text mapping model for the electronic medical record and obtain a global representation of the electronic medical record based on the embedded vector representation of the information block; A time series dynamic acquisition module is used to set a time series dynamic acquisition model for electronic medical records, and obtain an embedded vector representation of each time point of the information block according to the embedded vector representation of the information block; The time series dynamic acquisition model includes: ， in, For the The embedding vector representation of each time point, Get the weight matrix of the model dynamically for time series, For the The embedding vector representation of each time point is Electronic Medical Records No. Time series data at a time point, Dynamically obtain the model bias for time series; a symptom prediction module, configured to set up a disease and symptom prediction model and predict the user's disease type based on the global representation of the electronic medical record and the embedded vector representation of each time point; The efficacy evaluation module is used to recommend drugs to users based on the prediction results, set up an efficacy evaluation model for the recommended drugs, and evaluate the efficacy of the drugs after the users take them.

7. The electronic medical record analysis system based on natural language processing according to claim 6, characterized in that: The text mapping model for electronic medical records includes: ， in, Electronic Medical Records The global representation of is the number of information blocks, For the The adjustment factor for each block of information, is the activation function, is the weight matrix of the text mapping model, For the The embedding vector representation of each information block is: is the bias term of the text mapping model.

8. The electronic medical record analysis system based on natural language processing according to claim 6, characterized in that: Disease and symptom prediction models include: ， in, Based on electronic medical records Determine if the user has a disease The probability of is the first weight matrix of the disease and symptom prediction model, is the second weight matrix of the disease and symptom prediction model, For the The embedding vector representation of each time point is is the bias term of the disease and symptom prediction model.

Images