CN113113155A - Infectious disease trend prediction method based on neural network and SEIR model - Google Patents

Infectious disease trend prediction method based on neural network and SEIR model Download PDF

Info

Publication number
CN113113155A
CN113113155A CN202110658591.1A CN202110658591A CN113113155A CN 113113155 A CN113113155 A CN 113113155A CN 202110658591 A CN202110658591 A CN 202110658591A CN 113113155 A CN113113155 A CN 113113155A
Authority
CN
China
Prior art keywords
epidemic situation
patients
trend prediction
infection rate
seir
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
CN202110658591.1A
Other languages
Chinese (zh)
Inventor
王建勇
章毅
甘雨
吴雨
庞博
吴宇杭
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Sichuan University
Original Assignee
Sichuan University
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Sichuan University filed Critical Sichuan University
Priority to CN202110658591.1A priority Critical patent/CN113113155A/en
Publication of CN113113155A publication Critical patent/CN113113155A/en
Pending legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
    • G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
    • G16H50/00—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics
    • G16H50/80—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for detecting, monitoring or modelling epidemics or pandemics, e.g. flu
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/044—Recurrent networks, e.g. Hopfield networks
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/047—Probabilistic or stochastic networks
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods
    • G06N3/084—Backpropagation, e.g. using gradient descent

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Health & Medical Sciences (AREA)
  • Biomedical Technology (AREA)
  • Data Mining & Analysis (AREA)
  • General Health & Medical Sciences (AREA)
  • General Physics & Mathematics (AREA)
  • Mathematical Physics (AREA)
  • Biophysics (AREA)
  • Evolutionary Computation (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Molecular Biology (AREA)
  • Computing Systems (AREA)
  • General Engineering & Computer Science (AREA)
  • Artificial Intelligence (AREA)
  • Computational Linguistics (AREA)
  • Software Systems (AREA)
  • Public Health (AREA)
  • Medical Informatics (AREA)
  • Probability & Statistics with Applications (AREA)
  • Databases & Information Systems (AREA)
  • Pathology (AREA)
  • Epidemiology (AREA)
  • Primary Health Care (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)

Abstract

The invention relates to the field of artificial intelligence prediction, in particular to an infectious disease trend prediction method based on a neural network and an SEIR model, which comprises the following steps: step 1, acquiring a data set; step 2, preprocessing data; step 3, constructing an epidemic situation trend prediction model consisting of a virus infection rate prediction module and an epidemic situation trend prediction module; step 4, training the epidemic situation trend prediction model by using the data preprocessed in the step 2, and setting a loss function and a model parameter updating mode of the epidemic situation trend prediction model; step 5, predicting the epidemic situation trend by using the epidemic situation trend prediction model trained in the step 1 to the step 4; the method can effectively and dynamically predict the epidemic situation trend of the infectious diseases in real time by using less training data, and the prediction process does not need human intervention.

