Method and system for predicting spatial distribution of alveolar echinococcosis based on multi-source environmental factors

Abstract

The present application relates to the technical field of disease distribution prediction, and discloses a method and system for predicting the spatial distribution of alveolar echinococcosis based on multi-source environmental factors, comprising: obtaining multi-source environmental data of alveolar echinococcosis case points in a study area, processing the multi-source environmental data, and screening key environmental factors; obtaining a data set of sample points based on the key environmental factors, and dividing the data set into a training set and a test set; selecting a spatial prediction model, training and testing the spatial prediction model based on the training set and the test set, and obtaining an alveolar echinococcosis spatial prediction model; dividing the study area into a plurality of geographic units, obtaining the incidence probability of alveolar echinococcosis of each geographic unit based on the alveolar echinococcosis spatial prediction model, and determining the risk level. The present application can accurately predict the spatial distribution of alveolar echinococcosis by comprehensively considering multi-source environmental factors.

Description

Technical Field

[0001] This invention relates to the field of disease distribution prediction technology, and in particular to a method and system for predicting the spatial distribution of alveolar echinococcosis based on multiple environmental factors. Background Technology

[0002] Alveolar echinococcosis is a serious parasitic disease that severely endangers human health, and its incidence is closely related to environmental factors. Accurately understanding the spatial distribution patterns and risk levels of this disease is of great significance for disease prevention and control, and for optimizing resource allocation.

[0003] Existing methods for predicting the distribution of alveolar echinococcosis mostly rely on single environmental factors or simple statistical analysis, resulting in problems such as low prediction accuracy, limited coverage, and inability to quantify regional risk differences. With the development of environmental data acquisition technologies, the integrated utilization of multi-source environmental data has become crucial for improving prediction accuracy. However, there is currently a lack of a systematic and operational spatial distribution prediction scheme based on multi-source environmental factors, making it difficult to meet the actual needs of disease prevention and control.

[0004] Therefore, there is an urgent need for a method to predict the spatial distribution of alveolar echinococcosis based on multi-source environmental factors. This method can achieve accurate prediction of the spatial distribution and risk level of alveolar echinococcosis by integrating multi-dimensional environmental data, screening key influencing factors, and optimizing the prediction model. Summary of the Invention

[0005] The purpose of this invention is to provide a method and system for predicting the spatial distribution of alveolar echinococcosis based on multiple environmental factors, aiming to solve the problems of traditional methods that rely on a single environmental factor or simple statistical analysis, resulting in low prediction accuracy, limited coverage, and inability to quantify regional risk differences.

[0006] This invention provides a method for predicting the spatial distribution of alveolar echinococcosis based on multiple environmental factors, characterized by comprising: Multi-source environmental data of echinococcosis cases within the study area were acquired, and the multi-source environmental data were preprocessed, including data cleaning, outlier handling, and data standardization. Correlation analysis or principal component analysis was used to determine the correlation strength between each environmental factor and the incidence of alveolar echinococcosis, and key environmental factors were screened based on the correlation strength. Set up sample points, mark vesicular echinococcosis case points as positive sample points, and mark several non-vesicular echinococcosis case points in the same study area as negative sample points. Obtain the sample point dataset based on the key environmental factors, and divide the dataset into training set and test set. A spatial prediction model is selected, and the spatial prediction model is trained and tested based on the training set and the test set to obtain a spatial prediction model for echinococcosis. The study area is divided into several geographical units. The key environmental factor data of each geographical unit are input into the spatial prediction model of echinococcosis to obtain the incidence probability of echinococcosis in each geographical unit. Based on the incidence probability of echinococcosis, each geographical unit is divided into different risk levels, and a echinococcosis risk level planning map of the study area is generated.

[0007] Preferably, the multi-source environmental data includes: climate factor data, topographic factor data, vegetation factor data, soil factor data, and land use factor data; The climate factor data include annual average temperature, monthly average temperature range, annual precipitation, relative humidity, and sunshine duration. The topographic factor data includes: elevation, slope, aspect, and topographic relief. The vegetation factor data include: normalized vegetation index, enhanced vegetation index, and leaf area index; The soil factor data includes: soil type, soil pH value, and soil clay content; The land use factor data includes: land use type.

