GRU model and SEIR model-based method for predicting number of infectious disease daily infected people

Abstract

The invention discloses a method for predicting the number of infectious disease daily infected people based on a GRU model and an SEIR model. The method comprises the following steps: 1, acquiring a historical infected people number sequence, a historical positive rate sequence, a historical variant distribution sequence and a historical contact rate sequence; 2, inputting the historical infected people number sequence, the historical positive rate sequence, the historical variant distribution sequence and the historical contact rate sequence in the step 1 into a pre-trained infectious disease day infected people number prediction model to obtain a predicted day infected people number sequence; wherein the pre-trained infectious disease day infection number prediction model is constructed based on a GRU model and an SEIR model. The advantages of the neural network model and the traditional infectious disease model are fused, so that the prediction accuracy is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of infectious disease epidemic analysis, and more specifically to a method for predicting the number of infectious disease infections per day based on a GRU model and a SEIR model. Background Art

[0002] At present, the methods for predicting the number of people susceptible to infectious diseases either use traditional infectious disease models (such as SI, SIR, SIRS, SEIR) or use neural network models for prediction;

[0003] Although infectious disease models have certain reference value in predicting the number of susceptible people, the relevant parameters in infectious disease models are often fixed, which leads to low accuracy of model predictions; although neural network models perform well in predicting the number of susceptible people, the prediction accuracy still needs to be further improved; and there is a prominent problem with neural network models: they cannot be organically combined with traditional infectious disease models. Traditional infectious disease models (such as SI, SIR, SIRS, SEIR) have accumulated rich experience and knowledge through long-term research and practical verification, and have a deep understanding of the transmission mechanism and influencing factors of the epidemic. However, due to the complexity and particularity of neural network models, it is difficult to directly integrate them into these traditional infectious disease models, thus failing to fully utilize the valuable experience and knowledge of past research.

[0004] Therefore, how to provide a method for predicting the number of daily infections of an infectious disease, which can integrate the advantages of neural network models and traditional infectious disease models, thereby improving the prediction accuracy is an urgent problem that technical personnel in this field need to solve. Summary of the invention

[0005] In view of this, the purpose of the present invention is to provide a method for predicting the number of daily infectious diseases based on the GRU model and the SEIR model.

[0006] In order to achieve the above object, the present invention adopts the following technical solution:

[0007] In a first aspect, a method for predicting the number of people infected by an infectious disease per day based on a GRU model and a SEIR model is provided, comprising the following steps:

[0008] Step 1: Obtain the historical infection number sequence, historical positive rate sequence, historical variant strain distribution sequence and historical contact rate sequence;

[0009] Step 2: Input the historical number of infected people sequence, historical positive rate sequence, historical variant strain distribution sequence and historical contact rate sequence in step 1 into the pre-trained infectious disease daily infection number prediction model to obtain the predicted daily infection number sequence; wherein, the pre-trained infectious disease daily infection number prediction model is constructed based on the GRU model and the SEIR model.

[0010] Preferably, the infectious disease daily infection number prediction model includes a first GRU model, a second GRU model, a third GRU model, a fourth GRU model, a fifth GRU model, a sixth GRU model, a seventh GRU model, a first matrix splicing unit, a second matrix splicing unit, a third matrix splicing unit, a first multilayer perceptron, a second multilayer perceptron, a third multilayer perceptron, a SEIR model and a matrix addition unit; wherein the SEIR model is a SEIR model in the form of a standard deviation;

[0011] The output end of the second GRU model, the output end of the third GRU model, and the output end of the fourth GRU model are all connected to the input end of the first multilayer perceptron through the first matrix splicing unit;

[0012] The output end of the first GRU model and the output end of the first multi-layer perceptron are both connected to the input end of the second matrix splicing unit;

[0013] The output end of the second matrix splicing unit is connected to the input end of the SEIR model through the second multi-layer perceptron;

[0014] The output end of the fifth GRU model, the output end of the sixth GRU model, and the output end of the seventh GRU model are all connected to the input end of the third multilayer perceptron through the third matrix splicing unit;

[0015] The output end of the third multilayer perceptron and the output end of the SEIR model are both connected to the input end of the matrix addition unit;

[0016] The input end of the second GRU model is used to input a historical sequence of infected people;

[0017] The input end of the fifth GRU model is used to input a historical sequence of the number of infected persons;

[0018] The input end of the third GRU model is used to input a historical positive rate sequence;

[0019] The input end of the sixth GRU model is used to input a historical positive rate sequence;

[0020] The input end of the fourth GRU model is used to input the historical variant distribution sequence;

[0021] The input end of the seventh GRU model is used to input the historical variant distribution sequence;

[0022] The input end of the first GRU model is used to input a historical contact rate sequence;

[0023] The output end of the matrix addition unit is used to output a sequence of predicted daily infected persons.

