Mosquito outbreak risk prediction method, device, electronic device and storage medium

By constructing a prediction model of mosquito density and influencing factors, performing spatial downscale processing and sub-grid division, the problem of inaccurate prediction of mosquito outbreak risk in the existing technology is solved, and accurate prediction of mosquito-borne disease transmission trend is achieved, and mosquito-borne disease prevention is prevented. Control provides accurate solutions.

CN118136276BActive Publication Date: 2024-11-05INST OF GEOGRAPHICAL SCI & NATURAL RESOURCE RES CAS
View PDF 8 Cites 0 Cited by

Patent Information

Application Number
CN202410311621.5
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-03-19
Publication Date
2024-11-05
Estimated Expiration
2044-03-19

AI Technical Summary

Technical Problem

The existing technology is difficult to express the spatial differences within streets and townships in mosquito density prediction, resulting in inaccurate prediction of mosquito outbreak risk and inability to provide effective information for formulating accurate prevention and control plans.

Method used

By determining the mosquito density data and potential influencing factors in the target area, a prediction model is constructed, spatial downscale processing is performed, molecular grids are drawn, mosquito outbreak risk in each sub-grid, and mosquito-borne disease transmission risk area is determined.

Benefits of technology

It improves the accuracy of predicting the risk of mosquito outbreaks, accurately predicts the transmission trend of mosquito-borne diseases, and provides spatial information reference for the formulation of precise mosquito prevention and control plans.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118136276B_ABST
    Figure CN118136276B_ABST
Patent Text Reader

Abstract

The present application discloses a method, device, electronic device and storage medium for predicting the risk of mosquito outbreaks, and relates to the field of computer technology. Among them, the method includes: determining the density binary data corresponding to the mosquito density data in the target area, and determining multiple potential influencing factors affecting the mosquito density data in the target area; constructing a target prediction model for predicting the risk of mosquito outbreaks based on the density binary data and the attribute data of each potential influencing factor; performing spatial downscaling processing on the area to be tested to obtain multiple sub-grids, and predicting the predicted probability of mosquito outbreaks in each sub-grid based on the target prediction model; determining the transmission risk area of ​​mosquito-borne diseases in the area to be tested based on the predicted probability corresponding to each sub-grid. The technical solution provided by the present application can improve the prediction accuracy of the risk of mosquito outbreaks, accurately predict the transmission trend of mosquito-borne diseases, and provide spatial information reference for formulating precise mosquito prevention and control plans.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present application relates to the field of computer technology, and in particular to a method, device, electronic device and storage medium for predicting mosquito outbreak risk. Background Art

[0002] Mosquitoes are important biological vectors of some mosquito-borne diseases (such as dengue fever). The risk prediction of mosquito outbreaks is an important reference for determining the risk of mosquito-borne disease transmission in the study area. Based on the factors affecting mosquito density, the researchers attempted to achieve mosquito outbreak risk prediction and assessment by simulating the factors affecting mosquito density.

[0003] The spatial analysis scale of existing technologies is mainly to predict and evaluate the local mosquito density level from the dimensions of larger-scale areas such as towns and streets. Since mosquitoes themselves have a short flight distance and do not have the ability to migrate long distances, and their population density is affected to varying degrees by natural and socio-economic factors at different times and places, there is obvious heterogeneity in spatial distribution. Therefore, the mosquito density prediction results based on streets and towns units are difficult to express the spatial differences in mosquito outbreak risks within streets and towns, and thus cannot provide spatial information reference for the formulation of spatially accurate mosquito prevention and control plans. Therefore, how to improve the prediction accuracy of mosquito outbreak risks and improve the accuracy of predicting the spread of mosquito-borne diseases has become an urgent problem to be solved. Summary of the invention

[0004] The present application provides a method, device, electronic device and storage medium for predicting mosquito outbreak risk, which can improve the prediction accuracy of mosquito outbreak risk, accurately predict the spread of mosquito-borne diseases, and provide spatial information reference for formulating precise mosquito prevention and control plans.

[0005] In a first aspect, the present application provides a method for predicting mosquito outbreak risk, the method comprising:

[0006] Determine density binary data corresponding to mosquito density data in a target area, and determine a plurality of potential influencing factors affecting the mosquito density data in the target area;

[0007] Constructing a target prediction model for predicting mosquito outbreak risk based on the density binary data and the attribute data of each potential influencing factor;

[0008] Performing spatial downscaling processing on the area to be tested to obtain multiple sub-grids, and predicting the predicted probability of mosquito outbreak in each sub-grid based on the target prediction model;

[0009] The transmission risk area of ​​the mosquito-borne disease in the area to be tested is determined based on the predicted probability corresponding to each sub-grid.

[0010] Furthermore, the target prediction model for predicting the risk of mosquito outbreaks is constructed based on the density binary data and the attribute data of each potential influencing factor, including: screening out at least one target influencing factor from the multiple potential influencing factors based on the density binary data and the attribute data of each potential influencing factor; and constructing the target prediction model for predicting the risk of mosquito outbreaks based on the density binary data and the attribute data of the target influencing factor.

[0011] Furthermore, the method of selecting at least one target influencing factor from the multiple potential influencing factors based on the density binary data and the attribute data of each potential influencing factor includes: constructing a corresponding univariate prediction model for each potential influencing factor based on the density binary data and the attribute data of each potential influencing factor; determining the prediction accuracy of each univariate prediction model; determining a correlation matrix of the multiple potential influencing factors; and selecting at least one target influencing factor from the multiple potential influencing factors based on the correlation matrix and the prediction accuracy.

[0012] Furthermore, the target prediction model for predicting the risk of mosquito outbreaks is constructed based on the density binary data and the attribute data of the target influencing factors, including: performing at least two resampling processes on the density binary data to obtain corresponding at least two types of sample data sets; constructing multivariate prediction models based on the attribute data of the target influencing factors and each type of sample data set, thereby obtaining at least two multivariate prediction models; determining the prediction performance data of the at least two multivariate prediction models; and screening out the target prediction model from the at least two multivariate prediction models based on the prediction performance data.

[0013] Furthermore, the method of selecting at least one target influencing factor from the multiple potential influencing factors based on the correlation matrix and the prediction accuracy includes: for the current potential influencing factor, determining a preliminary influencing factor having a high correlation with the current potential influencing factor based on the correlation matrix, the preliminary influencing factor being at least one potential influencing factor among the multiple potential influencing factors; selecting candidate influencing factors that meet a preset accuracy standard from the preliminary influencing factors based on the prediction accuracy; traversing the multiple potential influencing factors in turn to obtain multiple candidate influencing factors, and determining the multiple candidate influencing factors as the target influencing factors.

