Dengue fashion trend prediction method and device, electronic equipment and computer storage medium

Abstract

The invention discloses a dengue fashion trend prediction method and device, electronic equipment and a computer storage medium. The dengue epidemic trend prediction method comprises the following steps: S1, collecting dengue historical case data, meteorological data, mosquito-borne monitoring data and social attention data; preprocessing the dengue history case data, the meteorological data, the mosquito-borne monitoring data and the social attention data; s2, inputting the meteorological data into an SARIMA model for prediction to obtain a predicted meteorological variable result; s3, combining the meteorological variable prediction result with mosquito-borne monitoring data and social attention data to construct a multiple regression model, training by using a least square method, estimating regression coefficients of respective variables, and minimizing prediction errors of the model; and S4, applying the trained multiple regression model to regional dengue prediction, and predicting the number of future dengue cases. The method can accurately predict the trend of the dengue fever.

Description

Technical Field

[0001] The present invention relates to the technical field of epidemiology, and more specifically, to a method, device, electronic device and computer storage medium for predicting dengue fever epidemic trends. Background Art

[0002] Dengue fever is an acute mosquito-borne infectious disease caused by the dengue virus, primarily transmitted by Aedes aegypti and Aedes albopictus mosquitoes. With the impact of global climate change and population migration, dengue fever outbreaks have shown a significant upward trend in tropical and subtropical regions in recent years. The spread of dengue fever is influenced by multiple factors, including climate, mosquito density, and socioeconomic conditions. In particular, climatic factors such as temperature, rainfall, and mosquito density are considered important drivers of dengue fever transmission. Therefore, predicting dengue fever epidemic trends based on these multiple variables is crucial for epidemic prevention and control.

[0003] Existing dengue prediction methods are mostly based on single-factor time series models, failing to fully account for the multifaceted and complex factors influencing dengue transmission, such as climate change, mosquito-borne disease surveillance, and public awareness. Traditional forecasting methods often fail to capture seasonal variations and long-term trends and perform poorly with nonlinear data. Furthermore, existing technologies lack accuracy in predicting short-term outbreaks, making it difficult to provide timely and reliable evidence for early intervention measures by public health authorities. Summary of the Invention

[0004] The present invention aims to provide a dengue fever epidemic trend prediction method, device, electronic device, and computer storage medium. The method can accurately predict dengue fever epidemic trends, effectively respond to short-term epidemic outbreaks, and provide a reliable scientific basis for public health departments to formulate intervention measures.

[0005] In order to solve the above technical problems, the technical solution adopted by the present invention is: In a first aspect, the present invention provides a method for predicting dengue fever epidemic trends, comprising the following steps: S1. Collect historical dengue fever case data, meteorological data, mosquito-borne surveillance data, and social attention data; pre-process the historical dengue fever case data, meteorological data, mosquito-borne surveillance data, and social attention data; S2. Input meteorological data into the SARIMA model for prediction to obtain the predicted meteorological variables. S3. Combine the predicted meteorological variables with mosquito surveillance data and social attention data to construct a multivariate regression model. Use the least squares method (OLS) to train the model, estimate the regression coefficients of each variable, and minimize the model's prediction error. S4. Apply the trained multivariate regression model to regional dengue forecasting to predict the number of future dengue cases.

[0006] The SARIMA model described in the present invention, whose full name is "Seasonal Autoregressive Integrated Moving Average Model", is a statistical model used to process data with seasonal characteristics in time series analysis.

[0007] Optionally, after step S4, the method further includes: S5. comparing the trained multivariate regression model with actual historical dengue fever case data and adjusting the trained multivariate regression model based on the validation results. The accuracy and stability of the model are verified by comparing the trained model with actual historical data. The parameters of the multivariate regression model are adjusted based on the validation results to improve prediction accuracy.

[0008] Optionally, the meteorological data of the present invention includes temperature, precipitation, and sunshine time.

[0009] Optionally, the mosquito-borne disease monitoring data is the mosquito Brayer coefficient.

[0010] Optionally, the social attention data is a Baidu index related to dengue fever.

[0011] The Baidu Index reflects public interest in and search behavior for dengue-related information, which, to a certain extent, reflects the changing trends in disease transmission and public panic. Research has found that people use search engines to search for information on symptoms, preventive measures, and other related information during the early stages of an outbreak, making search interest a potential early warning indicator. Therefore, incorporating the Baidu Index into multivariate regression models can help capture the dynamics of public perception of dengue risk, complementing the shortcomings of traditional meteorological and mosquito-borne surveillance data in reflecting social, psychological, and behavioral aspects, thereby improving the overall accuracy of dengue epidemic forecasts.