Description

Infectious disease trend prediction method based on neural network and SEIR model
Technical Field
The invention relates to the field of artificial intelligence prediction, in particular to an infectious disease trend prediction method based on a neural network and an SEIR model.
Background
With the development of machine learning in recent years, prediction methods can be divided into traditional methods without using a machine learning method and methods using machine learning, and the traditional methods use traditional static infectious disease mathematical models (such as an SIR model, an SEIR model and the like) to model and predict the spread of an epidemic situation.
In the traditional infectious disease epidemic situation trend prediction method, parameters in a prediction model cannot be self-adapted, so that dynamic prediction cannot be realized, and the influence of manual intervention is large due to the fact that the model parameters are set by experience; in the method using machine learning, time sequence information in data is not considered, or overall control is not carried out on various groups of people in the epidemic situation, so that comprehensive prediction on the trend situation of the epidemic situation is difficult.
Disclosure of Invention
Based on the problems, the invention provides the infectious disease trend prediction method based on the neural network and the SEIR model, the method can effectively and dynamically predict the infectious disease trend in real time by using less training data, and the prediction process does not need human intervention.
In order to solve the technical problems, the technical scheme adopted by the invention is as follows:
an infectious disease trend prediction method based on a neural network and an SEIR model comprises the following steps:
step 1, acquiring a data set;
step 2, preprocessing data;
step 3, constructing an epidemic situation trend prediction model consisting of a virus infection rate prediction module and an epidemic situation trend prediction module;
step 4, training the epidemic situation trend prediction model by using the data preprocessed in the step 2, and setting a loss function and a model parameter updating mode of the epidemic situation trend prediction model;
and 5, predicting the epidemic situation trend by using the epidemic situation trend prediction model trained in the steps 1 to 4.
Further, in the step 3, the preprocessed data is input into the virus infection rate prediction module and sequentially passes through an LSTM layer, a full link layer and a nonlinear transformation layer, the LSTM layer outputs the features containing the time sequence information, then the full link layer integrates the features, SoftPlus is used as an activation function, and finally the virus infection rates of the confirmed patients and the latent patients at a certain moment are output and predicted.
Further, the formula using SoftPlus as the activation function is:
Figure 982173DEST_PATH_IMAGE001
where exp is a non-linear function.
Further, in step 3, the formula for predicting the viral infection rates of the confirmed patients and the latent patients at a certain time is as follows:
Figure 768732DEST_PATH_IMAGE002
，
wherein,
Figure 828961DEST_PATH_IMAGE003
indicating the viral infection rate of the diagnosed patient,
Figure 943548DEST_PATH_IMAGE004
indicating the viral infection rate of the patients in the latent stage,
Figure 724422DEST_PATH_IMAGE005
to represent
Figure 176874DEST_PATH_IMAGE006
The virus infection rate of the patient is diagnosed at any moment,
Figure 893157DEST_PATH_IMAGE007
to represent
Figure 495040DEST_PATH_IMAGE006
Patients with latent period of timeThe rate of infection by the virus of (a),
Figure 204239DEST_PATH_IMAGE008
to represent
Figure 575177DEST_PATH_IMAGE006
The scaling of the infection rate of the virus at the moment,
Figure 462362DEST_PATH_IMAGE009
to represent
Figure 285961DEST_PATH_IMAGE006
The intensity of the epidemic situation prevention and control measures,
Figure 533272DEST_PATH_IMAGE010
to represent
Figure 758717DEST_PATH_IMAGE006
The viral infection rate of patients with time-latent phase is a multiple compared to the viral infection rate of patients with confirmed diagnosis.
Further, in the step 3, the output result of the virus infection rate prediction module is input into the epidemic situation trend prediction module, and the prediction of the epidemic situation trend is output through an SEIR model layer.
Further, in step 3, the prediction formula of the epidemic situation trend is as follows:
Figure 521530DEST_PATH_IMAGE011
，
wherein,
Figure 98005DEST_PATH_IMAGE012
indicates the number of people who are not infected,
Figure 634159DEST_PATH_IMAGE013