[0014] Furthermore, the method of selecting the target prediction model from the at least two multivariate prediction models based on the prediction performance data includes: determining the recall and specificity of each of the at least two multivariate prediction models for predicting the risk of mosquito outbreaks; selecting the first multivariate prediction model with the specificity greater than a preset threshold and the largest recall from the at least two multivariate prediction models; verifying the first multivariate prediction model using a verification data set to obtain a verification result, wherein the verification data set is a monitoring data set for monitoring the risk of mosquito outbreaks in a first area; if the spatial distribution consistency between the verification result and the corresponding monitoring result of the verification data set meets a preset consistency standard, the first multivariate prediction model is used as the target prediction model.

[0015] Furthermore, determining the risk area for transmission of mosquito-borne diseases in the area to be tested based on the predicted probability corresponding to each sub-grid includes: performing data processing on the predicted probability corresponding to each sub-grid to obtain the prediction result of each sub-grid; and performing spatial mapping on the prediction result of each sub-grid to obtain the risk area for transmission of mosquito-borne diseases in the area to be tested.

[0016] In a second aspect, the present application provides a device for predicting mosquito outbreak risk, the device comprising:

[0017] A data determination module, used to determine density binary data corresponding to mosquito density data in a target area, and to determine a plurality of potential influencing factors affecting the mosquito density data in the target area;

[0018] A model building module, used to build a target prediction model for predicting mosquito outbreak risk based on the density binary data and the attribute data of each potential influencing factor;

[0019] A model prediction module is used to perform spatial downscaling processing on the area to be tested to obtain multiple sub-grids, and predict the predicted probability of mosquito outbreak in each sub-grid based on the target prediction model;

[0020] The risk area determination module is used to determine the transmission risk area of ​​the mosquito-borne disease in the test area based on the predicted probability corresponding to each sub-grid.

[0021] In a third aspect, the present application provides an electronic device, comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can execute the method for predicting mosquito outbreak risks described in any embodiment of the present application.

[0022] In a fourth aspect, the present application provides a computer-readable storage medium, which stores computer instructions, and the computer instructions are used to enable a processor to implement the method for predicting mosquito outbreak risks described in any embodiment of the present application when executed.

[0023] In order to address the defects of the prior art in the background technology, an embodiment of the present application provides a method for predicting the risk of mosquito outbreaks. Execution of the method can bring the following beneficial effects: the present application takes into account the obvious heterogeneity of mosquitoes in spatial distribution, performs spatial downscaling on the measured area to obtain multiple sub-grids, and predicts the risk of mosquito outbreaks at a fine spatial scale, which can improve the prediction accuracy of mosquito outbreak risks, accurately predict the spread of mosquito-borne diseases, and provide spatial information reference for formulating precise mosquito prevention and control plans.

[0024] It should be noted that the above-mentioned computer instructions can be stored in whole or in part on a computer-readable storage medium. The computer-readable storage medium can be packaged together with the processor of the device for predicting mosquito outbreak risk, or it can be packaged separately from the processor of the device for predicting mosquito outbreak risk, and this application does not limit this.

[0025] The description of the second, third and fourth aspects in this application can refer to the detailed description of the first aspect; and the beneficial effects of the description of the second, third and fourth aspects can refer to the beneficial effect analysis of the first aspect, which will not be repeated here.

[0026] It should be understood that the content described in this section is not intended to identify the key or important features of the embodiments of the present application, nor is it intended to limit the scope of the present application. Other features of the present application will become easily understood through the following description.

[0027] It is understandable that before using the technical solutions disclosed in the embodiments of this application, the type, scope of use, and usage scenarios of the personal information involved in this application should be informed to the user and the user's authorization should be obtained in an appropriate manner in accordance with relevant laws and regulations. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0029] Figure 1 A schematic diagram of a first process flow of a method for predicting mosquito outbreak risk provided in an embodiment of the present application;

[0030] Figure 2 A second flow chart of the method for predicting mosquito outbreak risk provided in an embodiment of the present application;

[0031] Figure 3 A spatially downscaled mosquito outbreak risk distribution map provided in an embodiment of the present application;

[0032] Figure 4 A schematic diagram of the structure of a device for predicting mosquito outbreak risk provided in an embodiment of the present application;

[0033] Figure 5 It is a block diagram of an electronic device used to implement the method for predicting mosquito outbreak risk in an embodiment of the present application. DETAILED DESCRIPTION

[0034] In order to make the purpose, technical solution and advantages of the embodiments of the present application clearer, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of this application.

[0035] It should be noted that the terms "first", "second", "target", and "original" in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments of the present application described herein can be implemented in an order other than that illustrated or described herein. In addition, the terms "including", "having", and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product, or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products, or devices.

[0036] Figure 1 This is a schematic diagram of the first process of the method for predicting mosquito outbreak risk provided in an embodiment of the present application. This embodiment can be applied to predict the risk area of ​​mosquito-borne disease transmission in the test area based on the mosquito density data of the test area. The method for predicting mosquito outbreak risk provided in this embodiment can be executed by the device for predicting mosquito outbreak risk provided in an embodiment of the present application, which can be implemented by software and / or hardware and integrated in an electronic device that executes this method.

[0037] See also Figure 1 The method of this embodiment includes but is not limited to the following steps:

[0038] S110, determining density binary data corresponding to the mosquito density data in the target area, and determining a plurality of potential influencing factors affecting the mosquito density data in the target area.

[0039] In the embodiment of the present application, the target area can be a larger scale area divided by administrative units such as towns and streets. First, the mosquito density in the target area is monitored to obtain mosquito density data, or the mosquito density data in the target area is directly obtained from existing studies. The mosquito density data can be represented by the Breteau Index (BI) data of the larvae of Aedes albopictus, specifically including the original monitoring data value, monitoring location and monitoring time in the target area in a certain year and month. Then, the mosquito-borne disease transmission threshold used to indicate whether the mosquito density data in the target area is prone to cause mosquito-borne diseases is determined. The usual mosquito-borne disease transmission threshold is BI=5. Finally, the mosquito density data is converted into density binary data according to the mosquito-borne disease transmission threshold. The density binary data is used to indicate whether the mosquito density data is less than the mosquito-borne disease transmission threshold, and takes values ​​of 1 and 0. Specifically, if the mosquito density data is not less than the mosquito-borne disease transmission threshold, that is, BI≥5, indicating that it is easy to cause mosquito-borne diseases, then the density binary data corresponding to the mosquito density data is determined to be 1; if the mosquito density data is less than the mosquito-borne disease transmission threshold, that is, BI<5, indicating that it is not easy to cause mosquito-borne diseases, then the density binary data corresponding to the mosquito density data is determined to be 0.

