Epidemic disease early warning system and method based on college volunteer participation

Abstract

The invention discloses an epidemic disease early warning system and method based on college volunteer participation, and relates to the technical field of epidemic disease early warning. The epidemic disease early warning method based on college volunteer participation comprises the following steps: acquiring related data of a corresponding community based on college volunteers; solving an optimal solution of the speculation model by using a trust region algorithm; finishing feature processing, mapping a feature value into an output value, and obtaining a final prediction result of the corresponding community; and completing the precision evaluation of the corresponding community prediction result. According to the method, the early warning time is greatly shortened, independent early warning of each independent community in a regional range is completed in a point form, the early warning condition of each independent community in the region is combined, a prediction result is fused, an existing relatively mature neural network speculation model is combined, external data such as weather and medical treatment of the region are imported, and the early warning efficiency is improved. The operation of the direct data end is simple, and the subsequent speculation efficiency is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of epidemic warning, and specifically relates to an epidemic warning system and method based on the participation of college volunteers. Background Art

[0002] An epidemic refers to an infectious disease that can infect a large number of people. An epidemic can occur only in a certain region or can be a global pandemic. When the observed value of a certain disease exceeds the expected value, it is called an epidemic, and a global epidemic is defined as when the spread of a certain disease has crossed the boundaries between continents.

[0003] For example, the regional epidemic trend prediction and warning method and system disclosed in Chinese invention publication No. CN113744888B. Its warning method obtains the trend warning data of sample epidemics, processes the trend warning data of sample epidemics to obtain sample data, trains multiple sub-network models according to the sample data, and fuses the multiple sub-network models to obtain a sample epidemic trend prediction model. In this application, a distributed deep learning framework is adopted, and multiple sub-network models are fused. Each sub-network model is used to learn the relationship between different variables and the prediction results, and the potential connection between all variables, so as to improve the prediction accuracy of the model when the prediction step is large. In this application, the trend warning data of all epidemics in all regions is also obtained, and according to the trend warning data of all epidemics in all regions, transfer learning training is carried out on the sample epidemic trend prediction model to obtain the epidemic trend prediction models corresponding to all epidemics in all regions respectively. According to the epidemic trend prediction models corresponding to all epidemics in all regions respectively, the epidemic trends of all epidemics in all regions are predicted, and a visual prediction result is generated. In this application, long-term epidemic warning prediction services can be provided for all epidemics in all regions, and when there is a new epidemic, localized long-term epidemic warning prediction services can be quickly provided for the new epidemic through model transfer learning training. The trend warning data in this application at least includes: the search trend index of epidemic-related keywords, regional historical meteorological data, and regularly announced epidemic public monitoring data. The above data are all original data, which are accurate and easy to obtain, and do not involve user privacy data.

[0004] First of all, in the above technical solution, the direct source of its data is regional historical meteorological data and regularly announced epidemic public monitoring data. Since the public monitoring data has a wide range of sources, in the actual application process, a corresponding data review unit still needs to be set up to exclude error data that significantly exceeds the threshold range, resulting in excessive computing burdens both in the early stage of data processing and in the later stage of epidemic warning. The data processing range is too wide, resulting in its method being applicable to the overall warning judgment of a generalized large-scale region, causing unnecessary computing burdens on regional servers. Summary of the Invention

[0005] In view of the deficiencies of the prior art, the present invention provides an epidemic warning system and method based on the participation of college volunteers, which solves the problems raised in the above background art.

