Abnormal early warning method and device for outpatient service user data

Abstract

The invention discloses an outpatient service user data abnormity early warning method and device. The method comprises the following steps: acquiring a user outpatient service data set; preprocessing the user outpatient service data set to obtain user outpatient service standard data information; processing the outpatient service standard data information of the user by using a trained abnormity early warning model to obtain an abnormity early warning value; the abnormal early warning value is used for representing the occurrence probability of infectious diseases of the outpatient service user. Based on a large amount of outpatient service user medical data, early monitoring and prediction of infectious diseases and the like are achieved, and the supporting effect of the Internet and medical big data on the aspects of auxiliary epidemic situation research and judgment, early warning analysis, real-time monitoring and the like is fully played.

Description

Technical Field

[0001] The present invention relates to the fields of medical data processing, industrial big data and artificial intelligence, and in particular to an abnormal early warning method and device for outpatient user data. Background Art

[0002] With the widespread application of information technology, in today's medical system, electronic medical record systems have become an indispensable part of medical institutions at all levels, especially large and medium-sized hospitals. A large amount of medical data of outpatient users has been accumulated in the electronic medical record system, which specifically includes outpatient medical records, inpatient medical records, test reports, medication records and other types. These data include both structured parts, such as patients' vital signs such as body temperature and blood pressure, and a large amount of unstructured text data, such as detailed text information such as medical history, physical examination, diagnosis, and treatment written by doctors.

[0003] How to conduct early monitoring and prediction of infectious diseases based on a large amount of medical data from outpatient users, give full play to the supporting role of the Internet and medical big data in assisting epidemic research and judgment, early warning analysis, real-time monitoring, etc., achieve the goal of "early warning, early reporting, early diagnosis, and early treatment", prevent the import, spread, and export of epidemics, and control the spread of diseases, is an urgent problem to be solved. The current medical data processing methods have failed to conduct in-depth mining and utilization of the above-mentioned electronic medical record data. Summary of the invention

[0004] The present invention mainly solves the problem of how to conduct early monitoring and prediction of infectious diseases based on a large amount of outpatient user medical data, and give full play to the supporting role of the Internet and medical big data in assisting epidemic analysis, early warning analysis, real-time monitoring, etc. The present invention discloses an abnormal early warning method and device for outpatient user data.

[0005] In a first aspect of an embodiment of the present invention, a method for abnormal early warning of outpatient user data is disclosed, comprising:

[0006] S1, obtain the user's outpatient data set;

[0007] S2, preprocessing the user outpatient data set to obtain user outpatient standard data information;

[0008] S3, using the trained abnormal warning model, processing the user's outpatient standard data information to obtain an abnormal warning value; the abnormal warning value is used to characterize the probability of the outpatient user suffering from an infectious disease.

[0009] The preprocessing of the user outpatient data set to obtain user outpatient standard data information includes:

[0010] Performing text recognition processing on the unstructured data in the user outpatient data set to obtain corresponding text data;

[0011] Performing text vector conversion processing on the text data corresponding to the unstructured data to obtain a corresponding data vector;

[0012] Using the data vectors and structured data corresponding to the unstructured data in the user outpatient data set, construct user outpatient standard data information;

[0013] The user outpatient standard data information or user outpatient data information includes user data information, examination report data information, diagnosis report data information and medication record data information.

[0014] The abnormal warning model includes: a first processing network, a second processing network, and a third processing network;

[0015] The first processing network includes a first residual multi-head self-attention module, a second residual multi-head self-attention module, a third residual multi-head self-attention module, a fourth residual multi-head self-attention module and a multi-head mutual attention module;

[0016] The first residual multi-head self-attention module, the second residual multi-head self-attention module, the third residual multi-head self-attention module and the fourth residual multi-head self-attention module are connected to the multi-head mutual attention module, and are used to respectively receive the user data information, the examination report data information, the diagnosis report data information and the medication record data information in the input user outpatient standard data information, and perform self-attention feature extraction on the received data information to obtain corresponding feature information;

[0017] The multi-head mutual attention module is used to extract multi-head mutual attention features from the feature information corresponding to each piece of data information to obtain mutual attention features;

[0018] The first processing network is used to extract features from various types of input data information to obtain mutual attention features;

[0019] The output end of the first processing network is connected to the input end of the second processing network;

[0020] The second processing network is used to fuse the mutual attention features to obtain fused features; the output end of the second processing network is connected to the input end of the third processing network.