[0024] Preferably, the pre-trained infectious disease daily infection number prediction model is obtained based on the following steps:

[0025] S1: Construct a data set; wherein the data set includes several data pairs, each data pair includes input data and a prediction label; the input data includes a historical infection number sequence, a historical positive rate sequence, a historical variant strain distribution sequence, and a historical contact rate sequence; the prediction label is a predicted daily infection number sequence corresponding to the input data;

[0026] S2: Divide the data set into a training set, a validation set and a test set;

[0027] S3: performing standardization processing on the training set, the validation set and the test set;

[0028] S4: using a standardized training set to train the infectious disease daily infection number prediction model;

[0029] S5: Using a standardized validation set to adjust the model parameters of the infectious disease daily infection number prediction model; wherein the model parameters include the first GRU model, the second GRU model, the third GRU model, the fourth GRU model, the fifth GRU model, the sixth GRU model, the seventh GRU model, the first multilayer perceptron, the second multilayer perceptron, and the network parameters of the third multilayer perceptron;

[0030] S6: Use a standardized test set to evaluate the prediction effect of the infectious disease daily infection number prediction model to obtain the pre-trained infectious disease daily infection number prediction model.

[0031] Preferably, step 2 further comprises:

[0032] The historical infection number sequence, historical positivity rate sequence, historical variant strain distribution sequence and historical contact rate sequence in step one are standardized and then input into the pre-trained infectious disease daily infection number prediction model.

[0033] Preferably, the formula for the standardization process is:

[0034]

[0035] Among them, x' represents the data of the number of infections before standardization in the dataset, or represents the data of the positive rate before standardization in the dataset, or represents the data of the distribution of variants before standardization in the dataset, or represents the data of the contact rate before standardization in the dataset; x' represents the data of the number of infections after standardization in the dataset, or represents the data of the positive rate after standardization in the dataset, or represents the data of the distribution of variants after standardization in the dataset, or represents the data of the contact rate after standardization in the dataset; μ represents the average value of the data of the number of infections before standardization in the training set, or represents the average value of the data of the positive rate before standardization in the training set, or represents the average value of the data of the distribution of variants before standardization in the training set, or represents the average value of the data of the contact rate before standardization in the training set; σ represents the standard deviation of the data of the number of infections before standardization in the training set, or represents the standard deviation of the data of the positive rate before standardization in the training set, or represents the standard deviation of the data of the distribution of variants before standardization in the training set, or represents the standard deviation of the data of the contact rate before standardization in the training set.

[0036] Preferably, the initial number of susceptible persons S is used 0 , the initial number of lurkers E 0 、The number of initial infected people I 0 , the number of initial removers R 0 Solve the SEIR model in differential form to obtain the historical contact rate series in step 1 and the historical contact rate series in the data set;

[0037] Among them, the expression of the difference form of the SEIR model is:

[0038] S t+1 =S t -β t ·S t I t Δt;

[0039] E t+1 =E t +β t ·S t I t ·Δt-σ·E t Δt;

[0040] I t+1 =I t +σ·E t ·Δt-γ·I t Δt;

[0041] R t+1 =R t +γ·I t Δt;

[0042] N=S t +E t +I t +R t ;

[0043] in:

[0044] S t+1 , S t represents the number of susceptible persons at time t+1 and time t respectively; E t+1 、E t I represents the number of lurkers at time t+1 and time t respectively; t+1 ,I t represents the number of infected people at time t+1 and time t respectively; R t+1 , R t Represents the number of removers at time t+1 and time t respectively; β t+1 represents the contact rate data at time t+1; Δt represents the time step; σ represents the rate at which latent persons become infected persons; γ represents the rate at which infected persons recover or die; and N represents the total population.