[0006] To achieve the above objectives, the present invention is implemented through the following technical solutions. An epidemic warning system and method based on the participation of college volunteers, the method includes the following specific steps:

[0007] S1: Based on the actual investigation of the surrounding communities by college volunteers, obtain the number of susceptibles S, the number of latent individuals E, the number of unisolated infected individuals Up, the number of hospitalized infected individuals Ip, the number of deaths due to illness F, and the number of recovered individuals R in the corresponding communities;

[0008] Furthermore, N is the total population of the corresponding community. Considering that the population flow in the community is relatively small in the short term, in all model predictions of this application, without considering immigration, emigration, natural mortality, and birth rate, according to the existing SEIR infectious disease model:

[0009] N = S + E + Up + Ip + R + F, this formula satisfies the short-term warning situation disclosed in this application;

[0010] S2: Based on the provided model variables, construct and calculate a speculation model through the least squares method, and use the trust region algorithm to solve the optimal solution of the speculation model;

[0011] S3: According to the epidemic-related data obtained in S2, input it into the neural network for missing value processing, feature augmentation, normalization, and time series dataset processing of the data, complete the preprocessing of the relevant data, complete the feature processing by CNN, and extract the time features by LSTM, map the feature values to output values, and obtain the final prediction results of the corresponding communities;

[0012] S4: Calculate the mean square error and mean absolute error between the predicted cumulative number of confirmed cases obtained in S3 and the actual cumulative number of confirmed cases submitted by the volunteers' actual investigations, so as to complete the accuracy evaluation of the prediction results of the corresponding communities. If the accuracy deviation is too large, it is fed back to the corresponding volunteers for re-investigation. If the accuracy deviation is within the threshold range, a warning alarm of the corresponding level is issued to the corresponding community.

[0013] A further improvement of the technical solution of the present invention is that the optimal solution of the speculation model in S2 includes: obtaining the daily effective contact number α1 of latent individuals E, the daily effective contact number α2 of unisolated infected individuals Up, the incidence rate β of latent individuals E, the isolation rate θ of infected individuals, the case fatality rate λ1 of unisolated infected individuals Up, the case fatality rate λ2 of hospitalized infected individuals Ip, and the daily cure rate μ of infected individuals.

[0014] A further improvement of the technical solution of the present invention lies in that in the above-mentioned S2, the speculation model, under the condition of not considering immigration, emigration, natural mortality rate and birth rate, combined with the existing SEIR infectious disease model, has: N = S + E + Up + Ip + R + F.

[0015] Specifically, according to the actual model and the predicted values of the model, a target function is constructed through the idea of the least squares method, and the trust region algorithm is used to solve the optimal solution of the speculation model, obtaining the required {α1, α2, β, λ1, λ2, μ}.

[0016] The above formula is the speculation model;

[0017] The basic assumptions of the existing SEIR model include: assuming that the total population of the region remains constant at N, not considering population dynamic factors, nor considering factors such as population flow; assuming that the change rate of the number of susceptible people over time is proportional to the ratio of the number of susceptible people and the number of infected people at that time; the probability of converting from the infected population to the susceptible population is proportional to the number of infected people at the same time point.

[0018] And this application aims to provide basic initial infection data based on the participation of efficient volunteers, so it conforms to the basic assumptions of the SEIR model. Further, this application quickly infers subsequent speculation data such as the daily effective contact number α2, the incidence rate β of the latent population E, etc. required from the number of susceptible people S, the number of latent people E, the number of non-isolated patients Up with onset, the number of in-hospital patients Ip with onset, the number of deaths F due to illness, and the number of cured people R, greatly shortening the warning time, completing individual warnings for each independent community in the region in the form of points, and combining the warning situations of each independent community in the region, integrating the prediction results, and combining the existing relatively mature neural network speculation model, importing external data such as the weather and medical conditions in the region, completing the simplification of the operation at the direct data end, and improving the subsequent speculation efficiency.

[0019] It should be noted that this method is based on the characteristics of the wide distribution and large base of college volunteers, and the existing urban communities in our country are clearly distributed, and all personnel in a single community can be quickly and comprehensively contacted by multiple volunteers.

[0020] A further improvement of the technical solution of the present invention lies in that the mean square error in the above-mentioned S4 includes:

[0021]

[0022] In the formula, y i is the actual cumulative number of confirmed cases on the day obtained based on the investigation of college volunteers, is the final predicted cumulative number of confirmed cases in the corresponding community in S3; according to whether the MSE exceeds the preset threshold, the judgment of whether the prediction result in S3 is accurate is completed.