the number of patients in the latent period is shown,
Figure 714111DEST_PATH_IMAGE014
the number of patients who have been diagnosed is indicated,
Figure 457945DEST_PATH_IMAGE015
indicating the number of virus removers,
Figure 256136DEST_PATH_IMAGE016
is shown in
Figure 595982DEST_PATH_IMAGE006
The number of people not infected at that moment,
Figure 717391DEST_PATH_IMAGE017
is shown in
Figure 507492DEST_PATH_IMAGE006
The number of patients in the latent period at the moment,
Figure 402767DEST_PATH_IMAGE018
is shown in
Figure 936517DEST_PATH_IMAGE006
The number of patients can be diagnosed at any time,
Figure 853044DEST_PATH_IMAGE019
is shown in
Figure 876364DEST_PATH_IMAGE006
The number of virus removers at any time,
Figure 773782DEST_PATH_IMAGE020
indicating the viral infection rate of the diagnosed patient,
Figure 845643DEST_PATH_IMAGE021
indicating the viral infection rate of the patients in the latent stage,
Figure 630059DEST_PATH_IMAGE022
indicating the probability of a latent patient transforming into a diagnosed patient,
Figure 496384DEST_PATH_IMAGE023
means for diagnosing patientsAs a probability of the virus remover,
Figure 615519DEST_PATH_IMAGE024
、
Figure 491071DEST_PATH_IMAGE025
、
Figure 395573DEST_PATH_IMAGE026
、
Figure 432799DEST_PATH_IMAGE027
respectively represent
Figure 307739DEST_PATH_IMAGE028
Time of day
Figure 721403DEST_PATH_IMAGE012
、
Figure 480411DEST_PATH_IMAGE013
、
Figure 609910DEST_PATH_IMAGE014
、
Figure 579003DEST_PATH_IMAGE015
The predicted value of (a) is determined,
Figure 671724DEST_PATH_IMAGE029
and indicates the total population.
Further, in the step 4, the average square error is selected as a loss function during training.
Further, in step 4, the model parameter updating mode is set to be the learning rate of 0.005, a complete training set is a learning batch, the total number of learning iterations is 5000, the network training process updates parameters once for each learning batch, learning errors are calculated according to the output of the epidemic situation trend prediction module and the real label, the network parameters are updated by using the BP algorithm through errors, after each iteration learning is completed, the epidemic situation trend prediction module calculates the current prediction errors and compares the current prediction errors with the historical minimum errors, if the current prediction errors are smaller than the historical minimum errors, the current epidemic situation trend prediction module is saved, the historical minimum errors are updated to be the current errors, and then the training is continued until the total number of learning iterations is reached.
Compared with the prior art, the invention has the beneficial effects that:
1. the invention can realize automatic dynamic prediction on the infection rate of infectious diseases and the trend development of epidemic situation, thereby reducing manual operation and reducing the influence caused by manual intervention;
2. the method uses the time sequence information in the epidemic situation data, and combines the historical information and the current information to comprehensively predict;
3. and predicting various groups in the epidemic situation, and mastering the overall development trend of the epidemic situation.
Drawings
FIG. 1 is a flow chart of the present embodiment;
fig. 2 is a frame diagram of the epidemic situation trend prediction model in this embodiment.
Detailed Description
The invention will be further described with reference to the accompanying drawings. Embodiments of the present invention include, but are not limited to, the following examples.
Fig. 1 shows an infectious disease trend prediction method based on a neural network and an SEIR model, which includes the following steps:
step 1, acquiring a data set;
step 2, preprocessing data;
step 3, as shown in fig. 2, constructing an epidemic situation trend prediction model consisting of a virus infection rate prediction module and an epidemic situation trend prediction module;
step 4, training the epidemic situation trend prediction model by using the data preprocessed in the step 2, and setting a loss function and a model parameter updating mode of the epidemic situation trend prediction model;
and 5, predicting the epidemic situation trend by using the epidemic situation trend prediction model trained in the steps 1 to 4.
Further, in step 1, the data set includes suspected case data, confirmed case data, and removed case data.
Further, in step 2, the second step
Figure 409873DEST_PATH_IMAGE030
To
Figure 179115DEST_PATH_IMAGE030
+2 days of suspected case data, existing confirmed case data, removal case data as input
Figure 635504DEST_PATH_IMAGE031