[0040] Potential influencing factors can be factors that affect larval growth, development and reproduction, and include at least one of temperature, rainfall, population density, land use data and Normalized Difference Vegetative Index (NDVI) data. Potential influencing factors can affect larval density to a certain extent. For example, larval growth and development are highly dependent on temperature and rainfall. Temperature mainly affects larval density by affecting larval growth and development speed, adult feeding behavior and reproduction rate. Rainfall mainly affects larval density by providing a suitable aquatic environment for larvae to survive. Population density is an important factor in mosquito density prediction, which not only reflects the degree of population concentration in the area, but also reflects the relative amount of food available to female adults. Land use data can be used to characterize the consistency of spatial distribution of various land use types, and can represent the impact on mosquito density in the socioeconomic dimension. The Normalized Vegetation Index is an important indicator for detecting the green growth state of vegetation and spatial vegetation coverage, and vegetation can provide a good natural environment for mosquitoes to live and reproduce. It should be noted that the above content is only an example of potential influencing factors and does not limit the potential influencing factors.

[0041] S120. Construct a target prediction model for predicting mosquito outbreak risk based on the density binary data and the attribute data of each potential influencing factor.

[0042] Among them, the attribute data of potential influencing factors may refer to the specific numerical values ​​corresponding to the potential influencing factors. For example, when the potential influencing factor is temperature, the corresponding attribute data is the average temperature value in the target area, such as the attribute data is 18 degrees Celsius; for another example, when the potential influencing factor is rainfall, the corresponding attribute data is the cumulative rainfall in the target area, such as 300 mm.

[0043] Specifically, a target prediction model for predicting the risk of mosquito outbreaks is constructed based on the density binary data and the attribute data of each potential influencing factor, including: in order to ensure the prediction accuracy of the target prediction model, it is necessary to screen out at least one target influencing factor from multiple potential influencing factors based on the density binary data and the attribute data of each potential influencing factor; and a target prediction model for predicting the risk of mosquito outbreaks is constructed based on the density binary data and the attribute data of the target influencing factor.

[0044] In an embodiment of the present application, the importance of each potential influencing factor is analyzed based on the density binary data analysis and the attribute data of each potential influencing factor, and then a preset number of target influencing factors of relatively important importance are screened out from multiple potential influencing factors. The density binary data and the attribute data of the target influencing factor are used as a training set, and the machine learning model is trained and the parameters are tuned to obtain a target prediction model. Optionally, the machine learning model can be a random forest model.

[0045] S130, performing spatial downscaling processing on the area to be tested to obtain multiple sub-grids, and predicting the probability of mosquito outbreak in each sub-grid based on the target prediction model.

[0046] Among them, the area to be tested is the area where mosquito outbreak prediction is required, which can be a larger scale area divided by administrative units such as towns and streets.

[0047] Since mosquitoes have a short flight distance and do not have the ability to migrate long distances, and their population density is affected by natural and socio-economic factors at different times and locations to different degrees, resulting in obvious heterogeneity in spatial distribution, this embodiment performs spatial downscaling on the area to be tested, studies the spatial downscaling prediction and mapping of mosquito outbreak risk at a fine scale, and migrates the target prediction model and probability conversion reference threshold established in the dimensions of larger scale areas such as towns and streets to the sub-grids, obtains the spatial downscaling prediction results of mosquito outbreak risk in the fine-scale area to be tested and maps them. Optionally, the interval to be tested can be divided into multiple sub-grids according to the 1km×1km grid standard rules.

[0048] For each sub-grid, the attribute data corresponding to the target influencing factors in each sub-grid are obtained as the target variable data for spatial downscaling prediction; the attribute data corresponding to the target influencing factors are input into the target prediction model, and the target prediction model is used to perform spatial downscaling prediction on the mosquito outbreak risk of each 1km×1km grid in the test area to obtain the mosquito outbreak risk data in each sub-grid, and then the predicted probability of mosquito outbreak in each sub-grid is obtained.

[0049] The advantage of this step of spatially downscaling the area to be tested is that the statistical indicators of the area to be tested are spatially downscaled and converted into regular grids, which can better reflect the differences within the area to be tested (a larger scale area divided by administrative units). If the monitoring indicator BI ≥ ​​5 requires manual intervention of mosquitoes, if it is based on the area to be tested, then the entire area to be tested must be intervened, and the resources consumed are very large. This embodiment can accurately determine which grids need to be intervened within the area to be tested, which can greatly reduce the intervention cost, the cost of resource allocation, and improve the accuracy of resource allocation.

[0050] S140 , determining the risk area for mosquito-borne disease transmission in the area to be tested based on the predicted probability corresponding to each sub-grid.

[0051] Specifically, the risk areas for mosquito-borne disease transmission in the area to be tested are determined based on the predicted probability corresponding to each sub-grid, including: performing data processing on the predicted probability corresponding to each sub-grid to obtain the predicted result of each sub-grid; and performing spatial mapping on the predicted result of each sub-grid to obtain the risk areas for mosquito-borne disease transmission in the area to be tested.

[0052] In the embodiment of the present application, the prediction data of the risk of mosquito outbreak in each sub-grid can be represented by a probability value, and different probability values ​​represent the probability that the mosquito density in the predicted area is greater than the critical threshold of the mosquito density of mosquitoes that transmit mosquito-borne infectious diseases. Finally, it can be converted into the prediction results of each sub-grid according to the preset reference threshold, and then the risk area of ​​mosquito-borne disease transmission can be drawn. Optionally, the probability of each sub-grid in the tested area being greater than the threshold of mosquito-borne infectious disease transmission can be obtained, which is used to draw the risk area of ​​mosquito-borne disease transmission based on the prediction data of mosquito outbreak risk at a fine scale.

[0053] The technical solution provided in this embodiment determines the density binary data corresponding to the mosquito density data in the target area, and determines multiple potential influencing factors that affect the mosquito density data in the target area; constructs a target prediction model for predicting the risk of mosquito outbreaks based on the density binary data and the attribute data of each potential influencing factor; performs spatial downscaling processing on the area to be tested to obtain multiple sub-grids, and predicts the predicted probability of mosquito outbreaks in each sub-grid based on the target prediction model; and determines the transmission risk area of ​​mosquito-borne diseases in the area to be tested based on the predicted probability corresponding to each sub-grid. Taking into account the obvious heterogeneity of mosquitoes in spatial distribution, this application performs spatial downscaling processing on the area to be tested to obtain multiple sub-grids, and predicts the transmission of mosquito-borne diseases at a fine spatial scale, which can improve the prediction accuracy of mosquito outbreak risks, accurately predict the transmission trend of mosquito-borne diseases, and provide spatial information reference for formulating precise mosquito prevention and control plans.

[0054] The following further describes the method for predicting mosquito outbreak risk provided by the embodiments of the present application. Figure 2 A second flow chart of the method for predicting mosquito outbreak risk provided in the embodiment of the present application. The embodiment of the present application is optimized on the basis of the above embodiments, and the specific optimization is as follows: the present embodiment explains the training process of the target prediction model in detail.