[0008] Preferably, the preprocessing further includes: imputing missing values ​​in the climate factor data using the average value of data from adjacent stations; performing One-Hot encoding on land use types in the land use factor data; and resampling the multi-source environmental data at the same spatial resolution.

[0009] Preferably, when using correlation analysis or principal component analysis to determine the correlation strength between various environmental factors and the incidence of alveolar echinococcosis, if correlation analysis is used, the correlation strength is determined according to the following formula: ; Where r represents the Pearson correlation coefficient, xi and yi represent the observed values ​​in the two datasets respectively, and xˉ and yˉ represent the means of the two datasets respectively; If principal component analysis is used, the correlation strength is determined according to the following formula: ; Where Loadingij represents the correlation strength, Zj represents the standardized score of the j-th principal component, sigmai represents the standard deviation of the i-th variable, and lambdaj represents the eigenvalue of the j-th principal component.

[0010] Preferably, key environmental factors are screened based on the correlation strength, including: A correlation strength threshold is set, and the correlation strength is compared with the correlation strength threshold. If the correlation strength is greater than or equal to the correlation strength threshold, the multi-source environmental data corresponding to the correlation strength is determined to be a key environmental factor. If the correlation strength is less than the correlation strength threshold, the multi-source environmental data corresponding to the correlation strength will be excluded.

[0011] Preferably, the training set and the test set are divided in a ratio of 7:3 or 8:2.

[0012] Preferably, a spatial prediction model is selected, and the spatial prediction model is trained and tested based on the training set and the test set to obtain a spatial prediction model for echinococcosis, including: training the spatial prediction model based on the training set and obtaining the model parameters of the spatial prediction model; The model parameters are tested based on the test set to obtain the evaluation index of the model parameters; The evaluation indicators are judged. If all the evaluation indicators reach the specified threshold, the model parameters are fitted to obtain a spatial prediction model for echinococcosis. Otherwise, the model parameters are retrained and optimized until all the evaluation indicators reach the specified threshold. The evaluation metrics include the area under the receiver operating characteristic curve, accuracy, recall, precision, and F1 score.

[0013] Preferably, the spatial prediction model is a geographic weighted regression model, and the model formula is: ; Where y represents the incidence probability of alveolar echinococcosis, β0(μi,νi) represents the intercept term at position (μi,νi), βk(μi,νi) represents the regression coefficient of the k-th environmental factor at position (μi,νi), xk represents the value of the k-th environmental factor, and ϵi represents the error term.

[0014] Preferably, each geographical unit is divided into different risk levels based on the incidence probability of alveolar echinococcosis, including: Set a first incidence probability threshold, a second incidence probability threshold, and a third incidence probability threshold, with the first incidence probability threshold, the second incidence probability threshold, and the third incidence probability threshold increasing sequentially; The risk level of each geographical unit is determined based on the relationship between the incidence probability of alveolar echinococcosis and the first incidence probability threshold, the second incidence probability threshold and the third incidence probability threshold. If the incidence probability of alveolar echinococcosis is less than the first incidence probability threshold, then the corresponding geographical unit is determined to be a risk-free area. If the incidence probability of alveolar echinococcosis is greater than or equal to the first incidence probability threshold, and the incidence probability of alveolar echinococcosis is less than the first incidence probability threshold, then the corresponding geographical unit is determined to be a low-risk area. If the incidence probability of alveolar echinococcosis is greater than or equal to the second incidence probability threshold, and the incidence probability of alveolar echinococcosis is less than the third incidence probability threshold, then the corresponding geographical unit is determined to be a medium-risk area. If the incidence rate of alveolar echinococcosis is greater than or equal to the third incidence rate threshold, then the corresponding geographical unit is determined to be a high-risk area.