Data sorting is carried out on daily suspected case data, existing confirmed case data and removed case data serving as corresponding labels to obtain 62 parts of data, and then the 62 parts of data are further sorted into 4 data sets according to the scale of a training set: data _10, Data _20, Data _30, Data _ 40; the Data _10 division rule is that the first 10 Data is a training set, the 11 th input Data is initial input Data of a test set, and the label of the last 52 Data is a label of the test set; the Data _20, Data _30 and Data _40 division rules are the same.
Further, the virus infection rate prediction module is constructed according to the following principle:
in this example, viral infection rates were established for confirmed patients and for patients in the latent phase
Figure 266336DEST_PATH_IMAGE032
And
Figure 858992DEST_PATH_IMAGE033
meanwhile, the change situation at the later stage of the virus infection rate is closely related to prevention and control measures and diagnosis and treatment measures at the early stage of the epidemic situation, and the prediction of the virus infection rate needs to be jointly analyzed according to historical information and current information, so that the prediction of the virus infection rate is a kind of time sequence prediction problem, the LSTM in the deep cycle neural network has excellent time sequence learning capability, can store information in past data and combine the information with information in the current data, and is based on the LSTM and the information in the past data
Figure 799135DEST_PATH_IMAGE032
And
Figure 477241DEST_PATH_IMAGE033
the mathematical model designs a virus infection rate prediction module;
in this embodiment, the virus infection rate prediction module includes an LSTM layer, a full link layer, and a nonlinear transformation layer, and after the preprocessed data is input to the virus infection rate prediction module, the LSTM layer outputs the features containing the timing information, and then the full link layer integrates the features, using SoftPlus as the activation function, with the formula as
Figure 911764DEST_PATH_IMAGE034
Where exp is a non-linear function. The nonlinear expression capability of the network is enhanced, and finally the virus infection rates of confirmed patients and latent patients at a certain moment are output and predicted;
in this embodiment, the formula for predicting the virus infection rates of confirmed patients and patients in the latent stage at a certain time is as follows:
Figure 624505DEST_PATH_IMAGE035
，
wherein,
Figure 726761DEST_PATH_IMAGE003
indicating the viral infection rate of the diagnosed patient,
Figure 157742DEST_PATH_IMAGE004
indicating the viral infection rate of the patients in the latent stage,
Figure 130377DEST_PATH_IMAGE005
to represent
Figure 697625DEST_PATH_IMAGE006
The virus infection rate of the patient is diagnosed at any moment,
Figure 979571DEST_PATH_IMAGE007
to represent
Figure 304373DEST_PATH_IMAGE006
The viral infection rate of patients in the time incubation period,
Figure 205333DEST_PATH_IMAGE008
to represent
Figure 17300DEST_PATH_IMAGE006
The scaling of the infection rate of the virus at the moment,
Figure 345513DEST_PATH_IMAGE009
to represent
Figure 360873DEST_PATH_IMAGE006
The intensity of the epidemic situation prevention and control measures,
Figure 799945DEST_PATH_IMAGE010
to represent
Figure 466418DEST_PATH_IMAGE006
The viral infection rate of patients with time-latent phase is a multiple compared to the viral infection rate of patients with confirmed diagnosis.
Further, the epidemic situation trend prediction module is constructed according to the following principle:
output result of virus infection rate prediction module
Figure 965533DEST_PATH_IMAGE032
And
Figure 468190DEST_PATH_IMAGE033
inputting the epidemic situation trend prediction module, and outputting the prediction of the epidemic situation trend through an SEIR model layer.
In this embodiment, the prediction formula of the epidemic situation trend is as follows:
Figure 710952DEST_PATH_IMAGE011
，
wherein,
Figure 234862DEST_PATH_IMAGE012
indicates the number of people who are not infected,
Figure 904878DEST_PATH_IMAGE013
the number of patients in the latent period is shown,
Figure 629251DEST_PATH_IMAGE014
the number of patients who have been diagnosed is indicated,
Figure 675704DEST_PATH_IMAGE015
indicating the number of virus removers,
Figure 316770DEST_PATH_IMAGE036
is shown in
Figure 157687DEST_PATH_IMAGE006
The number of people not infected at that moment,
Figure 369357DEST_PATH_IMAGE017
is shown in
Figure 953922DEST_PATH_IMAGE006
The number of patients in the latent period at the moment,
Figure 246232DEST_PATH_IMAGE037
is shown in
Figure 133417DEST_PATH_IMAGE006
The number of patients can be diagnosed at any time,
Figure 222595DEST_PATH_IMAGE019