[0045] Preferably, the fourth GRU model is connected to the first matrix splicing unit through an accumulation unit; wherein the accumulation unit is used to vertically accumulate and sum the feature vectors of size N1*N2 output by the fourth GRU model to obtain a feature vector of size N1*1.

[0046] In a second aspect, an electronic device is provided, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, the method for predicting the daily number of infected persons of an infectious disease based on a GRU model and a SEIR model as described in any one of the above items is implemented.

[0047] In a third aspect, a non-transitory computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the method for predicting the daily number of infected people of an infectious disease based on the GRU model and the SEIR model as described in any of the above items is implemented.

[0048] In a fourth aspect, a computer program product is provided, comprising a computer program, which, when executed by a processor, implements the method for predicting the daily number of infected people of an infectious disease based on a GRU model and a SEIR model as described in any one of the above items.

[0049] It can be seen from the above technical solution that, compared with the prior art, the present invention discloses a method for predicting the number of people infected by an infectious disease per day. The present invention can integrate the advantages of a neural network model and a traditional infectious disease model, thereby improving the prediction accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying creative work.

[0051] Figure 1 A schematic diagram of the structure of a model for predicting the number of people infected by an infectious disease per day provided in an embodiment of the present invention;

[0052] Figure 2 The present invention is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0053] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0054] On the one hand, an embodiment of the present invention discloses a method for predicting the number of daily infections of an infectious disease based on a GRU model and a SEIR model, comprising the following steps:

[0055] Step 1: Obtain the historical infection number sequence, historical positive rate sequence, historical variant strain distribution sequence and historical contact rate sequence;

[0056] In a certain embodiment: the historical number of infections sequence has a sampling frequency of daily; the historical positivity rate sequence has a sampling frequency of weekly; the historical variant strain distribution sequence has a sampling frequency of weekly; and the historical contact rate sequence has a sampling frequency of daily.

[0057] Step 2: Input the historical number of infected people sequence, historical positive rate sequence, historical variant strain distribution sequence and historical contact rate sequence in step 1 into the pre-trained infectious disease daily infection number prediction model to obtain the predicted daily infection number sequence; wherein, the pre-trained infectious disease daily infection number prediction model is constructed based on the GRU model and the SEIR model.

[0058] In one embodiment, if Figure 1As shown, the infectious disease daily infection number prediction model includes a first GRU model, a second GRU model, a third GRU model, a fourth GRU model, a fifth GRU model, a sixth GRU model, a seventh GRU model, a first matrix splicing unit, a second matrix splicing unit, a third matrix splicing unit, a first multilayer perceptron, a second multilayer perceptron, a third multilayer perceptron, a SEIR model and a matrix addition unit; wherein the SEIR model is a SEIR model in the form of a standard deviation;

[0059] The output end of the second GRU model, the output end of the third GRU model, and the output end of the fourth GRU model are all connected to the input end of the first multilayer perceptron through the first matrix splicing unit;

[0060] The output end of the first GRU model and the output end of the first multi-layer perceptron are both connected to the input end of the second matrix splicing unit;

[0061] The output end of the second matrix splicing unit is connected to the input end of the SEIR model through the second multi-layer perceptron;

[0062] The output end of the fifth GRU model, the output end of the sixth GRU model, and the output end of the seventh GRU model are all connected to the input end of the third multilayer perceptron through the third matrix splicing unit;

[0063] The output end of the third multilayer perceptron and the output end of the SEIR model are both connected to the input end of the matrix addition unit;

[0064] The input end of the second GRU model is used to input a historical sequence of infected people;

[0065] The input end of the fifth GRU model is used to input a historical sequence of the number of infected persons;

[0066] The input end of the third GRU model is used to input a historical positive rate sequence;

[0067] The input end of the sixth GRU model is used to input a historical positive rate sequence;

[0068] The input end of the fourth GRU model is used to input the historical variant distribution sequence;

[0069] The input end of the seventh GRU model is used to input the historical variant distribution sequence;

[0070] The input end of the first GRU model is used to input a historical contact rate sequence;

[0071] The output end of the matrix addition unit is used to output a sequence of predicted daily infected persons.