[0012] Basic data cleaning was performed on the historical dengue fever case data, meteorological data, mosquito-borne surveillance data, and social awareness data to ensure data quality. Furthermore, the historical dengue fever case data, meteorological data, mosquito-borne surveillance data, and social awareness data were standardized, particularly the meteorological data and historical dengue fever case data, to eliminate the impact of unit differences and facilitate subsequent modeling and analysis.

[0013] In step S1, the preprocessing includes data cleaning and standardization.

[0014] Optionally, the data cleaning includes processing missing values and removing outliers.

[0015] As one implementation manner, the method adopted for the normalization process is minimum-maximum normalization.

[0016] The dengue fever epidemic trend prediction method described in the present invention can relatively accurately predict dengue fever epidemic trends and effectively respond to short-term outbreaks. This is mainly reflected in its sensitive capture of real-time and short-term key influencing factors. First, using the SARIMA model to predict meteorological data can relatively accurately capture short-term fluctuations in meteorological factors such as temperature and precipitation. These factors have a direct impact on mosquito-borne activity and virus transmission. Second, by introducing mosquito-borne monitoring data and social attention data (such as the Baidu Index), the model can promptly reflect changes in public concern about dengue fever risks and abnormal fluctuations in mosquito-borne density, which are important early warning indicators for epidemic outbreaks. Finally, by integrating data from these different sources, the multivariate regression model uses the least squares method to quickly estimate the weights of each variable and adjust the prediction results in real time, thereby identifying potential outbreak trends in the early stages of the epidemic and effectively responding to short-term outbreaks.

[0017] In a second aspect, the present invention provides a device for predicting dengue fever epidemic trends. The device comprises: Input unit, used to obtain dengue fever historical case data, meteorological data, mosquito-borne surveillance data and social attention data; A SARIMA model processing unit is used to process the meteorological data to obtain predicted meteorological variable results; A multiple regression model processing unit, configured to combine the predicted meteorological variable results with mosquito-borne disease monitoring data and social attention data to construct a multiple regression model; The output unit is used to output the predicted dengue fever epidemic trend.

[0018] In a third aspect, the present invention provides an electronic device. The electronic device includes at least one memory and at least one processor; The at least one memory is coupled to the at least one processor, the at least one memory is used to store a computer program, the at least one processor is used to call the computer program, and the computer program includes instructions. When the instructions are executed by the at least one processor, the electronic device executes the dengue fever epidemic trend prediction method as described in the first aspect above.

[0019] In a fourth aspect, the present invention provides a computer storage medium comprising computer instructions, which, when executed on an electronic device, cause the electronic device to execute the dengue fever epidemic trend prediction method described in the first aspect.

[0020] Compared with the prior art, the present invention has the following beneficial effects: The dengue fever epidemic trend prediction method described in the present invention comprehensively considers multiple factors that affect the spread of dengue fever, such as meteorology, vector monitoring, and social attention, uses the SARIMA model to process periodic data, and combines it with a multivariate regression model for accurate prediction. This not only significantly improves the accuracy of dengue fever epidemic trend prediction, but also can effectively respond to short-term outbreaks, providing a reliable scientific basis to help public health departments formulate prevention and control strategies in advance and reduce the negative impact of the epidemic on society and the economy. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] Figure 1 This is a process flow chart of the dengue fever epidemic trend prediction method described in the embodiment.

[0022] Figure 2 This is another processing flow chart of the dengue fever epidemic trend prediction method described in the embodiment.

[0023] Figure 3 This is a functional module block diagram of the device for predicting dengue fever epidemic trends according to the present invention.

[0024] Figure 4 2 is a structural block diagram of the electronic device described in the embodiment. DETAILED DESCRIPTION

[0025] The following will be combined with the drawings in the embodiments of this application to clearly and completely describe the technical solutions in the embodiments of this application. Obviously, the embodiments described are only part of the embodiments of this application, not all of the embodiments. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of this application.

[0026] The following describes a dengue fever epidemic trend prediction method according to an embodiment of the present invention.

[0027] Specifically, such as Figure 1 As shown, the dengue fever epidemic trend prediction method includes the following steps: S1. Collect historical dengue fever case data, meteorological data, mosquito-borne surveillance data, and social attention data; and preprocess the historical dengue fever case data, meteorological data, mosquito-borne surveillance data, and social attention data.