is shown in
Figure 469906DEST_PATH_IMAGE006
The number of virus removers at any time,
Figure 960930DEST_PATH_IMAGE020
indicating the viral infection rate of the diagnosed patient,
Figure 19016DEST_PATH_IMAGE021
indicating the viral infection rate of the patients in the latent stage,
Figure 329912DEST_PATH_IMAGE022
indicating the probability of a latent patient transforming into a diagnosed patient,
Figure 118263DEST_PATH_IMAGE023
representing the probability of a diagnosed patient transforming into a virus remover,
Figure 463794DEST_PATH_IMAGE038
、
Figure 692781DEST_PATH_IMAGE025
、
Figure 490973DEST_PATH_IMAGE026
、
Figure 345665DEST_PATH_IMAGE027
respectively represent
Figure 545703DEST_PATH_IMAGE028
Time of day
Figure 945591DEST_PATH_IMAGE012
、
Figure 231079DEST_PATH_IMAGE013
、
Figure 686200DEST_PATH_IMAGE014
、
Figure 616110DEST_PATH_IMAGE015
The predicted value of (a) is determined,
Figure 311533DEST_PATH_IMAGE029
and indicates the total population.
Furthermore, in step 4, the loss function is used to measure the difference level between the predicted result and the actual situation, so that the selected result directly affects the model training effect,in this embodiment, the average square error is selected as a loss function during training, and the difficulty, accuracy and importance of various data collection in reality are considered, so that the pair
Figure 208951DEST_PATH_IMAGE013
、
Figure 280812DEST_PATH_IMAGE014
、
Figure 65229DEST_PATH_IMAGE015
And respectively calculating MSEs by the data, and weighting the three types of MSEs to obtain an overall learning error of the network, wherein the overall learning error of the network is as follows:
Figure 197133DEST_PATH_IMAGE039
，
wherein,
Figure 53618DEST_PATH_IMAGE040
、
Figure 929170DEST_PATH_IMAGE041
、
Figure 568093DEST_PATH_IMAGE042
are respectively composed of
Figure 870898DEST_PATH_IMAGE013
、
Figure 680591DEST_PATH_IMAGE014
、
Figure 94255DEST_PATH_IMAGE015
MSE calculated by data, and the calculation method of the MSE is as follows:
Figure 102531DEST_PATH_IMAGE043
，
wherein,
Figure 576238DEST_PATH_IMAGE044
is composed of
Figure 420697DEST_PATH_IMAGE006
The actual value of the time of day,
Figure 372473DEST_PATH_IMAGE045
is composed of
Figure 235255DEST_PATH_IMAGE006
The predicted value of the time of day,
Figure 879863DEST_PATH_IMAGE046
is the total time length.
Further, in step 4, network setup
Figure 211618DEST_PATH_IMAGE047
，
Figure 967085DEST_PATH_IMAGE048
The model parameter updating mode is set to be that the learning rate is 0.005, a complete training set is a learning batch, the total learning iteration frequency is 5000, the network training process carries out parameter updating once for each learning batch, learning errors are calculated according to the output of the epidemic situation trend prediction module and real labels, the network parameters are updated by using errors through a BP algorithm, after each iteration learning is completed, the epidemic situation trend prediction module calculates the current prediction errors and compares the current prediction errors with the historical minimum errors, if the current prediction errors are smaller than the historical minimum errors, the current epidemic situation trend prediction module is stored, the historical minimum errors are updated to be the current errors, and then the training is continued until the total learning iteration frequency is reached.
Further, in step 5, since the prediction is to predict the epidemic situation trend in a future period of time, there is no future epidemic situation data, so that only one set of initial data including three consecutive days is input to the epidemic situation trend prediction model in the prediction
Figure 761339DEST_PATH_IMAGE013
、
Figure 186636DEST_PATH_IMAGE014
、
Figure 130321DEST_PATH_IMAGE015
Data, one day for one prediction period, and epidemic trend prediction model for each set of input to predict the next day
Figure 814112DEST_PATH_IMAGE013
、
Figure 526853DEST_PATH_IMAGE014
、
Figure 388630DEST_PATH_IMAGE015
And combining the prediction result with the last two days of data in the input data to form new input data so as to continue predicting the epidemic situation of the last day until the total prediction period is reached.
The above is an embodiment of the present invention. The specific parameters in the above embodiments and examples are only for the purpose of clearly illustrating the invention verification process of the inventor and are not intended to limit the scope of the invention, which is defined by the claims, and all equivalent structural changes made by using the contents of the specification and the drawings of the present invention should be covered by the scope of the present invention.