[0021] The second processing network includes a first input module, a first convolution module, a depth-separable convolution module, a first dimension-raising convolution module, a second dimension-raising convolution module, a third dimension-raising convolution module, a fourth dimension-raising convolution module, a second convolution module, a first pooling module, a third convolution module and a first fully connected module;

[0022] The input end of the first input module is used to receive the mutual attention feature; the output end of the first input module of the second processing network is connected to the input end of the first convolution module of the second processing network; the output end of the first convolution module of the second processing network is connected to the input end of the depthwise separable convolution module of the second processing network; the output end of the depthwise separable convolution module of the second processing network is connected to the input end of the first dimensionality-raising convolution module of the second processing network;

[0023] The output end of the first dimensionality-raising convolution module of the second processing network is connected to the input end of the second dimensionality-raising convolution module of the second processing network; the output end of the second dimensionality-raising convolution module of the second processing network is connected to the input end of the third dimensionality-raising convolution module of the second processing network; the output end of the third dimensionality-raising convolution module of the second processing network is connected to the input end of the fourth dimensionality-raising convolution module of the second processing network; the output end of the fourth dimensionality-raising convolution module of the second processing network is connected to the input end of the second convolution module of the second processing network; the output end of the second convolution module of the second processing network is connected to the input end of the first pooling module of the second processing network; the output end of the first pooling module of the second processing network is connected to the input end of the third convolution module of the second processing network; the output end of the third convolution module of the second processing network is respectively connected to the input ends of the fifth dimensionality-raising convolution module and the first fully connected module of the second processing network; the output end of the fifth dimensionality-raising convolution module is connected to the input end of the first fully connected module of the second processing network;

[0024] The output end of the first processing network is connected to the input end of the second processing network.

[0025] The third processing network includes: a second input module, a fourth convolution module, a fifth convolution module, a sixth convolution module, a second pooling module, a seventh convolution module, a second fully connected module and a third fully connected module;

[0026] The input end of the second input module of the third processing network is connected to the output end of the first fully connected module of the second processing network; the output end of the second input module of the third processing network is connected to the input end of the fourth convolution module of the third processing network; the output end of the fourth convolution module of the third processing network is connected to the input end of the fifth convolution module of the third processing network; the output end of the fifth convolution module of the third processing network is connected to the input end of the sixth convolution module of the third processing network; the output end of the sixth convolution module of the third processing network is connected to the input end of the second pooling module of the third processing network; the output end of the second pooling module of the third processing network is connected to the input end of the seventh convolution module of the third processing network; the output end of the seventh convolution module of the third processing network is connected to the input end of the second fully connected module of the third processing network; the output end of the second fully connected module of the third processing network is connected to the input end of the third fully connected module of the third processing network;

[0027] The third processing network is used to perform prediction processing on the fusion feature to obtain an abnormal warning value of the input user outpatient standard data information;

[0028] The output end of the third fully-connected module of the third processing network is used to output the abnormal warning value of the input user outpatient standard data information.