[0055] See also Figure 2 The method of this embodiment includes but is not limited to the following steps:

[0056] S210 , constructing a corresponding univariate prediction model for each potential influencing factor according to the density binary data and the attribute data of each potential influencing factor.

[0057] Among them, the univariate prediction model refers to a machine learning model trained based on the attribute data of a single potential influencing factor.

[0058] In the embodiment of the present application, the attribute data of each potential influencing factor is used as a sample together with the density binary data to train the initial machine learning model corresponding to the potential influencing factor. After training and parameter tuning, the univariate prediction model corresponding to the potential influencing factor is obtained. One potential influencing factor corresponds to one univariate prediction model. Optionally, the initial machine learning model can use a random forest model.

[0059] In a preferred embodiment, in order to perform spatial downscaling prediction and mapping of mosquito outbreak risks at a fine scale and improve the prediction accuracy of the target prediction model, it is necessary to improve the prediction accuracy of the univariate prediction model in this step. Therefore, the target area can be spatially downscaled to obtain several sub-areas. For each sub-area, if the mosquito density data in the sub-area is not less than the threshold value that causes the spread of mosquito-borne diseases, that is, BI≥5, the density binary data of the sub-area can be recorded as 1; if the mosquito density data in the sub-area is less than the threshold value that causes the spread of mosquito-borne diseases, that is, BI<5, the density binary data of the sub-area can be recorded as 0, and the density binary data of all sub-areas constitute the density binary data in the target area.

[0060] The attribute data and density binary data corresponding to the potential influencing factors in each sub-region are used as training samples of the initial machine learning model of the corresponding sub-region. The training samples in multiple sub-regions are used as training sets in the target area. The initial machine learning model is trained and parameter tuning is performed. Finally, after the training is completed, a univariate prediction model corresponding to each potential influencing factor is obtained.

[0061] For example, assuming that the potential influencing factors are temperature and precipitation, the target area is divided into sub-area A, sub-area B and sub-area C; the temperature value and density binary data of sub-area A are used as the first training sample, the temperature value and density binary data of sub-area B are used as the second training sample, and the temperature value and density binary data of sub-area C are used as the third training sample; the first training sample, the second training sample and the third training sample form a training set, and a random forest model is established for parameter tuning and training to obtain a single variable prediction model corresponding to the potential influencing factor of temperature. Similarly, the rainfall and density binary data of sub-area A are used as the fourth training sample, the rainfall and density binary data of sub-area B are used as the fifth training sample, and the rainfall and density binary data of sub-area C are used as the sixth training sample, and another random forest model is trained to obtain a single variable prediction model corresponding to the potential influencing factor of rainfall. That is, the corresponding random forest model is trained based on the attribute data and density binary data of each potential influencing factor to obtain a single variable prediction model corresponding to each potential influencing factor.

[0062] It should be noted that the above content only explains the sub-region division method and the training process of the univariate prediction model in the target region, and does not limit the number of sub-regions, the number of samples, and the number of potential influencing factors. In other words, the number of sub-regions, the number of samples, and the number of potential influencing factors in the target region can be multiple.

[0063] S220: Determine the prediction accuracy of each univariate prediction model.

[0064] In an embodiment of the present application, a univariate prediction model can predict density binary data. For each univariate prediction model, the attribute data to be predicted and the density binary data corresponding to the attribute data can be acquired in advance, and the attribute data can be input into the univariate prediction model. The univariate prediction model can output the prediction result of the corresponding density binary data; further, the prediction accuracy of the univariate prediction model can be determined based on the difference between the prediction result of the density binary data and the original monitoring result.

[0065] Preferably, the prediction accuracy of the univariate prediction model is evaluated by two indicators, one of which is specificity, which reflects the accuracy of the model's prediction results for the majority class samples, and its value is between 0-1, and as the value increases, the prediction results for the category are better. Another indicator is recall, which reflects the recall rate of the model. The higher the recall rate, the higher the accuracy of the model's prediction results for the minority class samples. Further, both G-means and Receiver Operating Characteristic-Area Under Curve (ROC-AUC) are taken into account to calculate the score corresponding to the accuracy of the univariate prediction model. The higher the score, the better the prediction performance of the model.

[0066] S230. Determine a correlation matrix of multiple potential influencing factors.

[0067] In the embodiment of the present application, if the correlation between two potential influencing factors is high, it may cause the prediction model to have multicollinearity problems, affecting the accuracy of the prediction results. In order to minimize the strong correlation between potential influencing factors, it can be considered from the stage of obtaining mosquito density data; for example: if multiple potential influencing factors include temperature and rainfall, due to the strong spatiotemporal autocorrelation and heterogeneity of temperature and rainfall, when obtaining mosquito density data, the reference adult generally flies less than 1km, and the 1km temperature and rainfall images can be used for 2x2 convolution calculation to express the spatial autocorrelation relationship between adjacent pixels and target pixels, and finally the mean of all pixels is taken to update the central pixel value. In terms of time, the temperature and rainfall of the current month and the lagged one month are selected to express the autocorrelation relationship at different times, and finally the mean of two months is taken to update the pixel value.

[0068] Specifically, for every two potential influencing factors among the multiple potential influencing factors, the correlation coefficient between every two potential influencing factors is determined, and the correlation coefficient is used to calculate the correlation between every two potential influencing factors. Optionally, the Pearson correlation coefficient between every two potential influencing factors can be calculated and used as a reference indicator to measure the correlation between every two potential influencing factors.

[0069] S240: Screen out at least one target influencing factor from multiple potential influencing factors according to the correlation matrix and the prediction accuracy.

[0070] In an embodiment of the present application, at least one target influencing factor is selected from multiple potential influencing factors according to a correlation matrix and prediction accuracy, specifically including: for the current potential influencing factor, determining the primary influencing factor with a high correlation with the current potential influencing factor according to the correlation matrix; selecting candidate influencing factors that meet the preset accuracy standard from the primary influencing factors according to the prediction accuracy; traversing multiple potential influencing factors in turn to obtain multiple candidate influencing factors, and determining the multiple candidate influencing factors as target influencing factors. Among them, the primary influencing factor is at least one potential influencing factor among multiple potential influencing factors, and the primary influencing factors corresponding to the current potential influencing factor include the current potential influencing factor.

[0071] In another optional embodiment, for the current potential influencing factor, if the correlation corresponding to the current potential influencing factor and the prediction accuracy of the corresponding univariate prediction model meet the preset conditions, the current potential influencing factor is used as a candidate influencing factor, and multiple potential influencing factors are traversed in sequence to obtain multiple candidate influencing factors, and the multiple candidate influencing factors are determined as target influencing factors. The preset conditions may be a preset accuracy threshold and correlation coefficient threshold.