[0072] In one embodiment, step 2 specifically includes the following steps:

[0073] StepA: Convert the historical infection number sequence (A t-L ,A t-L+1 ,…,A t-1 ) is input into the second GRU model to obtain the historical number of infected people sequence feature representation vector;

[0074] The historical positive rate sequence (B t-L ,B t-L+1 ,…,B t-1 ) is input into the third GRU model to obtain the historical positive rate sequence feature representation vector;

[0075] The historical variant distribution sequence (C t-L ,C t-L+1 ,…,C t-1 ) is input into the fourth GRU model to obtain the historical variant distribution sequence feature representation vector;

[0076] The historical number of infected people sequence feature representation vector, the historical positive rate sequence feature representation vector, and the historical variant strain distribution sequence feature representation vector are sequentially input into the first matrix splicing unit and the first multi-layer perceptron to obtain a multi-source comprehensive representation vector;

[0077] The historical contact rate series (β t-L ,β t-L+1 ,…,β t-1 ) is input into the first GRU model to obtain a historical contact rate sequence feature representation vector;

[0078] The multi-source comprehensive representation vector and the historical contact rate sequence feature representation vector are sequentially input into the second matrix concatenation unit and the second multi-layer perceptron to obtain the predicted contact rate sequence (β t , β t+1 ,…,β T );

[0079] The predicted contact rate series (β t , β t+1 ,…,β T ) is input into the SEIR model to obtain the SEIR predicted daily infection number sequence (I t ,I t+1 ,...,I T );

[0080] It is understandable that: when β t-1 , S t-1 、E t-1 ,I t-1 , R t-1When , Δt, σ, and γ are known, it can be inferred that S t ,E t ,I t ,R t The value of β t , S t 、E t ,I t , R t When , Δt, σ, and γ are known, it can be inferred that S t+1 ,E t+1 ,I t+1 ,R t+1 The value of β t+1 , S t+1 、E t+1 ,I t+1 , R t+1 When , Δt, σ, and γ are known, it can be inferred that S t+2 ,E t+2 ,I t+2 ,R t+2 The value of S can be inferred by recursively predicting it. T ,E T ,I T ,R T ; Then we can get the SEIR predicted daily infection number sequence (I t ,I t+1 ,...,I T );

[0081] Step B: Convert the historical infection number sequence (A t-L ,A t-L+1 ,…,A t-1 ) is input into the fifth GRU model to obtain the historical number of infected people sequence feature representation vector;

[0082] The historical positive rate sequence (B t-L ,B t-L+1 ,…,B t-1 ) is input to the sixth GRU model to obtain a historical positive rate sequence feature representation vector;

[0083] The historical variant distribution sequence (C t-L ,C t-L+1 ,…,C t-1 ) is input into the seventh GRU model to obtain the historical variant distribution sequence feature representation vector;

[0084] The historical number of infected people sequence feature representation vector, historical positive rate sequence feature representation vector, and historical variant strain distribution sequence feature representation vector obtained in Step B are sequentially input into the third matrix splicing unit and the third multi-layer perceptron to obtain a multi-source predicted daily number of infected people sequence;

[0085] Step C: Input the SEIR predicted daily infection number sequence and the multi-source predicted daily infection number sequence into the matrix addition unit to obtain the predicted daily infection number sequence in step 2.

[0086] In one embodiment, the pre-trained infectious disease daily infection number prediction model is obtained based on the following steps:

[0087] S1: Construct a data set; wherein the data set includes several data pairs, each data pair includes input data and a prediction label; the input data includes a historical infection number sequence, a historical positive rate sequence, a historical variant strain distribution sequence, and a historical contact rate sequence; the prediction label is a predicted daily infection number sequence corresponding to the input data;

[0088] It can be understood that the data set is obtained based on the following method:

[0089] Assuming that there are data from T moments in total (historical infection number data / historical positivity rate data / historical variant strain distribution data / historical contact rate data), the present invention extracts data from t to t+M as prediction labels, and extracts data from tL to t-1 as input data, then T-(M+L-1) {input data, prediction label} pairs can be obtained.

[0090] S2: Divide the data set into a training set, a validation set and a test set;

[0091] S3: performing standardization processing on the training set, the validation set and the test set;

[0092] S4: using a standardized training set to train the infectious disease daily infection number prediction model;

[0093] Specifically, it includes constructing a loss function, selecting an optimizer, and performing backpropagation training on the training set.