[0028] Meteorological data includes temperature, precipitation, and sunshine duration. Mosquito-borne disease surveillance data includes the Braier coefficient. Public awareness data includes the Baidu Index for dengue fever. Historical dengue fever case data, meteorological data (temperature, precipitation, sunshine duration), and mosquito-borne disease surveillance data (Brell coefficient) are collected through the Centers for Disease Control and Prevention and meteorological departments, respectively.

[0029] The preprocessing includes data cleaning and standardization. Data cleaning includes processing missing values and removing outliers.

[0030] Basic data cleaning was performed on the historical dengue fever case data, meteorological data, mosquito-borne surveillance data, and social awareness data to ensure data quality. Furthermore, the historical dengue fever case data, meteorological data, mosquito-borne surveillance data, and social awareness data were standardized, particularly the meteorological data and historical dengue fever case data, to eliminate the impact of unit differences and facilitate subsequent modeling and analysis.

[0031] For historical dengue fever case data, the original data table generally contains two columns: "Date" and "Number of Cases", for example: 2021-07-01, number of cases: 15; 2021-07-02, number of cases: —— (missing); 2021-07-03, number of cases: 18; 2021-07-04, number of cases: 120 (obviously abnormal, significantly higher than the previous and subsequent data); First, address missing values. For the missing data on July 2, 2021, linear interpolation can be used to fill the missing value by taking the mean of the number of cases on the preceding and following dates: (15 + 18) / 2 = 16.5. Second, address outliers. Since the number of cases on July 4, 2021, 120, is significantly higher than the adjacent data (the normal range is around 10-20), a box plot or Z-score method can be used to identify outliers. These values can then be removed or corrected using the median of the preceding and following dates (for example, replacing 120 with the median of 17).

[0032] Similarly, for meteorological data (including temperature, precipitation, and sunshine time), assuming that the temperature data for a certain day is: [30℃, 31℃, ——, 29℃], the missing value can be supplemented by the average value of adjacent days; if the precipitation record for a certain day is 500 mm, and the values before and after are both within the range of 10 to 50 mm, then 500 mm can be regarded as an entry error and needs to be eliminated or corrected to a reasonable value (for example, taking the average value before and after as 30 mm).

[0033] A similar approach is used for mosquito-borne surveillance data (mosquito Brayer coefficient) and social attention data (Baidu Index): missing values are filled with the mean or median, and data that deviates extremely from the normal range (for example, the Baidu Index suddenly soars to 5,000 on a certain day, while the normal range is between 200 and 600) is eliminated or replaced.

[0034] The purpose of normalization is to eliminate differences in dimensions and ranges between different data. The method used in this paper is minimum-maximum normalization. Taking minimum-maximum normalization as an example, assuming a range of dengue fever case data is [15, 16.5, 18, 17, 16], where the minimum is 15 and the maximum is 18, the conversion formula is: (x – 15) / (18 – 15). The converted data is: [0, 0.5, 1, 0.67, 0.33]. Similarly, for meteorological data, such as temperature data [29°C, 30°C, 31°C, 29.5°C], the minimum value of 29°C and the maximum value of 31°C can be substituted into the formula for normalization.

[0035] For mosquito-borne surveillance data and social attention data, the minimum and maximum values can be calculated separately and then normalized so that all data are within the range of 0 to 1, which is convenient for subsequent modeling.

[0036] S2. Input meteorological data into the SARIMA model for prediction to obtain the predicted meteorological variable results.

[0037] Suppose a daily temperature series (for example, [28°C, 29°C, 30°C, 29.5°C, …]) is extracted from meteorological data of the past three years. First, a seasonality test is performed on the series, and it is found that there is an obvious annual periodicity.

[0038] Use the SARIMA model, set the model parameters, and train the model to predict the temperature for the next 30 days.

[0039] The SARIMA model is expressed as SARIMA(p, d, q)(P, D, Q)_s, where p, d, and q represent the non-seasonal autoregressive (AR) order, differencing order, and moving average (MA) order, respectively. P, D, and Q represent the seasonal autoregressive, differencing, and moving average orders, and s represents the seasonal period (for example, if monthly data is used, s=12). Its mathematical expression can be written as: Φ P (B s ) φ p (B) (1 – B) d (1 – B s ) D X t = Θ Q (B s ) θ_q(B) ε t Among them, φ p (B) = 1 – φ1B – … – φ p B p is a non-seasonal autoregressive polynomial, θ q(B) = 1 +θ1B + … + θ q B q is a nonseasonal moving average polynomial; Φ P (B s ) and Θ Q (B s ) are the autoregressive and moving average polynomials of the seasonal part, respectively; (1– B) d and (1–B s ) D Indicates that the data is differentiated d times and seasonally D times to make the series stable; ε t is a white noise sequence.