[0023] A further improvement of the technical solution of the present invention lies in that the mean absolute error in S4 includes:

[0024]

[0025] In the formula, y i is the actual cumulative number of confirmed cases on the day obtained based on the survey of college volunteers, is the final predicted cumulative number of confirmed cases in the corresponding community in S3; according to whether the MAE exceeds the preset threshold, it is determined whether the prediction result in S3 is accurate. Both the MAE and the MSE are within the preset threshold, indicating that the prediction result and the survey result of the corresponding community are accurate.

[0026] Furthermore, in order to ensure the accuracy of the model calculation result in S2 and the neural network prediction result in S3, the mean square error and the mean absolute error are verified. In this process, the search trend index, regional historical meteorological data, and regularly announced epidemic public monitoring data introduced in the prior art are excluded, and the interference of external data is excluded, providing direct data proof for the final warning result of this warning method.

[0027] The present invention also provides an epidemic warning system based on the participation of college volunteers. The system includes:

[0028] A data acquisition module; multi-form network data input based on the survey results of college volunteers is stored in the database through structured processing;

[0029] A data processing module; based on the obtained epidemic-related data, the basic solution operation of the speculation model is completed to obtain relevant speculation data;

[0030] A data analysis module; based on a hybrid neural network of CNN and LSTM, eigenvalue extraction and eigenvalue mapping of relevant speculation data are completed to obtain the final prediction result.

[0031] A further improvement of the technical solution of the present invention lies in that in addition to providing data access to the database in the data acquisition module, it also includes: regularly crawling epidemic data through a crawler program and performing data cleaning in combination with the data processing module.

[0032] A further improvement of the technical solution of the present invention lies in that the system further includes: a visualization interface, which respectively displays the data conditions of each community in the form of line charts, bar charts, tables, and indicators, and displays different colors according to the relevant data conditions.

[0033] Beneficial effects

[0034] Compared with the prior art, the beneficial effects of the present invention are as follows: By quickly inferring subsequent speculation data such as the number of effective daily contacts α2, the incidence rate β of the latent population E, etc. required from the number of susceptible individuals S, the number of latent individuals E, the number of non-isolated individuals with onset Up, the number of hospitalized individuals with onset Ip, the number of deaths due to the disease F, and the number of cured individuals R, the warning time is significantly shortened. Separate warnings are completed for each independent community within the region in the form of points. Combining the warning situations of each independent community within the region, integrating the prediction results, and combining with the existing relatively mature neural network speculation model, importing external data such as the weather and medical conditions in the region, the operation at the direct data end is simplified, and the subsequent speculation efficiency is improved.

[0035] To ensure the accuracy of the model calculation results in S2 and the neural network prediction results in S3, mean square error and mean absolute error verification are performed, excluding the search trend index, regional historical meteorological data, and regularly announced epidemic public monitoring data introduced in the prior art, excluding the interference of external data, and providing direct data proof for the final warning result of this warning method. Brief Description of the Drawings

[0036] Figure 1 It is a schematic flowchart of an epidemic warning method based on the participation of college volunteers;

[0037] Figure 2 It is a schematic architecture diagram of an epidemic warning system based on the participation of college volunteers. Detailed Embodiments

[0038] Various exemplary embodiments, features, and aspects of the present application will be described in detail below with reference to the drawings. The same reference numerals in the drawings denote elements having the same or similar functions. Although various aspects of the embodiments are shown in the drawings, the drawings do not have to be drawn to scale unless otherwise specified.

[0039] The special term "exemplary" here means "serving as an example, embodiment, or illustration". Any embodiment described as "exemplary" here does not have to be construed as superior to or better than other embodiments.