[0029] The training process of the abnormal warning model includes:

[0030] A user historical outpatient data set is obtained; the user historical outpatient data set includes user outpatient data of a number of users within a set historical time period and corresponding label information; the label information is the infectious disease occurrence value corresponding to the user outpatient data; the user outpatient data is training data;

[0031] Initialize the number of training iterations;

[0032] Input the training data in the user's historical outpatient data set as input data into the abnormal warning model;

[0033] Using the abnormal warning model, the input data is processed to obtain a predicted value;

[0034] Performing difference calculation processing on the obtained predicted value and the label information corresponding to the input data to obtain a difference value;

[0035] Determine whether the difference value satisfies a convergence condition, and obtain a first determination result;

[0036] When the first judgment result is no, judging whether the training iteration number value is equal to the training number threshold, and obtaining a second judgment result;

[0037] When the second judgment result is no, determining that the model training state does not meet the training termination condition;

[0038] When the second judgment result is yes, determining that the model training state satisfies the training termination condition;

[0039] When the first judgment result is yes, determining that the model training state satisfies a training termination condition;

[0040] When the model training status does not meet the termination training condition, the parameter update model is used to update the parameters of the abnormal warning model, and the training iteration number value is increased by 1, triggering the execution of inputting the training data in the user's historical outpatient data set as input data into the abnormal warning model;

[0041] When the model training state satisfies the training termination condition, the training process of the abnormal warning model is completed to obtain a trained abnormal warning model.

[0042] The parameter updating model is:

[0043]

[0044] θ←θ+v;

[0045] In the formula, is the difference value calculated for the i-th training data in the user's historical outpatient data set, v is the parameter update value, θ is the parameter of the abnormal warning model, η is the initial parameter learning rate, α is the momentum angle parameter, 0≤α≤1, ▽ θ It means to find the partial derivative with respect to the variable θ.

[0046] According to a second aspect of the present invention, an abnormal warning device for outpatient user data is disclosed, the device comprising:

[0047] A memory storing executable program code;

[0048] a processor coupled to the memory;

[0049] The processor calls the executable program code stored in the memory to execute the abnormal warning method for outpatient user data.

[0050] According to a third aspect of the present invention, a computer storable medium is disclosed, wherein the computer storable medium stores computer instructions, and when the computer instructions are called by a computer, they are used to execute the abnormal warning method for outpatient user data.

[0051] According to a fourth aspect of the present invention, an information data processing terminal is disclosed. The information data processing terminal is used to implement the abnormal warning method for outpatient user data.

[0052] The beneficial effects of the present invention are:

[0053] The present invention discloses an abnormal early warning method and device for outpatient user data, which mainly solves the problem of how to conduct early monitoring and prediction of infectious diseases based on a large amount of outpatient user medical data, and give full play to the supporting role of the Internet and medical big data in assisting epidemic assessment, early warning analysis, real-time monitoring, etc.

[0054] The present invention discloses an abnormal early warning model, which extracts features of multiple unrelated dimensions by performing feature extraction processing on user data information, examination report data information, diagnosis report data information and medication record data information in the user outpatient standard data information, and fuses the features by using a second processing network to obtain fused features, and finally obtains a prediction result by using a third processing module, thereby realizing rapid and accurate early warning of abnormal situations. In the training process of the model, a special parameter update model is established to realize rapid and accurate convergence of the model. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] Figure 1 It is a flow chart for implementing the method of the present invention. DETAILED DESCRIPTION

[0056] In order to better understand the content of the present invention, an embodiment is given here.

[0057] Figure 1 It is a flow chart for implementing the method of the present invention.

[0058] In a first aspect of an embodiment of the present invention, a method for abnormal early warning of outpatient user data is disclosed, comprising:

[0059] S1, obtain the user's outpatient data set;

[0060] S2, preprocessing the user outpatient data set to obtain user outpatient standard data information;

[0061] S3, using the trained abnormal warning model to process the user's outpatient standard data information to obtain an abnormal warning value; the abnormal warning value is used to characterize the probability of the outpatient user suffering from an infectious disease;

[0062] The preprocessing of the user outpatient data set to obtain user outpatient standard data information includes:

[0063] Performing text recognition processing on the unstructured data in the user outpatient data set to obtain corresponding text data;

[0064] Performing text vector conversion processing on the text data corresponding to the unstructured data to obtain a corresponding data vector;

[0065] Using the data vectors and structured data corresponding to the unstructured data in the user outpatient data set, construct user outpatient standard data information;

[0066] The user outpatient standard data information or user outpatient data information includes user data information, examination report data information, diagnosis report data information and medication record data information;