[0015] This invention also discloses a spatial distribution prediction system for alveolar echinococcosis based on multiple environmental factors, used to apply the above-mentioned spatial distribution prediction method for alveolar echinococcosis based on multiple environmental factors, comprising: The data acquisition module is configured to acquire multi-source environmental data of echinococcosis cases within the study area and preprocess the multi-source environmental data, including data cleaning, outlier handling, and data standardization. The environmental factor screening module is configured to use correlation analysis or principal component analysis to determine the correlation strength between each environmental factor and the incidence of alveolar echinococcosis, and to screen out key environmental factors based on the correlation strength. The sample data acquisition module is configured to set sample points, mark vesicular echinococcosis case points as positive sample points, mark several non-vesicular echinococcosis case points in the same study area as negative sample points, acquire the dataset of sample points based on the key environmental factors, and divide the dataset into training set and test set. The prediction model building module is configured to select a spatial prediction model, train and test the spatial prediction model based on the training set and the test set, and obtain a spatial prediction model for echinococcosis. The spatial distribution prediction module is configured to divide the study area into several geographical units, input the key environmental factor data of each geographical unit into the spatial prediction model of echinococcosis, obtain the incidence probability of echinococcosis in each geographical unit, divide each geographical unit into different risk levels based on the incidence probability of echinococcosis, and generate a echinococcosis risk level planning map of the study area.

[0016] Compared with existing technologies, the advantages of this invention lie in its integration of multi-source environmental data, including climate, topography, vegetation, soil, and land use, to comprehensively cover the key environmental influencing factors of echinococcosis incidence. Compared with single-factor prediction, it significantly improves the comprehensiveness and accuracy of prediction. Key environmental factors are screened through correlation analysis or principal component analysis, eliminating irrelevant or weakly correlated factors, reducing data redundancy, lowering model computational complexity, and ensuring prediction accuracy. The use of a geographically weighted regression model as the spatial prediction model fully considers the spatial heterogeneity of environmental factors, accurately quantifies the incidence probability in different geographical units, and provides a reliable basis for risk level classification. Attached Figure Description

[0017] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.

[0018] Figure 1 This is a flowchart illustrating the spatial distribution prediction method for alveolar echinococcosis based on multiple environmental factors according to the present invention. Figure 2 This is a functional block diagram of the spatial distribution prediction system for alveolar echinococcosis based on multiple environmental factors according to the present invention. Detailed Implementation

[0019] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of this application.

[0020] like Figure 1 As shown, this invention provides a method for predicting the spatial distribution of alveolar echinococcosis based on multiple environmental factors, including: Multi-source environmental data of echinococcosis cases within the study area were acquired, and the multi-source environmental data were preprocessed, including data cleaning, outlier handling, and data standardization. Correlation analysis or principal component analysis was used to determine the correlation strength between each environmental factor and the incidence of alveolar echinococcosis, and key environmental factors were screened based on the correlation strength. Set up sample points, mark vesicular echinococcosis case points as positive sample points, and mark several non-vesicular echinococcosis case points in the same study area as negative sample points. Obtain the sample point dataset based on the key environmental factors, and divide the dataset into training set and test set. A spatial prediction model is selected, and the spatial prediction model is trained and tested based on the training set and the test set to obtain a spatial prediction model for echinococcosis. The study area is divided into several geographical units. The key environmental factor data of each geographical unit are input into the spatial prediction model of echinococcosis to obtain the incidence probability of echinococcosis in each geographical unit. Based on the incidence probability of echinococcosis, each geographical unit is divided into different risk levels, and a echinococcosis risk level planning map of the study area is generated.

[0021] This invention enables accurate understanding of the spatial distribution patterns of alveolar echinococcosis using multi-source environmental data. Comprehensive preprocessing of the multi-source environmental data ensures data quality and reliability, laying a solid foundation for subsequent analysis. The process of screening key environmental factors precisely focuses on factors significantly influencing the incidence of alveolar echinococcosis, avoiding interference from irrelevant information and improving the accuracy and efficiency of the model. Training and testing sets were constructed in the setting of sample points and the division of the dataset, allowing the spatial prediction model of alveolar echinococcosis to be fully trained and effectively validated. The spatial prediction model of alveolar echinococcosis obtained after training and testing exhibits high accuracy and practicality. Dividing the study area into geographical units and inputting key environmental factor data allows for the accurate determination of the incidence probability in each geographical unit, thereby classifying risk levels and generating risk level planning maps. This provides strong support for public health departments to formulate targeted prevention and control strategies, helps to rationally allocate prevention and control resources, improves the effectiveness and precision of prevention and control measures, thereby reducing the incidence of alveolar echinococcosis and protecting the health of the people.