[0040] In addition, to better illustrate the present application, numerous specific details are given in the following detailed embodiments. Those skilled in the art should understand that the present application can be implemented without some specific details. In some instances, methods, means, and elements well known to those skilled in the art are not described in detail to highlight the gist of the present application.

[0041] The present invention provides an epidemic warning method based on the participation of college volunteers, and the method includes the following specific steps:

[0042] S1: Based on the actual investigation of college volunteers in the surrounding communities, obtain the number of susceptibles S, the number of latent individuals E, the number of unisolated infected individuals Up, the number of hospitalized infected individuals Ip, the number of deaths due to illness F, and the number of recovered individuals R in the corresponding communities;

[0043] Furthermore, N is the total population of the corresponding community. Considering that the population flow in the community is relatively small in the short term, in all model predictions of this application, without considering immigration, emigration, natural mortality, and birth rates, according to the existing SEIR infectious disease model:

[0044] N = S + E + Up + Ip + R + F, and this equation satisfies the short-term warning situation disclosed in this application;

[0045] S2: Based on the provided model variables, construct and calculate a speculation model through the least squares method, and use the trust region algorithm to solve the optimal solution of the speculation model;

[0046] S3: According to the epidemic-related data obtained in S2, input it into the neural network for missing value processing, feature augmentation, normalization, and time series dataset processing of the data, complete the preprocessing of the relevant data, perform feature processing by CNN, and extract time features by LSTM. Map the feature values to output values to obtain the final prediction results of the corresponding communities;

[0047] S4: Calculate the mean square error and mean absolute error between the predicted cumulative number of confirmed cases obtained in S3 and the actual cumulative number of confirmed cases submitted by the volunteers' actual investigation, so as to complete the accuracy evaluation of the prediction results of the corresponding communities. If the accuracy deviation is too large, feedback to the corresponding volunteers for re-investigation. If the accuracy deviation is within the threshold range, issue a warning alert of the corresponding level to the corresponding community.

[0048] The optimal solution of the speculation model in S2 includes: obtaining the daily effective contact number α1 of latent individuals E, the daily effective contact number α2 of unisolated infected individuals Up, the incidence rate β of latent individuals E, the isolation rate θ of infected individuals, the case fatality rate λ1 of unisolated infected individuals Up, the case fatality rate λ2 of hospitalized infected individuals Ip, and the daily cure rate μ of infected individuals.

[0049] The speculation model in S2 satisfies N = S + E + Up + Ip + R + F in combination with the existing SEIR infectious disease model without considering immigration, emigration, natural mortality, and birth rates.

[0050] Furthermore, in this embodiment, under the condition of satisfying the above equation, define the differential equations of other types of populations as:

[0051]

[0052] Combined with the above formula, specifically, according to the actual model and the predicted values of the model, an objective function is constructed based on the idea of the least squares method, and the trust region algorithm is used to solve the optimal solution of the speculation model, obtaining the required {α1, α2, β, λ1, λ2, μ};

[0053] The basic assumptions of the existing SEIR model include: assuming that the total population of the region remains constant at N, without considering population dynamics factors or population mobility factors; assuming that the rate of change of the number of susceptible individuals over time is proportional to the ratio of the number of susceptible individuals and the number of infected individuals at that time; the probability of conversion from the infected population to the susceptible population is proportional to the number of infected individuals at the same time point;

[0054] This application aims to provide basic initial infection data based on the participation of efficient volunteers, so it conforms to the basic assumptions of the SEIR model. Furthermore, this application quickly infers subsequent speculation data such as the daily effective contact number α2 and the incidence rate β of the latent population E from the number of susceptible individuals S, the number of latent individuals E, the number of unisolated individuals with onset Up, the number of hospitalized individuals with onset Ip, the number of deaths due to illness F, and the number of cured individuals R, significantly shortening the warning time. Separate warnings are completed for each independent community in the region in the form of points, and combined with the warning situations of each independent community in the region, the prediction results are integrated. Combining with the existing relatively mature neural network speculation model, external data such as the weather and medical conditions in the region are imported to complete the simplification of the operation at the direct data end, improving the subsequent speculation efficiency.