[0094] S5: Using a standardized validation set to adjust the model parameters of the infectious disease daily infection number prediction model; wherein the model parameters include the first GRU model, the second GRU model, the third GRU model, the fourth GRU model, the fifth GRU model, the sixth GRU model, the seventh GRU model, the first multilayer perceptron, the second multilayer perceptron, and the network parameters of the third multilayer perceptron;

[0095] S6: Use a standardized test set to evaluate the prediction effect of the infectious disease daily infection number prediction model to obtain the pre-trained infectious disease daily infection number prediction model.

[0096] Specifically, the present invention uses the rmse or peak prediction of the test set to evaluate the prediction effect of the model.

[0097] In one embodiment, step 2 further comprises:

[0098] The historical infection number sequence, historical positivity rate sequence, historical variant strain distribution sequence and historical contact rate sequence in step one are standardized and then input into the pre-trained infectious disease daily infection number prediction model.

[0099] In one embodiment, the formula for the normalization process is:

[0100]

[0101] Among them, x' represents the data of the number of infections before standardization in the dataset, or represents the data of the positive rate before standardization in the dataset, or represents the data of the distribution of variants before standardization in the dataset, or represents the data of the contact rate before standardization in the dataset; x' represents the data of the number of infections after standardization in the dataset, or represents the data of the positive rate after standardization in the dataset, or represents the data of the distribution of variants after standardization in the dataset, or represents the data of the contact rate after standardization in the dataset; μ represents the average value of the data of the number of infections before standardization in the training set, or represents the average value of the data of the positive rate before standardization in the training set, or represents the average value of the data of the distribution of variants before standardization in the training set, or represents the average value of the data of the contact rate before standardization in the training set; σ represents the standard deviation of the data of the number of infections before standardization in the training set, or represents the standard deviation of the data of the positive rate before standardization in the training set, or represents the standard deviation of the data of the distribution of variants before standardization in the training set, or represents the standard deviation of the data of the contact rate before standardization in the training set.

[0102] In one embodiment, the number of initial susceptible persons S 0 , the initial number of lurkers E 0 、The number of initial infected people I 0 , the number of initial removers R 0 Solve the SEIR model in differential form to obtain the historical contact rate series in step 1 and the historical contact rate series in the data set;

[0103] Specifically: due to the number of infected people I at any time t (here any time t is the current time or a time before the current time) t , the initial number of susceptible people S 0 , the initial number of lurkers E 0 、The number of initial infected people I 0 , the number of initial removers R 0, time step Δt, σ and γ (σ and γ are fixed values, which can often be obtained through relevant epidemiological analysis) are all known quantities. Then, using these known quantities, we can solve the equation system to obtain the β in the difference form of the SEIR model. t ; Among them, β t That is, the contact rate data at any time t (the arbitrary time t here refers to the current time or a time before the current time); (the SEIR contact rate parameter estimation here refers to more mature methods such as least squares method and numerical simulation technology for estimation);

[0104] Based on the contact rate data β at any time t (the arbitrary time t here is the current time or a time before the current time) t , then the historical contact rate sequence in step 1 and the historical contact rate sequence in the data set can be selected;

[0105] Among them, the expression of the difference form of the SEIR model is:

[0106] S t+1 =S t -β t ·S t I t Δt;

[0107] E t+1 =E t +β t ·S t I t ·Δt-σ·E t Δt;

[0108] I t+1 =I t +σ·E t ·Δt-γ·I t Δt;

[0109] R t+1 =R t +γ·I t Δt;

[0110] N=S t +E t +I t +R t ;

[0111] in:

[0112] S t+1 , S t represents the number of susceptible persons at time t+1 and time t respectively; E t+1 、E tI represents the number of lurkers at time t+1 and time t respectively; t+1 ,I t represents the number of infected people at time t+1 and time t respectively; R t+1 , R t Represents the number of removers at time t+1 and time t respectively; β t+1 represents the contact rate data at time t+1; Δt represents the time step; σ represents the rate at which latent persons become infected persons; γ represents the rate at which infected persons recover or die; and N represents the total population.

[0113] In a certain embodiment, the fourth GRU model is connected to the first matrix splicing unit through an accumulation unit; wherein the accumulation unit is used to vertically accumulate and sum the feature vectors of size N1*N2 output by the fourth GRU model to obtain the feature vectors of size N1*1.