[0022] In some embodiments of this application, the multi-source environmental data includes: climate factor data, topographic factor data, vegetation factor data, soil factor data, and land use factor data; wherein, the climate factor data includes annual average temperature, monthly average temperature range, annual precipitation, relative humidity, and sunshine duration; the topographic factor data includes: altitude, slope, aspect, and topographic relief; the vegetation factor data includes: normalized difference vegetation index, enhanced vegetation index, and leaf area index; the soil factor data includes: soil type, soil pH value, and soil clay content; and the land use factor data includes: land use type.

[0023] In this embodiment, climate factors are obtained from meteorological reanalysis databases (such as ERA5); topographic factors are extracted from digital elevation model (DEM, such as SRTM / ASTER GDEM) data; vegetation factors are calculated from remote sensing satellite data (such as Landsat, MODIS); soil factors are obtained from global soil databases (such as HWSD); and land use / cover is obtained from global land use datasets (such as Globeland30).

[0024] In some embodiments of this application, the preprocessing further includes: imputing missing values ​​in the climate factor data using the average value of data from adjacent stations; performing One-Hot encoding on land use types in the land use factor data; and resampling the multi-source environmental data at the same spatial resolution.

[0025] Understandably, meticulous and targeted preprocessing of multi-source environmental data further enhances its usability and accuracy. For imputation of missing values ​​in climate factor data, the method of averaging data from adjacent stations ensures data continuity and makes the data as close to reality as possible, avoiding analytical bias caused by missing data. One-Hot coding is applied to land use types in land use factor data, transforming categorical data into a numerical form that is easy for the model to process, enabling the model to better identify and analyze the relationship between different land use types and the incidence of echinococcosis. Resampling of multi-source environmental data at the same spatial resolution unifies the spatial scale of the data, eliminating the impact of differences in spatial resolution, and allowing different types of environmental data to be integrated and analyzed under the same standard. The comprehensive application of these preprocessing steps makes multi-source environmental data more reliable and effective in subsequent analysis and modeling stages. In the correlation analysis and key environmental factor screening process, high-quality data can more accurately reflect the true relationship between various environmental factors and the incidence of echinococcosis, thereby identifying more representative and influential key environmental factors.

[0026] In some embodiments of this application, when using correlation analysis or principal component analysis to determine the correlation strength between various environmental factors and the incidence of alveolar echinococcosis, if correlation analysis is used, the correlation strength is determined according to the following formula: ; Where r represents the Pearson correlation coefficient, xi and yi represent the observed values ​​in the two datasets respectively, and xˉ and yˉ represent the means of the two datasets respectively; If principal component analysis is used, the correlation strength is determined according to the following formula: ; Where Loadingij represents the correlation strength, Zj represents the standardized score of the j-th principal component, sigmai represents the standard deviation of the i-th variable, and lambdaj represents the eigenvalue of the j-th principal component.

[0027] Understandably, using correlation analysis or principal component analysis to determine the correlation strength between environmental factors and the incidence of alveolar echinococcosis lays a solid foundation for subsequent screening of key environmental factors and model construction. The Pearson correlation coefficient formula used in correlation analysis accurately measures the degree of linear correlation between two variables. By calculating this coefficient, the closeness of the association between each environmental factor and the incidence of alveolar echinococcosis can be clearly understood, thus identifying those environmental factors that have a significant impact on the incidence. Principal component analysis takes another approach, extracting principal components from the data and calculating the correlation strength. This method can transform multiple related environmental factors into a few unrelated principal components, thereby simplifying the data structure and reducing data dimensionality. Simultaneously, by determining the correlation strength between each environmental factor and the principal components, the importance and influence of each environmental factor in the entire dataset can be further clarified.