[0072] For example, if among multiple potential influencing factors, one of the potential influencing factors has a higher prediction accuracy, for example, the recall value reaches the recall rate threshold, indicating that the potential influencing factor may be an important factor affecting mosquito density; at this time, the correlation coefficient of the potential influencing factor is judged. If the correlation coefficients of the potential influencing factor and other potential influencing factors are less than the correlation coefficient threshold, it means that the correlation between the potential influencing factor and other potential influencing factors is not large. If the attribute data of the potential influencing factor is used to predict the density binary data without causing obvious collinearity problems, the potential influencing factor can be used as the target influencing factor. Based on the same method, the influencing factors that meet the above preset conditions among multiple potential influencing factors can all be used as target influencing factors.

[0073] S250, performing at least two resampling processes on the density binary data to obtain corresponding at least two types of sample data sets.

[0074] Since mosquito density has obvious heterogeneity in spatial distribution and obvious seasonality in time, it jointly affects the imbalance of mosquito density category samples. The target area contains multiple sub-areas, and the number of sub-areas with mosquito density data greater than the threshold of mosquito-borne disease transmission is much smaller than the number of sub-areas without cases, that is, there is an imbalance problem in the sample data, and the imbalance problem of sample data will reduce the prediction accuracy of the prediction model. The results obtained by the existing balanced data prediction model cannot reflect the imbalance characteristics of the original monitoring data. Therefore, in order to solve the problem of imbalanced sample data, at least two resampling methods are used in this embodiment to process the density binary data in the target area to obtain corresponding at least two types of sample data sets.

[0075] This embodiment uses four types of resampling as examples, namely no sampling, undersampling, oversampling and combined sampling. Among them, oversampling is to achieve a balance in the number of samples of the two categories by randomly copying some minority class samples. Undersampling is to randomly delete majority class samples to achieve a balance in the number of samples of the two categories. Combined sampling is to randomly delete majority class samples and randomly copy minority class samples to achieve a balance in the number of samples of the two categories.

[0076] S260: construct multivariate prediction models according to the attribute data of the target influencing factors and each type of sample data set, thereby obtaining at least two multivariate prediction models.

[0077] Exemplarily, if the target influencing factors include factor A, factor B and factor C, the attribute data corresponding to factor A, factor B and factor C are fused together, and the fusion process can be to fuse the attribute data corresponding to factor A, factor B and factor C in the same sub-region together. Further, the density binary data in the target region is undersampled, and the density binary data after undersampling and the fused attribute data are used as training samples to train the random forest model to obtain a multivariate prediction model corresponding to the undersampling. Similarly, the density binary data after oversampling and the fused attribute data are used as training samples to train the random forest model to obtain a multivariate prediction model corresponding to the oversampling. Similarly, the density binary data after combined sampling and the fused attribute data are used as training samples to train the random forest model to obtain a multivariate prediction model corresponding to the combined sampling. In addition, the original unsampled density binary data and the fused attribute data are used as training samples to train the random forest model to obtain a multivariate prediction model corresponding to the unsampled process.

[0078] S270: Determine prediction performance data of at least two multivariate prediction models.

[0079] The prediction performance data includes at least one of sensitivity (i.e., accuracy), specificity, recall rate, G-means, and AUC. The recall rate value is between 0 and 1. The larger the value, the lower the probability of sample omission. It is used as an important reference indicator for evaluating model accuracy.

[0080] S280: Filter out a target prediction model from at least two multivariate prediction models based on the prediction performance data.

[0081] Specifically, a target prediction model is selected from at least two multivariate prediction models based on the prediction performance data, including: determining the recall and specificity of each of the at least two multivariate prediction models for predicting the occurrence of mosquito-borne diseases; selecting a first multivariate prediction model with a specificity greater than a preset threshold (such as 0.5) and a maximum recall from at least two multivariate prediction models; using a validation data set to validate the first multivariate prediction model to obtain a validation result, where the validation data set is a monitoring data set for monitoring the risk of mosquito outbreaks in a first area; if the spatial distribution consistency between the validation result and the corresponding monitoring result of the validation data set meets a preset consistency standard (such as the highest accuracy), the first multivariate prediction model is used as the target prediction model.

[0082] Taking into account the prediction accuracy of at least two multivariate prediction models for each type of sample data set, it is believed that the optimal model should be characterized by a specificity greater than 0.5 and a maximum recall rate, or a non-significant change in recall rate and a significant change in specificity. In order to compare the consistency and difference in the spatial distribution of the monitoring results and the verification results of BI ≥ ​​5, we considered defining the first 10 percentiles of the prediction probability as the reference threshold for converting the binary continuous value prediction probability to the binary results of BI ≥ ​​5 and BI < 5.

[0083] The technical solution provided in this embodiment is to construct a corresponding univariate prediction model for each potential influencing factor according to the density binary data and the attribute data of each potential influencing factor; determine the prediction accuracy of each univariate prediction model; determine the correlation matrix of multiple potential influencing factors; screen out at least one target influencing factor from multiple potential influencing factors according to the correlation matrix and the prediction accuracy; perform at least two resampling processes on the density binary data to obtain corresponding at least two types of sample data sets; construct a multivariate prediction model according to the attribute data of the target influencing factor and each type of sample data set, so as to obtain at least two multivariate prediction models; determine the prediction performance data of at least two multivariate prediction models; and screen out the target prediction model from at least two multivariate prediction models based on the prediction performance data. This application takes into account the obvious heterogeneity of mosquito density in spatial distribution, and performs spatial downscaling processing on the target area to obtain multiple sub-areas, which can improve the prediction accuracy of the target prediction model; and because mosquito density has obvious seasonality in time, the original sample data of mosquitoes (i.e., density binary data) is subjected to multiple resampling processing to obtain multi-class sample data sets, which can solve the defects caused by unbalanced sample data, improve the prediction accuracy of mosquito density, accurately predict the spread of mosquito-borne diseases, and provide spatial information reference for formulating precise prevention and control plans for mosquito density.