[0067] The user data information includes personal data information such as the user's age, gender, medical history, etc.;

[0068] The inspection report data information includes vital signs data and test data;

[0069] The diagnostic report data information includes outpatient medical record data, inpatient medical record data and data vectors corresponding to unstructured data;

[0070] The abnormal warning model includes: the first processing network, the second processing network, and the third processing network;

[0071] The first processing network includes a first residual multi-head self-attention module, a second residual multi-head self-attention module, a third residual multi-head self-attention module, a fourth residual multi-head self-attention module and a multi-head mutual attention module;

[0072] The first residual multi-head self-attention module, the second residual multi-head self-attention module, the third residual multi-head self-attention module, and the fourth residual multi-head self-attention module are connected to the multi-head mutual attention module, and are used to respectively receive user data information, examination report data information, diagnosis report data information, and medication record data information in the input user outpatient data information, and perform self-attention feature extraction on the received data information to obtain corresponding feature information;

[0073] The multi-head mutual attention module is used to extract multi-head mutual attention features from the feature information corresponding to each piece of data information to obtain mutual attention features;

[0074] The first processing network is used to extract features from various types of input data information to obtain mutual attention features;

[0075] The second processing network is used to perform fusion processing on the mutual attention features to obtain fusion features; the output end of the second processing network is connected to the input end of the third processing network;

[0076] The second processing network includes a first input module, a first convolution module, a depth-separable convolution module, a first dimension-raising convolution module, a second dimension-raising convolution module, a third dimension-raising convolution module, a fourth dimension-raising convolution module, a second convolution module, a first pooling module, a third convolution module and a first fully connected module;

[0077] The input end of the first input module is used to receive the mutual attention feature; the output end of the first input module of the second processing network is connected to the input end of the first convolution module of the second processing network; the output end of the first convolution module of the second processing network is connected to the input end of the depthwise separable convolution module of the second processing network; the output end of the depthwise separable convolution module of the second processing network is connected to the input end of the first dimensionality-raising convolution module of the second processing network;

[0078] The output end of the first dimensionality-raising convolution module of the second processing network is connected to the input end of the second dimensionality-raising convolution module of the second processing network; the output end of the second dimensionality-raising convolution module of the second processing network is connected to the input end of the third dimensionality-raising convolution module of the second processing network; the output end of the third dimensionality-raising convolution module of the second processing network is connected to the input end of the fourth dimensionality-raising convolution module of the second processing network; the output end of the fourth dimensionality-raising convolution module of the second processing network is connected to the input end of the second convolution module of the second processing network; the output end of the second convolution module of the second processing network is connected to the input end of the first pooling module of the second processing network; the output end of the first pooling module of the second processing network is connected to the input end of the third convolution module of the second processing network; the output end of the third convolution module of the second processing network is respectively connected to the input ends of the fifth dimensionality-raising convolution module and the first fully connected module of the second processing network; the output end of the fifth dimensionality-raising convolution module is connected to the input end of the first fully connected module of the second processing network;

[0079] The output end of the first processing network is connected to the input end of the second processing network;

[0080] The third processing network includes: a second input module, a fourth convolution module, a fifth convolution module, a sixth convolution module, a second pooling module, a seventh convolution module, a second fully connected module and a third fully connected module;

[0081] The input end of the second input module of the third processing network is connected to the output end of the first fully connected module of the second processing network; the output end of the second input module of the third processing network is connected to the input end of the fourth convolution module of the third processing network; the output end of the fourth convolution module of the third processing network is connected to the input end of the fifth convolution module of the third processing network; the output end of the fifth convolution module of the third processing network is connected to the input end of the sixth convolution module of the third processing network; the output end of the sixth convolution module of the third processing network is connected to the input end of the second pooling module of the third processing network; the output end of the second pooling module of the third processing network is connected to the input end of the seventh convolution module of the third processing network; the output end of the seventh convolution module of the third processing network is connected to the input end of the second fully connected module of the third processing network; the output end of the second fully connected module of the third processing network is connected to the input end of the third fully connected module of the third processing network.