[0028] In this embodiment, Pearson correlation analysis was used to calculate the correlation coefficients between various environmental factors and the incidence of alveolar echinococcosis, with a correlation strength threshold of 0.3. The correlation coefficients for annual average temperature, altitude, normalized difference vegetation index, soil pH, and land use type were all greater than 0.3, and were therefore identified as key environmental factors; the remaining factors were excluded because their correlation coefficients were less than 0.3.

[0029] In some embodiments of this application, the key environmental factors are screened based on the correlation strength, including: setting a correlation strength threshold, comparing the correlation strength with the correlation strength threshold, and if the correlation strength is greater than or equal to the correlation strength threshold, then determining the multi-source environmental data corresponding to the correlation strength as a key environmental factor; if the correlation strength is less than the correlation strength threshold, then excluding the multi-source environmental data corresponding to the correlation strength.

[0030] Understandably, setting a correlation strength threshold to screen key environmental factors can effectively filter out environmental data with weak correlation to the incidence of alveolar echinococcosis, allowing subsequent analysis and modeling to focus more on key factors. This screening method can avoid interference from too much irrelevant or weakly correlated data on the model, thereby improving the model's accuracy and efficiency.

[0031] In some embodiments of this application, the training set and the test set are divided in a ratio of 7:3 or 8:2.

[0032] In some embodiments of this application, a spatial prediction model is selected, and the spatial prediction model is trained and tested based on the training set and the test set to obtain a spatial prediction model for alveolar echinococcosis. This includes: training the spatial prediction model based on the training set to obtain model parameters; testing the model parameters based on the test set to obtain evaluation metrics for the model parameters; judging the evaluation metrics; if all evaluation metrics reach a specified threshold, then fitting the model parameters to obtain a spatial prediction model for alveolar echinococcosis; otherwise, retraining and optimizing the model parameters until all evaluation metrics reach the specified threshold. The evaluation metrics include the area under the receiver operating characteristic (ROC) curve, accuracy, recall, precision, and F1 score.

[0033] Understandably, by appropriately dividing the training and testing sets and rigorously training and testing the spatial prediction model, a high-performance and reliable spatial prediction model for alveolar echinococcosis can be obtained. Dividing the training and testing sets in a 7:3 or 8:2 ratio ensures that the model has sufficient training data to learn features while also having suitable data to test the model's generalization ability.

[0034] Model parameters are acquired during training and then evaluated using a test set. The model's performance is judged based on evaluation metrics. Using multi-dimensional evaluation metrics such as the area under the receiver operating characteristic (ROC) curve, accuracy, recall, precision, and F1 score provides a comprehensive and objective reflection of model performance. Model parameter fitting is only performed when all evaluation metrics reach specified thresholds, ensuring that the final spatial prediction model for alveolar echinococcosis has high accuracy and reliability.

[0035] If the evaluation indicators fail to reach the specified thresholds, the model parameters are retrained and optimized. This iterative approach continuously improves the model's performance, allowing it to better adapt to actual data and prediction needs. The resulting spatial prediction model for alveolar echinococcosis can more accurately predict the spatial distribution of the disease, providing public health departments with more precise information. Based on the model's predictions, public health departments can prepare for prevention and control in advance, deploying preventative measures in areas with potentially high disease incidence, such as strengthening public education and conducting screening, thereby effectively reducing the incidence of alveolar echinococcosis.

[0036] In some embodiments of this application, the spatial prediction model is a geographic weighted regression model, and the model formula is: ; Where y represents the incidence probability of alveolar echinococcosis, β0(μi,νi) represents the intercept term at position (μi,νi), βk(μi,νi) represents the regression coefficient of the k-th environmental factor at position (μi,νi), xk represents the value of the k-th environmental factor, and ϵi represents the error term.