[0055] It should be noted that this method is based on the characteristics of the wide distribution and large base of college volunteers, and the existing urban communities in our country are clearly distributed, and all personnel in a single community can be quickly and comprehensively contacted by multiple volunteers.

[0056] The mean square error in S4 includes:

[0057]

[0058] In the formula, y i is the actual cumulative number of confirmed cases on the day obtained from the college volunteer survey, is the final predicted cumulative number of confirmed cases in the corresponding community in S3; according to whether the MSE exceeds the preset threshold, it is determined whether the prediction result in S3 is accurate.

[0059] The mean absolute error in S4 includes:

[0060]

[0061] In the formula, y i is the actual cumulative number of confirmed cases on the day obtained from the college volunteer survey, is the cumulative number of confirmed cases finally predicted for the corresponding community in S3; according to whether the MAE exceeds the preset threshold, it is determined whether the prediction result in S3 is accurate. Both the MAE and the MSE are within the preset threshold, indicating that the prediction result and the survey result of the corresponding community are accurate.

[0062] Furthermore, in order to ensure the accuracy of the model calculation results in S2 and the neural network prediction results in S3, mean square error and mean absolute error verification are carried out. In this process, the search trend index, regional historical meteorological data, and regularly announced epidemic public monitoring data introduced in the prior art are excluded, and the interference of external data is excluded, providing direct data proof for the final warning result of this warning method.

[0063] The present invention also provides an epidemic warning system based on the participation of college volunteers, including:

[0064] A data acquisition module; multi-form network data input based on the survey results of college volunteers, and stored in the database through structured processing;

[0065] A data processing module; based on the obtained epidemic-related data, complete the basic solution operation of the speculation model to obtain relevant speculation data;

[0066] A data analysis module; based on a hybrid neural network of CNN and LSTM, complete the eigenvalue extraction and eigenvalue mapping of relevant speculation data to obtain the final prediction result.

[0067] In addition to providing data access to the database in the data acquisition module, it also includes: regularly crawling epidemic data through a crawler program and performing data cleaning in combination with the data processing module.

[0068] The system also includes: a visualization interface, which displays the data of each community in the form of line charts, bar charts, tables, and indicators respectively, and displays different colors according to the relevant data.

[0069] In specific implementation, the present application provides a computer storage medium and a corresponding data processing unit. Among them, the computer storage medium can store a computer program, and when the computer program is executed by the data processing unit, it can run the content of the present invention of an epidemic warning system and method based on the participation of college volunteers and some or all of the steps in each embodiment. The storage medium can be a magnetic disk, an optical disk, a read-only memory (ROM), or a random access memory (RAM), etc.

[0070] Those skilled in the art can clearly understand that the technical solutions in the embodiments of the present invention can be implemented by means of a computer program and its corresponding general hardware platform. Based on such an understanding, the technical solutions in the embodiments of the present invention, in essence or the part that contributes to the prior art, can be embodied in the form of a computer program, that is, a software product. This computer program software product can be stored in a storage medium and includes several instructions to enable a device including a data processing unit (which can be a personal computer, a server, a single-chip microcomputer, an MCU, or a network device, etc.) to execute the methods described in various embodiments or some parts of the embodiments of the present invention.

[0071] The present invention provides an epidemic warning system and method based on the participation of college volunteers. There are many methods and ways to specifically implement this technical solution. The above is only the preferred implementation manner of the present invention. It should be noted that for those of ordinary skill in the art, without departing from the principle of the present invention, several improvements and refinements can be made, and these improvements and refinements should also be regarded as the protection scope of the present invention. Each component not clearly defined in this embodiment can be implemented by the prior art.