[0040] In this example, the parameters are set for the nonseasonal order (1,1,1) (i.e., p=1, d=1, q=1) and the seasonal order (1,0,0) or (0,0,0). The example uses a seasonal order of (1,0,0) and a period s of 12, where Φ1 = the parameter and D = 0. During training, these parameters are estimated using historical daily temperature series. The resulting model can then be used to iteratively predict the temperature for the next 30 days. This involves calculating the residual and prediction error at the current moment using historical data. The model expression is then recursively used to predict future moments, yielding predicted values for the future temperature. This approach not only accounts for short-term temperature fluctuations but also incorporates annual cyclical characteristics into the forecast, thereby improving prediction accuracy.

[0041] For example, the model predicts that the temperature for the next week will be [29.0℃, 29.2℃, 29.1℃, 29.3℃, 29.2℃, 29.0℃, 29.1℃]. These predicted values will be used as an input variable in the subsequent multiple regression model.

[0042] S3. Combine the predicted meteorological variables with mosquito-borne disease surveillance data and social attention data to construct a multivariate regression model. Use the least squares method (OLS) for training, estimate the regression coefficients of each variable, and minimize the model's prediction error.

[0043] In step S3, the predicted meteorological variables (e.g., predicted temperature), the mosquito-borne disease surveillance data collected simultaneously (assuming the Brayer coefficient is 2.5 on a certain day), and the social attention data (e.g., Baidu Index is 350) are combined as independent variables. The constructed multiple regression model can be expressed as follows: Number of dengue cases = β0 + β1 × predicted temperature + β2 × Brayer coefficient + β3 × Baidu index + ε For example, using historical data (such as records from the past year), the coefficients estimated by the least squares method (OLS) are: β0 = -50, β1 = 2.0, β2 = 10, β3 = 0.05.

[0044] The multiple regression model is now: Predicted number of cases = -50 + 2.0 × temperature + 10 × Brayer coefficient + 0.05 × Baidu index Assuming that the predicted temperature for a certain day is 29.2℃, the Brayer coefficient is 2.5, and the Baidu Index is 350, the predicted number of cases is: -50 + 2.0×29.2 + 10×2.5 + 0.05×350 = -50 + 58.4 + 25 + 17.5 = 50.9.

[0045] After OLS training, the model parameters are determined and the goal is to minimize the prediction error.

[0046] S4. Apply the trained multivariate regression model to regional dengue forecasting to predict the number of future dengue cases.

[0047] In step S4, the trained regression model is used to predict the number of dengue fever cases in a certain area in the future. For example, if the input variables for predicting a certain day in the future are: predicted temperature 29.0℃, Brayer coefficient 2.3, and Baidu index 320, then the model is substituted to calculate: Predicted number of cases = -50 + 2.0×29.0 + 10×2.3 + 0.05×320 = -50 + 58 + 23 +16 = 47.

[0048] Therefore, the number of dengue fever cases on that day was predicted to be approximately 47.

[0049] Further, if Figure 2 As shown, the dengue fever epidemic trend prediction method further includes: S5. comparing the trained multivariate regression model with actual historical dengue fever case data and adjusting the trained multivariate regression model based on the verification results. The accuracy and stability of the model are verified by comparing the model with actual historical data. The parameters of the multivariate regression model are adjusted based on the verification results to improve prediction accuracy.

[0050] In step S5, the prediction results are compared with the actual historical case data collected. For example, if the actual number of cases recorded on a certain day is 55, but the model predicts 47, there is an error. By calculating the root mean square error (RMSE) between the predicted and actual number of cases over a period of time (for example, one month), the error distribution is analyzed. If it is found that the resulting multivariate regression model systematically underestimates or overestimates the number of cases, it may be necessary to review the regression variables or retrain the model parameters. Adjustments can include introducing new variables, correcting the weights of existing variables, or using weighted least squares methods to retrain the model until the prediction error falls within the expected range.

[0051] For example, after adjustment, the re-estimated coefficients may become: β0 = -45, β1 = 1.8, β2 = 11, β3 = 0.06, thereby improving the overall fit of the model. The modified multiple regression model is: Predicted number of cases = -45 + 1.8 × temperature + 11 × Brayer coefficient + 0.06 × Baidu index. After recalculation, the predicted number of cases is 52.

[0052] This embodiment also provides a device for predicting dengue fever epidemic trends.

[0053] The following describes an apparatus for executing the above-mentioned dengue fever epidemic trend prediction method.