[0114] On the other hand, an embodiment of the present invention discloses an electronic device, such as Figure 2 As shown, the electronic device may include: a processor 201, a communication interface 202, a memory 203 and a communication bus 204, wherein the processor 201, the communication interface 202 and the memory 203 communicate with each other through the communication bus 204. The processor 201 may call the logic instructions in the memory 203 to execute the method for predicting the number of infected persons per day of an infectious disease based on the GRU model and the SEIR model.

[0115] In addition, the logic instructions in the above-mentioned memory 203 can be implemented in the form of software functional units and can be stored in a computer-readable storage medium when sold or used as an independent product. Based on such an understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk and other media that can store program codes.

[0116] On the other hand, the present invention also provides a computer program product, which includes a computer program. The computer program can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can execute the infectious disease daily infection number prediction method based on the GRU model and the SEIR model provided by the above-mentioned methods.

[0117] On the other hand, the present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, is implemented to execute the method for predicting the daily number of infectious diseases based on the GRU model and the SEIR model provided by the above-mentioned methods.

[0118] The device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Ordinary technicians in this field can understand and implement it without paying creative labor.

[0119] Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0120] In this specification, each embodiment is described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the embodiments can be referred to each other. For the device disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant parts can be referred to the method part.

[0121] The above description of the disclosed embodiments enables one skilled in the art to implement or use the present invention. Various modifications to these embodiments will be apparent to one skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but rather to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for predicting the number of people infected by infectious diseases based on the GRU model and the SEIR model, characterized in that: The steps include: Step 1: Obtain the historical infection number sequence, historical positive rate sequence, historical variant strain distribution sequence and historical contact rate sequence; Step 2: Input the historical number of infected people sequence, historical positive rate sequence, historical variant strain distribution sequence and historical contact rate sequence in step 1 into the pre-trained infectious disease daily infection number prediction model to obtain the predicted daily infection number sequence; wherein, the pre-trained infectious disease daily infection number prediction model is constructed based on the GRU model and the SEIR model.

2. According to claim 1, a method for predicting the number of people infected by an infectious disease based on a GRU model and a SEIR model, characterized in that: The infectious disease daily infection number prediction model includes a first GRU model, a second GRU model, a third GRU model, a fourth GRU model, a fifth GRU model, a sixth GRU model, a seventh GRU model, a first matrix splicing unit, a second matrix splicing unit, a third matrix splicing unit, a first multilayer perceptron, a second multilayer perceptron, a third multilayer perceptron, a SEIR model and a matrix addition unit; wherein the SEIR model is a SEIR model in the form of a standard deviation; The output end of the second GRU model, the output end of the third GRU model, and the output end of the fourth GRU model are all connected to the input end of the first multilayer perceptron through the first matrix splicing unit; The output end of the first GRU model and the output end of the first multi-layer perceptron are both connected to the input end of the second matrix splicing unit; The output end of the second matrix splicing unit is connected to the input end of the SEIR model through the second multi-layer perceptron; The output end of the fifth GRU model, the output end of the sixth GRU model, and the output end of the seventh GRU model are all connected to the input end of the third multilayer perceptron through the third matrix splicing unit; The output end of the third multilayer perceptron and the output end of the SEIR model are both connected to the input end of the matrix addition unit; The input end of the second GRU model is used to input a historical sequence of infected people; The input end of the fifth GRU model is used to input a historical sequence of the number of infected persons; The input end of the third GRU model is used to input a historical positive rate sequence; The input end of the sixth GRU model is used to input a historical positive rate sequence; The input end of the fourth GRU model is used to input the historical variant distribution sequence; The input end of the seventh GRU model is used to input the historical variant distribution sequence; The input end of the first GRU model is used to input a historical contact rate sequence; The output end of the matrix addition unit is used to output a sequence of predicted daily infected persons.