Claims

1. An epidemic warning method based on the participation of college volunteers, characterized in that, The method includes the following specific steps: S1: Based on the actual investigation of college volunteers in the surrounding communities, obtain the number of susceptibles S, the number of latent individuals E, the number of unquarantined infected individuals Up, the number of hospitalized infected individuals Ip, the number of deaths due to illness F, and the number of recovered individuals R in the corresponding communities; S2: Based on the provided model variables, construct and calculate a prediction model through the least squares method, and use the trust region algorithm to solve the optimal solution of the prediction model; S3: According to the epidemic-related data obtained in S2, input it into the neural network for missing value processing, feature augmentation, normalization, and time series dataset processing of the data, complete the preprocessing of the relevant data, perform feature processing by CNN, and extract time features by LSTM, map the feature values to output values, and obtain the final prediction results for the corresponding communities; S4: Calculate the mean square error and mean absolute error between the predicted cumulative number of confirmed cases obtained in S3 and the actual cumulative number of confirmed cases submitted by the actual investigation of the volunteers, so as to complete the accuracy evaluation of the prediction results of the corresponding communities. If the accuracy deviation is too large, feedback to the corresponding volunteers for re-investigation. If the accuracy deviation is within the threshold range, issue a warning alarm of the corresponding level to the corresponding community.

2. The epidemic warning method based on the participation of college volunteers according to claim 1, characterized in that The optimal solution of the prediction model in S2 includes: obtaining the daily effective contact number α1 of latent individuals E, the daily effective contact number α2 of unquarantined infected individuals Up, the incidence rate β of latent individuals E, the isolation rate θ of infected individuals, the case fatality rate λ1 of unquarantined infected individuals Up, the case fatality rate λ2 of hospitalized infected individuals Ip, and the daily cure rate μ of infected individuals.

3. The epidemic warning method based on the participation of college volunteers according to claim 1 is characterized in that The prediction model in S2 satisfies the following equation in combination with the existing SEIR epidemic model without considering immigration, emigration, natural mortality, and birth rate: N = S + E + Up + Ip + R + F.

4. A method for epidemic warning based on the participation of college volunteers according to claim 1, characterized in that, The mean square error in S4 includes: where y i is the actual cumulative number of confirmed cases on the day obtained from the investigation of college volunteers, is the final predicted cumulative number of confirmed cases in the corresponding community in S3; according to whether the MSE exceeds the preset threshold, it is determined whether the prediction result in S3 is accurate.

5. The method for epidemic warning based on the participation of college volunteers according to claim 4, wherein, The mean absolute error in S4 includes: where y i is the actual cumulative number of confirmed cases on the day obtained from the survey of college volunteers, is the final predicted cumulative number of confirmed cases in the corresponding community in S3; according to whether the MAE exceeds the preset threshold, the accuracy of the prediction result in S3 is judged. Both the MAE and the MSE are within the preset threshold, indicating that the prediction result and the survey result of the corresponding community are accurate.

6. An epidemic warning system based on the participation of college volunteers, which is applied to the epidemic warning method based on the participation of college volunteers as described in claim 5, is characterized in that, The system includes: A data acquisition module; based on the multi-form network data input from the investigation results of college volunteers, it is stored in the database through structured processing; A data processing module; based on the obtained epidemic-related data, complete the solution operation of the prediction model and obtain relevant prediction data; A data analysis module; based on the hybrid neural network of CNN and LSTM, complete the extraction and mapping of feature values of the relevant prediction data to obtain the final prediction results.

7. The epidemic warning system based on the participation of college volunteers according to claim 6, wherein In addition to providing data access to the database in the data acquisition module, it also includes: regularly crawling epidemic data through a crawler program and performing data cleaning in combination with the data processing module.

8. The epidemic warning system based on the participation of college volunteers according to claim 6, characterized in that, The system also includes: a visualization interface, which displays the data of each community in the form of line charts, bar charts, tables, and indicators respectively, and shows different colors according to the relevant data.

Images