[0054] Figure 3 A functional module block diagram of the device provided in this embodiment is provided. The device for predicting the epidemic trend of dengue fever includes: Input unit 10, for obtaining dengue fever historical case data, meteorological data, mosquito-borne surveillance data and social attention data; A SARIMA model processing unit 20 is used to process the meteorological data to obtain a predicted meteorological variable result; A multiple regression model processing unit 30 is used to combine the predicted meteorological variable results with mosquito-borne monitoring data and social attention data to construct a multiple regression model; The output unit 40 is used to output the predicted dengue fever epidemic trend.

[0055] The device provided in this embodiment and the dengue fever epidemic trend prediction method provided in this application have the same concept. The specific implementation process is detailed in the full text of the specification and will not be repeated here.

[0056] This embodiment also provides an electronic device. The electronic device capable of executing the above-mentioned dengue fever epidemic trend prediction method is described below.

[0057] Figure 4A structural block diagram of an electronic device 100 for implementing a dengue fever epidemic trend prediction method is shown. The electronic device 100 includes: at least one memory 101 and at least one processor 102; the at least one memory 101 is coupled to the at least one processor 102, the at least one memory 101 is used to store a computer program, and the at least one processor 102 is used to call the computer program. The computer program includes instructions. When the instructions are executed by the at least one processor, the electronic device executes the above-mentioned dengue fever epidemic trend prediction method.

[0058] This embodiment also provides a computer storage medium. The following describes the computer storage medium containing the dengue fever epidemic trend prediction method.

[0059] A computer storage medium includes computer instructions. When the computer instructions are executed on an electronic device, the electronic device executes the above-mentioned dengue fever epidemic trend prediction method.

[0060] Obviously, the above embodiments of the present invention are merely examples for the purpose of clearly illustrating the present invention, and are not intended to limit the embodiments of the present invention. Those skilled in the art will appreciate that other variations or modifications can be made based on the above description. It is not necessary and impossible to enumerate all embodiments here. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention shall be included within the scope of protection of the claims of the present invention.

Claims

1. A method for predicting dengue fever epidemic trends, characterized in that: The following steps are involved: S1. Collect historical dengue fever case data, meteorological data, mosquito-borne surveillance data, and social attention data; pre-process the historical dengue fever case data, meteorological data, mosquito-borne surveillance data, and social attention data; S2. Input meteorological data into the SARIMA model for prediction to obtain the predicted meteorological variables. S3. Combine the predicted meteorological variables with mosquito surveillance data and social attention data to construct a multivariate regression model. Use the least squares method (OLS) to train the model, estimate the regression coefficients of each variable, and minimize the model's prediction error. S4. Apply the trained multivariate regression model to regional dengue forecasting to predict the number of future dengue cases.

2. The dengue fever epidemic trend prediction method according to claim 1, characterized in that: After step S4, the method further includes: S5. comparing with actual dengue fever historical case data, and adjusting the trained multivariate regression model according to the verification result.

3. The dengue fever epidemic trend prediction method according to claim 1, characterized in that: The meteorological data includes temperature, precipitation, and sunshine time.

4. The dengue fever epidemic trend prediction method according to claim 1, characterized in that: The mosquito-borne monitoring data is the mosquito Brayer coefficient.

5. The dengue fever epidemic trend prediction method according to claim 1, characterized in that: The social attention data is the Baidu index related to dengue fever.

6. The dengue fever epidemic trend prediction method according to claim 1, characterized in that: In step S1, the preprocessing includes data cleaning and standardization.

7. The dengue fever epidemic trend prediction method according to claim 6, characterized in that: The data cleaning includes processing missing values and removing outliers.

8. A device for predicting dengue fever epidemic trends, characterized in that: include: Input unit, used to obtain dengue fever historical case data, meteorological data, mosquito-borne surveillance data and social attention data; A SARIMA model processing unit is used to process the meteorological data to obtain predicted meteorological variable results; A multiple regression model processing unit, configured to combine the predicted meteorological variable results with mosquito-borne disease monitoring data and social attention data to construct a multiple regression model; The output unit is used to output the predicted dengue fever epidemic trend.

9. An electronic device, characterized in that: comprising at least one memory and at least one processor; The at least one memory is coupled to the at least one processor, the at least one memory is used to store a computer program, the at least one processor is used to call the computer program, and the computer program includes instructions. When the instructions are executed by the at least one processor, the electronic device executes the dengue fever epidemic trend prediction method as described in any one of claims 1 to 7.

10. A computer storage medium, characterized in that The method comprises computer instructions, which, when executed on an electronic device, enable the electronic device to execute the dengue fever epidemic trend prediction method according to any one of claims 1 to 7.

Images