Claims (8)

1. An infectious disease trend prediction method based on a neural network and an SEIR model is characterized by comprising the following steps:
step 1, acquiring a data set;
step 2, preprocessing data;
step 3, constructing an epidemic situation trend prediction model consisting of a virus infection rate prediction module and an epidemic situation trend prediction module;
step 4, training the epidemic situation trend prediction model by using the data preprocessed in the step 2, and setting a loss function and a model parameter updating mode of the epidemic situation trend prediction model;
and 5, predicting the epidemic situation trend by using the epidemic situation trend prediction model trained in the steps 1 to 4.
2. An infectious disease trend prediction method based on neural networks and SEIR models as claimed in claim 1, wherein: in the step 3, the preprocessed data is input into a virus infection rate prediction module and sequentially passes through an LSTM layer, a full connection layer and a nonlinear transformation layer, the LSTM layer outputs the characteristics containing the time sequence information, then the full connection layer integrates the characteristics, and finally the virus infection rates of the confirmed patients and the latent patients at a certain moment are output and predicted by using SoftPlus as an activation function.
3. An infectious disease trend prediction method based on neural networks and SEIR models as claimed in claim 2, wherein: the formula using SoftPlus as the activation function is:
Figure 480728DEST_PATH_IMAGE001
where exp is a non-linear function.
4. An infectious disease trend prediction method based on neural networks and SEIR models as claimed in claim 3, wherein: in step 3, the formula for predicting the virus infection rates of the confirmed patients and the latent patients at a certain time is as follows:
Figure 457912DEST_PATH_IMAGE002
，
wherein,
Figure 119837DEST_PATH_IMAGE003
indicating the viral infection rate of the diagnosed patient,
Figure 789853DEST_PATH_IMAGE004
indicating the viral infection rate of the patients in the latent stage,
Figure 655172DEST_PATH_IMAGE005
to represent
Figure 436046DEST_PATH_IMAGE006
The virus infection rate of the patient is diagnosed at any moment,
Figure 952478DEST_PATH_IMAGE007
to represent
Figure 793395DEST_PATH_IMAGE006
The viral infection rate of patients in the time incubation period,
Figure 146010DEST_PATH_IMAGE008
to represent
Figure 464996DEST_PATH_IMAGE006
The scaling of the infection rate of the virus at the moment,
Figure 835935DEST_PATH_IMAGE009
to represent
Figure 847753DEST_PATH_IMAGE006
The intensity of the epidemic situation prevention and control measures,
Figure 419155DEST_PATH_IMAGE011
to represent
Figure 807411DEST_PATH_IMAGE006
The viral infection rate of patients with time-latent phase is a multiple compared to the viral infection rate of patients with confirmed diagnosis.
5. An infectious disease trend prediction method based on neural networks and SEIR models as claimed in claim 2, wherein: and in the step 3, the output result of the virus infection rate prediction module is input into an epidemic situation trend prediction module and is output for predicting the epidemic situation trend through an SEIR model layer.
6. An infectious disease trend prediction method based on neural networks and SEIR models as claimed in claim 5, wherein: in the step 3, the prediction formula of the epidemic situation trend is as follows:
Figure 32856DEST_PATH_IMAGE012
，
wherein,
Figure 215576DEST_PATH_IMAGE014
indicates the number of people who are not infected,
Figure 542783DEST_PATH_IMAGE015
the number of patients in the latent period is shown,
Figure 469151DEST_PATH_IMAGE016
the number of patients who have been diagnosed is indicated,
Figure 549102DEST_PATH_IMAGE017
indicating the number of virus removers,
Figure 168302DEST_PATH_IMAGE019
is shown in
Figure 717227DEST_PATH_IMAGE006
The number of people not infected at that moment,
Figure 916127DEST_PATH_IMAGE020
is shown in
Figure 116164DEST_PATH_IMAGE006
The number of patients in the latent period at the moment,
Figure 906265DEST_PATH_IMAGE022
is shown in
Figure 926174DEST_PATH_IMAGE006
The number of patients can be diagnosed at any time,
Figure 210656DEST_PATH_IMAGE023
is shown in
Figure 265200DEST_PATH_IMAGE006
The number of virus removers at any time,
Figure 960623DEST_PATH_IMAGE024
indicating the viral infection rate of the diagnosed patient,
Figure 733407DEST_PATH_IMAGE025
indicating the viral infection rate of the patients in the latent stage,
Figure 553071DEST_PATH_IMAGE026
indicating the probability of a latent patient transforming into a diagnosed patient,
Figure 462121DEST_PATH_IMAGE027
representing the probability of a diagnosed patient transforming into a virus remover,
Figure 594025DEST_PATH_IMAGE029
、
Figure 322947DEST_PATH_IMAGE030
、
Figure 214811DEST_PATH_IMAGE031
、
Figure 243946DEST_PATH_IMAGE032
respectively represent
Figure 281173DEST_PATH_IMAGE033
Time of day
Figure 28549DEST_PATH_IMAGE014
、
Figure 927366DEST_PATH_IMAGE015
、
Figure 811008DEST_PATH_IMAGE016
、
Figure 19136DEST_PATH_IMAGE017
The predicted value of (a) is determined,
Figure 988229DEST_PATH_IMAGE034
and indicates the total population.
7. An infectious disease trend prediction method based on neural networks and SEIR models as claimed in claim 1, wherein: in step 4, the average square error is selected as a loss function during training.
8. An infectious disease trend prediction method based on neural networks and SEIR models as claimed in claim 1, wherein: in the step 4, the model parameter updating mode is set to be that the learning rate is 0.005, a complete training set is a learning batch, the total learning iteration frequency is 5000, the network training process updates parameters once for each learning batch, learning errors are calculated according to the output of the epidemic situation trend prediction module and the real label, the network parameters are updated by using BP algorithm with errors, after each iteration learning is completed, the epidemic situation trend prediction module calculates the current prediction errors and compares the current prediction errors with the historical minimum errors, if the current prediction errors are smaller than the historical minimum errors, the current epidemic situation trend prediction module is stored, the historical minimum errors are updated to be the current errors, and then the training is continued until the total learning iteration frequency is reached.
CN202110658591.1A 2021-06-15 2021-06-15 Infectious disease trend prediction method based on neural network and SEIR model Pending CN113113155A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202110658591.1A CN113113155A (en) 2021-06-15 2021-06-15 Infectious disease trend prediction method based on neural network and SEIR model