[0082] The third processing network is used to perform prediction processing on the fusion feature to obtain an abnormal warning value of the input user outpatient data information;

[0083] The output end of the third fully-connected module of the third processing network is used to output the abnormal warning value of the input user outpatient data information.

[0084] The training process of the abnormal warning model includes:

[0085] A user historical outpatient data set is obtained; the user historical outpatient data set includes user outpatient data of a number of users within a set historical time period and corresponding label information; the label information is the infectious disease occurrence value corresponding to the user outpatient data; the user outpatient data is training data;

[0086] Initialize the number of training iterations;

[0087] Input the training data in the user's historical outpatient data set as input data into the abnormal warning model;

[0088] Using the abnormal warning model, the input data is processed to obtain a predicted value;

[0089] Performing difference calculation processing on the obtained predicted value and the label information corresponding to the input data to obtain a difference value;

[0090] Determine whether the difference value satisfies a convergence condition, and obtain a first determination result;

[0091] When the first judgment result is no, judging whether the training iteration number value is equal to the training number threshold, and obtaining a second judgment result;

[0092] When the second judgment result is no, determining that the model training state does not meet the training termination condition;

[0093] When the second judgment result is yes, determining that the model training state satisfies the training termination condition;

[0094] When the first judgment result is yes, determining that the model training state satisfies a training termination condition;

[0095] When the model training status does not meet the termination training condition, the parameter update model is used to update the parameters of the abnormal warning model, and the training iteration number value is increased by 1, triggering the execution of inputting the training data in the user's historical outpatient data set as input data into the abnormal warning model;

[0096] When the model training state satisfies the training termination condition, the training process of the abnormal warning model is completed to obtain a trained abnormal warning model.

[0097] The parameter updating model is:

[0098]

[0099] θ←θ+v;

[0100] In the formula, is the difference value calculated for the i-th training data in the user's historical outpatient data set, v is the parameter update value, θ is the parameter of the abnormal warning model, η is the initial parameter learning rate, α is the momentum angle parameter, 0≤α≤1, ▽ θ It means to find partial derivatives with respect to variable θ;

[0101] The difference value satisfies the convergence condition, which means that the difference value is less than a preset convergence threshold; the difference value does not satisfy the convergence condition, which means that the difference value is not less than the preset convergence threshold.

[0102] The difference calculation process may be implemented using a loss function.

[0103] The loss function may adopt a cross entropy loss function.

[0104] The user outpatient data set includes user outpatient data information of several users within a set time period; the user outpatient data information includes structured data and unstructured data; the structured data includes the user's age, gender, vital signs data, medication record data, outpatient medical record data, inpatient medical record data, and test report data; the structured data are all digital coding values ​​of the above information of the user; the unstructured data includes unstructured text data, such as detailed text information such as medical history, physical examination, diagnosis, and treatment written by a physician;

[0105] The text vector conversion process is implemented using the CountVectorizer (bag-of-words model) of the Scikit-learn library in Python.

[0106] The user historical outpatient data set includes user outpatient data of a number of users within a set historical time period and corresponding infectious disease occurrence values; the user outpatient data is constructed based on data vectors corresponding to structured data and unstructured data in the user historical outpatient data;

[0107] The infectious disease occurrence value is a labeling quantity set according to the prevalence of various infectious diseases within a set historical time period.

[0108] The first residual multi-head self-attention module, the second residual multi-head self-attention module, the third residual multi-head self-attention module, and the fourth residual multi-head self-attention module can be a Transformer model.

[0109] The multi-head mutual attention module is implemented using the Mutli Self-Attention module.

[0110] According to a second aspect of the present invention, an abnormal warning device for outpatient user data is disclosed, the device comprising:

[0111] A memory storing executable program code;

[0112] a processor coupled to the memory;

[0113] The processor calls the executable program code stored in the memory to execute the abnormal warning method for outpatient user data.