[0084] In a practical application, the prediction performance data comparison of the multivariate prediction model corresponding to different resampling methods is shown in Table 1:

[0085] Table 1 Prediction performance data of multivariate prediction models corresponding to different resampling methods

[0086] Sampling method Accuracy Specificity Recall G-means AUC Unprocessed 0.9432 0.9892 0.2469 0.4911 0.8643 Undersampling 0.7701 0.7682 0.7977 0.7821 0.8468 Oversampling 0.9330 0.9689 0.3903 0.6124 0.8614 Combined Sampling 0.9162 0.9396 0.5604 0.7244 0.8609

[0087] Table 1 shows that the recall rate of the multivariate prediction model corresponding to unsampling for BI≥5 samples is as low as 0.2469, while the recall rate of the multivariate prediction model corresponding to undersampling for BI≥5 samples is as high as 0.7977, which significantly reduces the probability of sample omission compared with the multivariate prediction model corresponding to unsampling. The prediction results of each indicator of the multivariate prediction model corresponding to oversampling and the multivariate prediction model corresponding to combined sampling are between the results of the multivariate prediction model corresponding to unsampling and the multivariate prediction model corresponding to undersampling. There is no significant difference in the ROC-AUC prediction index among the four models, all of which are maintained at around 0.85. Although the prediction accuracy of the multivariate prediction model corresponding to undersampling for BI<5 samples is about 18% lower than that of other multivariate prediction models, the prediction accuracy of BI≥5 samples is about 223%, 104% and 42% higher than that of the multivariate prediction model corresponding to unsampling, the multivariate prediction model corresponding to oversampling and the multivariate prediction model corresponding to combined sampling respectively; G-means is about 59%, 28% and 8% higher than that of the multivariate prediction model corresponding to unsampling, the multivariate prediction model corresponding to oversampling and the multivariate prediction model corresponding to combined sampling respectively. The recall rate of the multivariate prediction model corresponding to undersampling for BI≥5 samples is significantly better than that of the multivariate prediction model corresponding to oversampling, the multivariate prediction model corresponding to combined sampling and the multivariate prediction model corresponding to unsampling, and the results of other indicators are not significantly lower than those of other multivariate prediction models and remain at about 0.8. It has good prediction ability as a whole and is considered to be the optimal prediction model for this implementation.

[0088] In order to verify the consistency of the spatial distribution of the prediction results of BI≥5 in streets and towns based on the multivariate prediction model corresponding to undersampling and the monitoring results, the probability of BI≥5 in streets and towns was predicted monthly and the spatial distribution maps of the prediction results in April, June and September were drawn. Table 2 Except for L Street, F Street and J Town, the prediction results of BI≥5 were consistent with the monitoring results. Compared with the monitoring results, except for 6 streets and towns with wrong predictions, the other 17 prediction results were consistent with the monitoring results of streets and towns. Compared with the monitoring results, except for 7 streets and towns with wrong predictions, the other 12 prediction results were basically consistent with the monitoring results. The change of larval density is restricted by the density of mosquitoes at each growth stage within the population, the influence of changes in the external environment and the influence of time change trends, which will increase the difficulty of accurately predicting BI≥5. Although the accuracy of the prediction results of BI≥5 of the multivariate prediction model corresponding to undersampling varies in different months, the spatial location of the prediction results of each month is basically consistent with the monitoring results, which can accurately reflect the spatial location of streets and towns with BI≥5 in the monitoring data.

[0089] Table 2 Specific data of verification results and test results of multivariate prediction model corresponding to undersampling

[0090]

[0091] The multivariate prediction model corresponding to undersampling has the highest recall rate for BI≥5 samples as the target prediction model, and the other evaluation indicators are maintained at around 0.8, with good overall prediction performance. The multivariate prediction model corresponding to undersampling is used to predict the probability of mosquito density BI≥5 in 1km grid unit in 2019 on a monthly basis, and a kilometer grid and measurement result distribution map is drawn.

[0092] Figure 3 This is a spatial downscaled mosquito outbreak risk distribution map. Figure 3 From 1 and 4, we can see that the comparison between the prediction results of streets and towns and the grid prediction results shows that most of the streets and towns in the central urban area are below the BI<5 level, and the spatial distribution of BI≥5 of the two prediction results is quite similar, both showing a distribution characteristic of low in the middle and high around. However, it can still be found that a small number of BI≥5 km grids are distributed inside the streets and towns in the central urban area, such as W Street. Figure 3 2 and 5 show that the comparison between the street and town prediction results and the grid prediction results shows that although the prediction result of K town in the northeast is BI ≥ ​​5, the grid prediction results in this area are low or even no grid with BI ≥ ​​5 appears. Table 3 shows the prediction results of mosquito outbreak risk in the tested area. The street and town prediction results in the table show that N town has BI ≥ ​​5 from June to August, but Figure 3 The spatial distribution maps of 4, 5 and 6 during the same period show that the number of BI ≥ ​​5 grids in the region is relatively sparse and shows a trend of increasing first and then decreasing over time. Its area share gradually decreases from 2.1% in July to 0.8% in August, and the position of BI ≥ ​​5 grids changes between months. Table 3 shows that the area share of BI ≥ ​​5 grids in the prediction results of BI ≥ ​​5 streets and towns is less than 50%, and more area is occupied by BI < 5 grids. The monitoring results of BI < 5 streets and towns show that except for M Town, R Town, K Town, and S Street, the rest of the streets and towns have BI ≥ ​​5 grids distributed inside, and the area share ranges from 0.3% to 22.6%. Comprehensive Table 3 and Figure 3 The grid prediction results show that there are large differences between the fine-scale prediction results and the street and town prediction results in terms of the spatial distribution position and area proportion of the BI ≥ ​​5 grids. The difference can be divided into the following two basic situations according to BI = 5: 1. Only some grid prediction results BI ≥ ​​5 exist in the street and town monitoring with BI ≥ ​​5; 2. There may also be a small number of BI ≥ ​​5 grids within the street and town monitoring with BI < 5.

[0093] Table 3 Prediction results of mosquito outbreak risk in the tested areas

[0094]

[0095]

[0096] Figure 4 A schematic diagram of a device for predicting mosquito outbreak risk provided in an embodiment of the present application is shown in FIG. Figure 4 As shown, the apparatus 400 may include:

[0097] The data determination module 410 is used to determine the density binary data corresponding to the mosquito density data in the target area, and determine a plurality of potential influencing factors affecting the mosquito density data in the target area;

[0098] A model building module 420, for building a target prediction model for predicting mosquito outbreak risk based on the density binary data and the attribute data of each potential influencing factor;

[0099] The model prediction module 430 is used to perform spatial downscaling processing on the area to be tested to obtain multiple sub-grids, and predict the predicted probability of mosquito outbreak in each sub-grid based on the target prediction model;

[0100] The risk area determination module 440 is used to determine the transmission risk area of ​​the mosquito-borne disease in the area to be tested based on the predicted probability corresponding to each sub-grid.

[0101] Furthermore, the above-mentioned model building module 420 can be specifically used to: screen out at least one target influencing factor from the multiple potential influencing factors based on the density binary data and the attribute data of each potential influencing factor; and construct the target prediction model for predicting the risk of mosquito outbreaks based on the density binary data and the attribute data of the target influencing factor.

[0102] Furthermore, the above-mentioned model building module 420 may include: an influencing factor screening unit and a prediction model building unit;

[0103] The influencing factor screening unit is used to construct a corresponding univariate prediction model for each potential influencing factor according to the density binary data and the attribute data of each potential influencing factor; determine the prediction accuracy of each univariate prediction model; determine the correlation matrix of the multiple potential influencing factors; and screen out the at least one target influencing factor from the multiple potential influencing factors according to the correlation matrix and the prediction accuracy.