3. The method for predicting the number of people infected by an infectious disease based on a GRU model and a SEIR model according to claim 2, characterized in that: The pre-trained infectious disease daily infection number prediction model is obtained based on the following steps: S1: Construct a data set; wherein the data set includes several data pairs, each data pair includes input data and a prediction label; the input data includes a historical infection number sequence, a historical positive rate sequence, a historical variant strain distribution sequence, and a historical contact rate sequence; the prediction label is a predicted daily infection number sequence corresponding to the input data; S2: Divide the data set into a training set, a validation set and a test set; S3: performing standardization processing on the training set, the validation set and the test set; S4: using a standardized training set to train the infectious disease daily infection number prediction model; S5: Using a standardized validation set to adjust the model parameters of the infectious disease daily infection number prediction model; wherein the model parameters include the first GRU model, the second GRU model, the third GRU model, the fourth GRU model, the fifth GRU model, the sixth GRU model, the seventh GRU model, the first multilayer perceptron, the second multilayer perceptron, and the network parameters of the third multilayer perceptron; S6: Use a standardized test set to evaluate the prediction effect of the infectious disease daily infection number prediction model to obtain the pre-trained infectious disease daily infection number prediction model.

4. The method for predicting the number of people infected by an infectious disease based on a GRU model and a SEIR model according to claim 3, characterized in that: Step 2 further includes: The historical infection number sequence, historical positivity rate sequence, historical variant strain distribution sequence and historical contact rate sequence in step one are standardized and then input into the pre-trained infectious disease daily infection number prediction model.

5. The method for predicting the number of people infected by an infectious disease based on a GRU model and a SEIR model according to claim 4, characterized in that: The formula for the standardization process is: Among them, x' represents the data of the number of infections before standardization in the dataset, or represents the data of the positive rate before standardization in the dataset, or represents the data of the distribution of variants before standardization in the dataset, or represents the data of the contact rate before standardization in the dataset; x' represents the data of the number of infections after standardization in the dataset, or represents the data of the positive rate after standardization in the dataset, or represents the data of the distribution of variants after standardization in the dataset, or represents the data of the contact rate after standardization in the dataset; μ represents the average value of the data of the number of infections before standardization in the training set, or represents the average value of the data of the positive rate before standardization in the training set, or represents the average value of the data of the distribution of variants before standardization in the training set, or represents the average value of the data of the contact rate before standardization in the training set; σ represents the standard deviation of the data of the number of infections before standardization in the training set, or represents the standard deviation of the data of the positive rate before standardization in the training set, or represents the standard deviation of the data of the distribution of variants before standardization in the training set, or represents the standard deviation of the data of the contact rate before standardization in the training set.

6. The method for predicting the number of people infected by an infectious disease based on a GRU model and a SEIR model according to claim 5, characterized in that: Solve the difference SEIR model using the initial number of susceptible people S0, the initial number of latent people E0, the initial number of infected people I0, and the initial number of removed people R0 to obtain the historical contact rate sequence in step 1 and the historical contact rate sequence in the data set; Among them, the expression of the difference form of the SEIR model is: S t+1 =S t -β t ·S t ·I t ·Δt; E t+1 =E t +b t ·S t ·I t ·Δt-σ·E t ·Δt; I t+1 =I t +s·E t ·Δt-γ·I t ·Δt; R t+1 =R t +γ·I t ·Δt; N=S t +E t +I t +R t ; in: S t+1 , S t represents the number of susceptible persons at time t+1 and time t respectively; E t+1 、E t I represents the number of lurkers at time t+1 and time t respectively; t+1 ,I t represents the number of infected people at time t+1 and time t respectively; R t+1 , R t Represents the number of removers at time t+1 and time t respectively; β t+1 represents the contact rate data at time t+1; Δt represents the time step; σ represents the rate at which latent persons become infected persons; γ represents the rate at which infected persons recover or die; and N represents the total population.

7. The method for predicting the number of people infected by an infectious disease based on a GRU model and a SEIR model according to claim 2, characterized in that: The fourth GRU model is connected to the first matrix splicing unit through an accumulation unit; wherein the accumulation unit is used to vertically accumulate and sum the feature vectors of size N1*N2 output by the fourth GRU model to obtain a feature vector of size N1*1.

8. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the computer program, it implements the method for predicting the number of infectious diseases per day based on the GRU model and the SEIR model as described in any one of claims 1 to 7.

9. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method for predicting the number of people infected by an infectious disease per day based on a GRU model and a SEIR model as described in any one of claims 1 to 7 is implemented.

10. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the method for predicting the number of people infected by an infectious disease per day based on a GRU model and a SEIR model as described in any one of claims 1 to 7 is implemented.

Images