[0114] According to a third aspect of the present invention, a computer storable medium is disclosed, wherein the computer storable medium stores computer instructions, and when the computer instructions are called by a computer, they are used to execute the abnormal warning method for outpatient user data.

[0115] According to a fourth aspect of the present invention, an information data processing terminal is disclosed. The information data processing terminal is used to implement the abnormal warning method for outpatient user data.

[0116] The above description is only an embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent substitution, improvement, etc. made within the spirit and principle of the present invention should be included in the scope of the claims of the present invention.

Claims

1. A method for abnormal early warning of outpatient user data, characterized in that: include: S1, obtain the user's outpatient data set; The user outpatient data set includes user outpatient data information; S2, preprocessing the user outpatient data set to obtain user outpatient standard data information; S3, using the trained abnormal warning model to process the user's outpatient standard data information to obtain an abnormal warning value; The abnormal warning value is used to characterize the probability of an outpatient user developing an infectious disease.

2. The abnormal early warning method for outpatient user data according to claim 1, characterized in that: The preprocessing of the user outpatient data set to obtain user outpatient standard data information includes: Performing text recognition processing on the unstructured data in the user outpatient data set to obtain corresponding text data; Performing text vector conversion processing on the text data corresponding to the unstructured data to obtain a corresponding data vector; Using the data vectors and structured data corresponding to the unstructured data in the user outpatient data set, construct user outpatient standard data information; The user outpatient standard data information or user outpatient data information includes user data information, examination report data information, diagnosis report data information and medication record data information.

3. The abnormal early warning method for outpatient user data according to claim 2, characterized in that: The abnormal warning model includes: a first processing network, a second processing network, and a third processing network; The first processing network includes a first residual multi-head self-attention module, a second residual multi-head self-attention module, a third residual multi-head self-attention module, a fourth residual multi-head self-attention module and a multi-head mutual attention module; The first residual multi-head self-attention module, the second residual multi-head self-attention module, the third residual multi-head self-attention module and the fourth residual multi-head self-attention module are connected to the multi-head mutual attention module, and are used to respectively receive the user data information, the examination report data information, the diagnosis report data information and the medication record data information in the input user outpatient standard data information, and perform self-attention feature extraction on the received data information to obtain corresponding feature information; The multi-head mutual attention module is used to extract multi-head mutual attention features from the feature information corresponding to each piece of data information to obtain mutual attention features; The first processing network is used to extract features from various types of input data information to obtain mutual attention features; The output end of the first processing network is connected to the input end of the second processing network; The second processing network is used to fuse the mutual attention features to obtain fused features; the output end of the second processing network is connected to the input end of the third processing network.

4. The abnormal early warning method for outpatient user data according to claim 3, characterized in that: The second processing network includes a first input module, a first convolution module, a depth-separable convolution module, a first dimension-raising convolution module, a second dimension-raising convolution module, a third dimension-raising convolution module, a fourth dimension-raising convolution module, a second convolution module, a first pooling module, a third convolution module and a first fully connected module; The input end of the first input module is used to receive the mutual attention feature; the output end of the first input module of the second processing network is connected to the input end of the first convolution module of the second processing network; the output end of the first convolution module of the second processing network is connected to the input end of the depthwise separable convolution module of the second processing network; the output end of the depthwise separable convolution module of the second processing network is connected to the input end of the first dimensionality-raising convolution module of the second processing network; The output end of the first dimensionality-raising convolution module of the second processing network is connected to the input end of the second dimensionality-raising convolution module of the second processing network; the output end of the second dimensionality-raising convolution module of the second processing network is connected to the input end of the third dimensionality-raising convolution module of the second processing network; the output end of the third dimensionality-raising convolution module of the second processing network is connected to the input end of the fourth dimensionality-raising convolution module of the second processing network; the output end of the fourth dimensionality-raising convolution module of the second processing network is connected to the input end of the second convolution module of the second processing network; the output end of the second convolution module of the second processing network is connected to the input end of the first pooling module of the second processing network; the output end of the first pooling module of the second processing network is connected to the input end of the third convolution module of the second processing network; the output end of the third convolution module of the second processing network is respectively connected to the input ends of the fifth dimensionality-raising convolution module and the first fully connected module of the second processing network; the output end of the fifth dimensionality-raising convolution module is connected to the input end of the first fully connected module of the second processing network; The output end of the first processing network is connected to the input end of the second processing network.