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202110658591.1A CN113113155A (en) 2021-06-15 2021-06-15 Infectious disease trend prediction method based on neural network and SEIR model

Publications (1)

Publication Number Publication Date
CN113113155A true CN113113155A (en) 2021-07-13

Family

ID=76723702

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202110658591.1A Pending CN113113155A (en) 2021-06-15 2021-06-15 Infectious disease trend prediction method based on neural network and SEIR model

Country Status (1)

Country Link
CN (1) CN113113155A (en)

Cited By (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112529329A (en) * 2020-12-21 2021-03-19 广东珠江智联信息科技股份有限公司 Infectious disease prediction method based on BP algorithm and SEIR model
CN113345599A (en) * 2021-08-04 2021-09-03 医渡云（北京）技术有限公司 Epidemic situation prediction method, epidemic situation prediction device, storage medium and electronic equipment
CN113539517A (en) * 2021-08-05 2021-10-22 浙江大学 Prediction method of time sequence intervention effect
CN113593716A (en) * 2021-07-30 2021-11-02 延安大学 Infectious disease trend prediction method and device
CN113744888A (en) * 2021-09-02 2021-12-03 深圳万海思数字医疗有限公司 Regional epidemic trend prediction early warning method and system
CN114496296A (en) * 2022-02-07 2022-05-13 华北理工大学 SEIARN propagation prediction system based on LSTM track crossing
CN117672544A (en) * 2023-12-14 2024-03-08 广东省疾病预防控制中心 Infectious disease transmission prediction method and system based on SEIR model
CN117690601A (en) * 2024-02-02 2024-03-12 江西省胸科医院(江西省第三人民医院) Tuberculosis epidemic trend prediction system based on big data analysis

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20170316324A1 (en) * 2016-04-27 2017-11-02 Virginia Polytechnic Institute And State University Computerized Event-Forecasting System and User Interface
CN111524611A (en) * 2020-04-24 2020-08-11 腾讯科技（深圳）有限公司 Method, device and equipment for constructing infectious disease trend prediction model
CN111739658A (en) * 2020-07-20 2020-10-02 医渡云（北京）技术有限公司 Method and device for predicting infectious disease trend based on input case
CN112163705A (en) * 2020-09-29 2021-01-01 重庆邮电大学 Epidemic situation prediction and early warning method based on CNN + SEIR and LSTM
CN112529329A (en) * 2020-12-21 2021-03-19 广东珠江智联信息科技股份有限公司 Infectious disease prediction method based on BP algorithm and SEIR model
CN112712903A (en) * 2021-01-15 2021-04-27 杭州中科先进技术研究院有限公司 Infectious disease monitoring method based on human-computer three-dimensional cooperative sensing

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20170316324A1 (en) * 2016-04-27 2017-11-02 Virginia Polytechnic Institute And State University Computerized Event-Forecasting System and User Interface
CN111524611A (en) * 2020-04-24 2020-08-11 腾讯科技（深圳）有限公司 Method, device and equipment for constructing infectious disease trend prediction model
CN111739658A (en) * 2020-07-20 2020-10-02 医渡云（北京）技术有限公司 Method and device for predicting infectious disease trend based on input case
CN112163705A (en) * 2020-09-29 2021-01-01 重庆邮电大学 Epidemic situation prediction and early warning method based on CNN + SEIR and LSTM
CN112529329A (en) * 2020-12-21 2021-03-19 广东珠江智联信息科技股份有限公司 Infectious disease prediction method based on BP algorithm and SEIR model
CN112712903A (en) * 2021-01-15 2021-04-27 杭州中科先进技术研究院有限公司 Infectious disease monitoring method based on human-computer three-dimensional cooperative sensing