[0104] The prediction model construction unit is used to perform at least two resampling processes on the density binary data to obtain corresponding at least two types of sample data sets; construct multivariate prediction models according to the attribute data of the target influencing factors and each type of sample data set, so as to obtain at least two multivariate prediction models; determine the prediction performance data of the at least two multivariate prediction models; and screen out the target prediction model from the at least two multivariate prediction models based on the prediction performance data.

[0105] Furthermore, the above-mentioned influencing factor screening unit can be specifically used to: for the current potential influencing factor, determine the preliminary influencing factor that has a high correlation with the current potential influencing factor according to the correlation matrix, and the preliminary influencing factor is at least one potential influencing factor among the multiple potential influencing factors; select candidate influencing factors that meet the preset accuracy standard from the preliminary influencing factors according to the prediction accuracy; traverse the multiple potential influencing factors in turn to obtain multiple candidate influencing factors, and determine the multiple candidate influencing factors as the target influencing factors.

[0106] Furthermore, the above-mentioned prediction model construction unit can be specifically used to: determine the recall rate and specificity of each of the at least two multivariate prediction models for predicting the mosquito outbreak risk; select the first multivariate prediction model with the specificity greater than a preset threshold and the largest recall rate from the at least two multivariate prediction models; use a verification data set to verify the first multivariate prediction model to obtain a verification result, and the verification data set is a monitoring data set for monitoring the mosquito outbreak risk in the first area; if the spatial distribution consistency between the verification result and the corresponding monitoring result of the verification data set meets the preset consistency standard, then the first multivariate prediction model is used as the target prediction model.

[0107] Furthermore, the risk area determination module 440 can be specifically used to: perform data processing on the prediction probability corresponding to each sub-grid to obtain the prediction result of each sub-grid; perform spatial mapping on the prediction result of each sub-grid to obtain the transmission risk area of ​​mosquito-borne diseases in the area to be tested.

[0108] The device for predicting mosquito outbreak risk provided in this embodiment can be applied to the method for predicting mosquito outbreak risk provided in any of the above embodiments, and has corresponding functions and beneficial effects.

[0109] Figure 5 It is a block diagram of an electronic device for implementing a method for predicting the risk of mosquito outbreaks in an embodiment of the present application. The electronic device 10 is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workbenches, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital processing, cellular phones, smart phones, wearable devices (such as helmets, glasses, watches, etc.) and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present application described and / or required herein.

[0110] like Figure 5As shown, the electronic device 10 includes at least one processor 11, and a memory connected to the at least one processor 11, such as a read-only memory (ROM) 12, a random access memory (RAM) 13, etc., wherein the memory stores a computer program that can be executed by at least one processor, and the processor 11 can perform various appropriate actions and processes according to the computer program stored in the read-only memory (ROM) 12 or the computer program loaded from the storage unit 18 to the random access memory (RAM) 13. In the RAM 13, various programs and data required for the operation of the electronic device 10 can also be stored. The processor 11, the ROM 12, and the RAM 13 are connected to each other through a bus 14. An input / output (I / O) interface 15 is also connected to the bus 14.

[0111] A number of components in the electronic device 10 are connected to the I / O interface 15, including: an input unit 16, such as a keyboard, a mouse, etc.; an output unit 17, such as various types of displays, speakers, etc.; a storage unit 18, such as a disk, an optical disk, etc.; and a communication unit 19, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 19 allows the electronic device 10 to exchange information / data with other devices through a computer network such as the Internet and / or various telecommunication networks.

[0112] The processor 11 may be a variety of general and / or special processing components with processing and computing capabilities. Some examples of the processor 11 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, a digital signal processor (DSP), and any appropriate processor, controller, microcontroller, etc. The processor 11 executes the various methods and processes described above, such as a method for predicting mosquito outbreak risks.

[0113] In some embodiments, the method for predicting the risk of a mosquito outbreak may be implemented as a computer program, which is tangibly contained in a computer-readable storage medium, such as a storage unit 18. In some embodiments, part or all of the computer program may be loaded and / or installed on the electronic device 10 via the ROM 12 and / or the communication unit 19. When the computer program is loaded into the RAM 13 and executed by the processor 11, one or more steps of the method for predicting the risk of a mosquito outbreak described above may be performed. Alternatively, in other embodiments, the processor 11 may be configured to execute the method for predicting the risk of a mosquito outbreak in any other appropriate manner (e.g., by means of firmware).

[0114] Various embodiments of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chips (SOCs), load programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments may include: being implemented in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which can be a special purpose or general purpose programmable processor that can receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device.

[0115] The computer programs for implementing the methods of the present application may be written in any combination of one or more programming languages. These computer programs may be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, so that when the computer programs are executed by the processor, the functions / operations specified in the flow charts and / or block diagrams are implemented. The computer programs may be executed entirely on the machine, partially on the machine, partially on the machine and partially on a remote machine as a stand-alone software package, or entirely on a remote machine or server.

[0116] In the context of the present application, a computer readable storage medium may be a tangible medium that may contain or store a computer program for use by or in conjunction with an instruction execution system, device, or equipment. A computer readable storage medium may include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, device, or equipment, or any suitable combination of the foregoing. Alternatively, a computer readable storage medium may be a machine readable signal medium. A more specific example of a machine readable storage medium may include an electrical connection based on one or more lines, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0117] To provide interaction with a user, the systems and techniques described herein may be implemented on an electronic device having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and a pointing device (e.g., a mouse or trackball) through which the user can provide input to the electronic device. Other types of devices may also be used to provide interaction with the user; for example, the feedback provided to the user may be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user may be received in any form (including acoustic input, voice input, or tactile input).

[0118] The systems and techniques described herein can be implemented in a computing system that includes backend components (e.g., as a data server) or a computing system that includes middleware components (e.g., an application server) or a computing system that includes frontend components (e.g., a user computer with a graphical user interface or a web browser through which a user can interact with implementations of the systems and techniques described herein) or a computing system that includes any combination of such backend components, middleware components, or frontend components. The components of the system can be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: a local area network (LAN), a wide area network (WAN), a blockchain network, and the Internet.

[0119] A computing system may include a client and a server. The client and the server are generally remote from each other and usually interact through a communication network. The client and server relationship is generated by computer programs running on the corresponding computers and having a client-server relationship with each other. The server may be a cloud server, also known as a cloud computing server or cloud host, which is a host product in the cloud computing service system to solve the defects of difficult management and weak business scalability in traditional physical hosts and VPS services.

[0120] Note that the above are only preferred embodiments of the present application and the technical principles used. Those skilled in the art will understand that the present application is not limited to the specific embodiments described herein, and that various obvious changes, readjustments, and substitutions can be made by those skilled in the art without departing from the scope of protection of the present application. For example, those skilled in the art can use the various forms of processes shown above to reorder, add, or delete steps; the steps recorded in the present application can be performed in parallel, sequentially, or in different orders, as long as the desired results of the technical solution of the present application can be achieved, and this document does not limit them here.