5. The abnormal early warning method for outpatient user data according to claim 4, characterized in that: The third processing network includes: a second input module, a fourth convolution module, a fifth convolution module, a sixth convolution module, a second pooling module, a seventh convolution module, a second fully connected module and a third fully connected module; The input end of the second input module of the third processing network is connected to the output end of the first fully connected module of the second processing network; the output end of the second input module of the third processing network is connected to the input end of the fourth convolution module of the third processing network; the output end of the fourth convolution module of the third processing network is connected to the input end of the fifth convolution module of the third processing network; the output end of the fifth convolution module of the third processing network is connected to the input end of the sixth convolution module of the third processing network; the output end of the sixth convolution module of the third processing network is connected to the input end of the second pooling module of the third processing network; the output end of the second pooling module of the third processing network is connected to the input end of the seventh convolution module of the third processing network; the output end of the seventh convolution module of the third processing network is connected to the input end of the second fully connected module of the third processing network; the output end of the second fully connected module of the third processing network is connected to the input end of the third fully connected module of the third processing network; The third processing network is used to perform prediction processing on the fusion feature to obtain an abnormal warning value of the input user outpatient standard data information; The output end of the third fully-connected module of the third processing network is used to output the abnormal warning value of the input user outpatient standard data information.

6. The abnormal early warning method for outpatient user data according to claim 5, characterized in that: The training process of the abnormal warning model includes: A user historical outpatient data set is obtained; the user historical outpatient data set includes user outpatient data of a number of users within a set historical time period and corresponding label information; the label information is the infectious disease occurrence value corresponding to the user outpatient data; the user outpatient data is training data; Initialize the number of training iterations; Input the training data in the user's historical outpatient data set as input data into the abnormal warning model; Using the abnormal warning model, the input data is processed to obtain a predicted value; Performing difference calculation processing on the obtained predicted value and the label information corresponding to the input data to obtain a difference value; Determine whether the difference value satisfies a convergence condition, and obtain a first determination result; When the first judgment result is no, judging whether the training iteration number value is equal to the training number threshold, and obtaining a second judgment result; When the second judgment result is no, determining that the model training state does not meet the training termination condition; When the second judgment result is yes, determining that the model training state satisfies the training termination condition; When the first judgment result is yes, determining that the model training state satisfies a training termination condition; When the model training status does not meet the termination training condition, the parameter update model is used to update the parameters of the abnormal warning model, and the training iteration number value is increased by 1, triggering the execution of inputting the training data in the user's historical outpatient data set as input data into the abnormal warning model; When the model training state satisfies the training termination condition, the training process of the abnormal warning model is completed to obtain a trained abnormal warning model.

7. The abnormal early warning method for outpatient user data according to claim 6, characterized in that: The parameter updating model is: In the formula, is the difference value calculated for the i-th training data in the user's historical outpatient data set, v is the parameter update value, θ is the parameter of the abnormal warning model, η is the initial parameter learning rate, α is the momentum angle parameter, 0≤α≤1, It means to find the partial derivative with respect to the variable θ.

8. An abnormal warning device for outpatient user data, characterized in that: The device comprises: A memory storing executable program code; a processor coupled to the memory; The processor calls the executable program code stored in the memory to execute the abnormal warning method for outpatient user data according to any one of claims 1 to 7.

9. A computer storable medium, characterized in that: The computer storable medium stores computer instructions, and when the computer instructions are called by a computer, they are used to execute the abnormal warning method for outpatient user data according to any one of claims 1 to 7.

10. An information data processing terminal, characterized in that: The information data processing terminal is used to implement the abnormal early warning method for outpatient user data according to any one of claims 1 to 7.

Images