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
甘雨 等: "新冠肺炎疫情趋势预测模型", 《智能系统学报》 *

Cited By (12)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112529329A (en) * 2020-12-21 2021-03-19 广东珠江智联信息科技股份有限公司 Infectious disease prediction method based on BP algorithm and SEIR model
CN113593716A (en) * 2021-07-30 2021-11-02 延安大学 Infectious disease trend prediction method and device
CN113593716B (en) * 2021-07-30 2024-02-09 延安大学 Infectious disease trend prediction method and device
CN113345599A (en) * 2021-08-04 2021-09-03 医渡云（北京）技术有限公司 Epidemic situation prediction method, epidemic situation prediction device, storage medium and electronic equipment
CN113539517A (en) * 2021-08-05 2021-10-22 浙江大学 Prediction method of time sequence intervention effect
CN113539517B (en) * 2021-08-05 2024-04-16 浙江大学 Method for predicting time sequence intervention effect
CN113744888A (en) * 2021-09-02 2021-12-03 深圳万海思数字医疗有限公司 Regional epidemic trend prediction early warning method and system
CN113744888B (en) * 2021-09-02 2023-09-22 深圳万海思数字医疗有限公司 Regional epidemic trend prediction and early warning method and system
CN114496296A (en) * 2022-02-07 2022-05-13 华北理工大学 SEIARN propagation prediction system based on LSTM track crossing
CN117672544A (en) * 2023-12-14 2024-03-08 广东省疾病预防控制中心 Infectious disease transmission prediction method and system based on SEIR model
CN117690601A (en) * 2024-02-02 2024-03-12 江西省胸科医院(江西省第三人民医院) Tuberculosis epidemic trend prediction system based on big data analysis
CN117690601B (en) * 2024-02-02 2024-05-24 江西省胸科医院(江西省第三人民医院) Tuberculosis epidemic trend prediction system based on big data analysis

Similar Documents

Publication Publication Date Title
CN113113155A (en) Infectious disease trend prediction method based on neural network and SEIR model
CN111524611A (en) Method, device and equipment for constructing infectious disease trend prediction model
CN111445997A (en) Infectious disease dynamics model based on machine learning algorithm and analysis method
CN102469103B (en) Trojan event prediction method based on BP (Back Propagation) neural network
CN102945310A (en) Epidemic propagation network modeling and inference of based on autonomic computing
CN110379522A (en) A kind of disease popularity trend predicting system and method
JPH0636061A (en) Learning system of hierarchic neural network
CN113035368A (en) Disease propagation prediction method based on differential migration diagram neural network
CN112116002A (en) Determination method, verification method and device of detection model
CN115640337B (en) Abnormal data prediction method, system and related equipment for symbolic graph network
CN114283320A (en) Target detection method based on full convolution and without branch structure
CN115271975A (en) Information processing method of medical insurance service
CN117292844A (en) Epidemic situation prediction method and system based on epidemic situation trend and management and control policy association model
CN110310199A (en) Borrow or lend money construction method, system and the debt-credit Risk Forecast Method of risk forecast model
CN113053536B (en) Infectious disease prediction method, system and medium based on hidden Markov model
CN110322342A (en) Borrow or lend money construction method, system and the debt-credit Risk Forecast Method of risk forecast model
CN112397202B (en) Prediction method for height growth curve of teenager and child
CN110555472A (en) electrocardiosignal data labeling method based on active learning
CN112185581A (en) New coronary pneumonia SIT model inversion method based on Python and BP neural network
Wen et al. Cascade SEIRD: Forecasting the Spread of COVID-19 with Dynamic Parameters Update
CN117198555B (en) Disease transmission prediction method based on viral load and neural network
Wu et al. Leveraging Neural Networks to Profile Health Care Providers with Application to Medicare Claims
CN118072976B (en) System and method for predicting respiratory tract diseases of children based on data analysis
CN115952838B (en) Self-adaptive learning recommendation system-based generation method and system
CN115302507B (en) Intelligent decision method for disassembly process of digital twin-driven industrial robot

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
RJ01 Rejection of invention patent application after publication
RJ01 Rejection of invention patent application after publication

Application publication date: 20210713