[0121] The above specific implementations do not constitute a limitation on the protection scope of this application. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions and improvements made within the spirit and principles of this application should be included in the protection scope of this application.

Claims

1. A method for predicting mosquito outbreak risk, characterized in that: The method comprises: Determine density binary data corresponding to mosquito density data in a target area, and determine a plurality of potential influencing factors affecting the mosquito density data in the target area; wherein the density binary data is used to indicate whether the mosquito density data is less than a mosquito-borne disease transmission threshold; Constructing a target prediction model for predicting the risk of mosquito outbreaks according to the density binary data and the attribute data of each potential influencing factor, including: selecting at least one target influencing factor from the multiple potential influencing factors according to the density binary data and the attribute data of each potential influencing factor; constructing the target prediction model for predicting the risk of mosquito outbreaks according to the density binary data and the attribute data of the target influencing factor; Performing spatial downscaling processing on the area to be tested to obtain multiple sub-grids, and predicting the predicted probability of mosquito outbreak in each sub-grid based on the target prediction model; Determine the transmission risk area of ​​the mosquito-borne disease in the test area based on the predicted probability corresponding to each sub-grid; Among them, the target prediction model for predicting the risk of mosquito outbreaks is constructed based on the density binary data and the attribute data of the target influencing factors, including: performing at least two resampling processes on the density binary data to obtain corresponding at least two types of sample data sets; constructing multivariate prediction models based on the attribute data of the target influencing factors and each type of sample data set, thereby obtaining at least two multivariate prediction models; determining the prediction performance data of the at least two multivariate prediction models; screening out the target prediction model from the at least two multivariate prediction models based on the prediction performance data; the target prediction model uses the attribute data corresponding to the target influencing factor as input, predicts the risk of mosquito outbreaks in each sub-grid, and obtains the predicted probability of mosquito outbreaks.

2. The method for predicting mosquito outbreak risk according to claim 1, characterized in that: The step of selecting at least one target influencing factor from the plurality of potential influencing factors according to the density binary data and the attribute data of each potential influencing factor includes: Constructing a corresponding univariate prediction model for each potential influencing factor according to the density binary data and the attribute data of each potential influencing factor; Determine the prediction accuracy of each univariate prediction model; Determining a correlation matrix of the plurality of potential influencing factors; The at least one target influencing factor is screened out from the multiple potential influencing factors according to the correlation matrix and the prediction accuracy.

3. The method for predicting mosquito outbreak risk according to claim 2, characterized in that: The step of selecting the at least one target influencing factor from the plurality of potential influencing factors according to the correlation matrix and the prediction accuracy comprises: For the current potential influencing factor, determine, according to the correlation matrix, a preliminary influencing factor having a high correlation with the current potential influencing factor, wherein the preliminary influencing factor is at least one potential influencing factor among the multiple potential influencing factors; Selecting candidate influencing factors that meet a preset accuracy standard from the preliminary influencing factors according to the prediction accuracy; The multiple potential influencing factors are traversed in sequence to obtain multiple candidate influencing factors, and the multiple candidate influencing factors are determined as the target influencing factors.

4. The method for predicting mosquito outbreak risk according to claim 1, characterized in that: The step of selecting the target prediction model from the at least two multivariate prediction models based on the prediction performance data comprises: Determining the recall and specificity of each of the at least two multivariate prediction models for predicting mosquito outbreak risk; Selecting a first multivariate prediction model having a specificity greater than a preset threshold and a maximum recall rate from the at least two multivariate prediction models; Using a validation data set to validate the first multivariate prediction model to obtain a validation result, wherein the validation data set is a monitoring data set for monitoring the mosquito outbreak risk in the first area; If the spatial distribution consistency between the verification result and the monitoring result corresponding to the verification data set meets the preset consistency standard, the first multivariate prediction model is used as the target prediction model.

5. The method for predicting mosquito outbreak risk according to claim 1, characterized in that: The step of determining the transmission risk area of ​​the mosquito-borne disease in the area to be tested based on the predicted probability corresponding to each sub-grid comprises: Performing data processing on the prediction probability corresponding to each sub-grid to obtain a prediction result of each sub-grid; The prediction results of each sub-grid are spatially mapped to obtain the transmission risk area of ​​mosquito-borne diseases in the area to be tested.

6. A device for predicting mosquito outbreak risk, characterized in that: The device comprises: A data determination module, used to determine density binary data corresponding to mosquito density data in a target area, and to determine a plurality of potential influencing factors affecting the mosquito density data in the target area; wherein the density binary data is used to indicate whether the mosquito density data is less than a mosquito-borne disease transmission threshold; A model building module, used to build a target prediction model for predicting mosquito outbreak risk based on the density binary data and the attribute data of each potential influencing factor; A model prediction module is used to perform spatial downscaling processing on the area to be tested to obtain multiple sub-grids, and predict the predicted probability of mosquito outbreak in each sub-grid based on the target prediction model; A risk area determination module, used to determine the transmission risk area of ​​the mosquito-borne disease in the test area based on the predicted probability corresponding to each sub-grid; The model building module is specifically used to select at least one target influencing factor from the multiple potential influencing factors according to the density binary data and the attribute data of each potential influencing factor; and to build the target prediction model for predicting the risk of mosquito outbreaks according to the density binary data and the attribute data of the target influencing factor; The model building module includes a prediction model building unit, which is used to perform at least two resampling processes on the density binary data to obtain corresponding at least two types of sample data sets; construct multivariate prediction models according to the attribute data of the target influencing factor and each type of sample data set, so as to obtain at least two multivariate prediction models; determine the prediction performance data of the at least two multivariate prediction models; screen out the target prediction model from the at least two multivariate prediction models based on the prediction performance data; the target prediction model uses the attribute data corresponding to the target influencing factor as input, predicts the risk of mosquito outbreaks in each sub-grid, and obtains the predicted probability of mosquito outbreaks.

7. An electronic device, characterized in that: The electronic device comprises: at least one processor; and a memory communicatively connected to the at least one processor; wherein, The memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can execute the method for predicting mosquito outbreak risk as described in any one of claims 1 to 5.

8. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a processor to implement the method for predicting mosquito outbreak risk as described in any one of claims 1 to 5 when executed.

Citation Information

Patent Citations

  • Dengue fever prevention and control method and system

    CN108711455A

  • Dengue fever infectious disease prediction method based on meteorological model

    CN112397205A

  • Dengue propagation risk prediction method and affected population size determination method

    CN112951440A

  • Disease risk level prediction method and device, readable storage medium and electronic equipment

    CN114300141A

  • Method and system for identifying dengue mosquito-borne suitable region and main influence factors

    CN114707079A