[0037] It is understandable that the geographically weighted regression model considers the heterogeneity of spatial location, assigning different regression coefficients to each location based on the characteristics of environmental factors in different geographical locations. This allows the model to more accurately capture the relationship between the incidence probability of alveolar echinococcosis and environmental factors, as the impact of environmental conditions on the disease may vary in different regions. The model formula clearly shows the specific degree of influence of each environmental factor on the incidence probability of alveolar echinococcosis, which helps in in-depth analysis of the disease's pathogenesis. The geographically weighted regression model can also visually demonstrate the characteristics and patterns of the spatial distribution of alveolar echinococcosis. Through the analysis and visualization of model parameters, it is possible to understand which regions have a greater impact on the incidence probability of the disease and which regions have a smaller impact. Public health departments can use this information to carry out targeted prevention and control work, improving the efficiency of prevention and control resource utilization.

[0038] In this embodiment, a geographically weighted regression model is selected for training. The model parameters are fitted based on the training set to obtain an initial model. The initial model is then tested using a test set, and evaluation metrics are calculated: the area under the receiver operating characteristic (ROC) curve is 0.89, the accuracy is 85%, the recall is 83%, the precision is 84%, and the F1 score is 83.5%. All metrics meet the specified thresholds (thresholds are set as follows: ROC curve ≥ 0.85, accuracy, recall, precision, and F1 score ≥ 80%). Therefore, no further optimization is needed, and this model is determined to be the final spatial prediction model for alveolar echinococcosis.

[0039] In some embodiments of this application, each geographical unit is divided into different risk levels based on the incidence probability of echinococcosis, including: setting a first incidence probability threshold, a second incidence probability threshold, and a third incidence probability threshold, wherein the first incidence probability threshold, the second incidence probability threshold, and the third incidence probability threshold increase sequentially; determining the risk level of each geographical unit according to the relationship between the incidence probability of echinococcosis and the first incidence probability threshold, the second incidence probability threshold, and the third incidence probability threshold; if the incidence probability of echinococcosis is less than the first incidence probability threshold, the corresponding geographical unit is determined to be a risk-free area; if the incidence probability of echinococcosis is greater than or equal to the first incidence probability threshold, and the incidence probability of echinococcosis is less than the first incidence probability threshold, the corresponding geographical unit is determined to be a low-risk area; if the incidence probability of echinococcosis is greater than or equal to the second incidence probability threshold, and the incidence probability of echinococcosis is less than the third incidence probability threshold, the corresponding geographical unit is determined to be a medium-risk area; if the incidence probability of echinococcosis is greater than or equal to the third incidence probability threshold, the corresponding geographical unit is determined to be a high-risk area.

[0040] Understandably, dividing geographical units into different risk levels allows public health departments to quickly and intuitively grasp the risk status of alveolar echinococcosis in various regions. For risk-free areas, the frequency and resource investment in routine monitoring can be appropriately reduced, concentrating more effort and resources on high-risk areas. For low-risk areas, a certain level of monitoring can be maintained, while strengthening health education for local residents to raise their awareness of prevention. For medium-risk areas, public health departments can implement targeted prevention and control measures, such as strengthening environmental management and conducting regular screening of key populations. This allows for the timely detection of potential sources of infection before a large-scale outbreak, enabling effective interventions to prevent further spread of the disease. High-risk areas are the key areas for prevention and control, requiring a significant concentration of human, material, and financial resources. Increased investment in local medical resources can improve diagnostic and treatment capabilities; strengthened control of disease vectors can reduce transmission routes; and various channels can be used to disseminate knowledge about alveolar echinococcosis prevention and control to local residents, improving their self-protection capabilities. This method of classifying risk levels based on the probability of incidence helps optimize the allocation of prevention and control resources, avoiding waste and misuse. This will make prevention and control efforts more targeted, improve the effectiveness of prevention and control, and minimize the threat of alveolar echinococcosis to public health.

[0041] In this embodiment, for example, the study area is divided into 1km × 1km geographical units, resulting in 5000 geographical units. Key environmental factor data for each geographical unit are input into a prediction model to obtain the incidence probability of each unit. A first incidence probability threshold of 0.1, a second incidence probability threshold of 0.3, and a third incidence probability threshold of 0.5 are set: geographical units with an incidence probability < 0.1 are considered risk-free areas (2100 units); geographical units with an incidence probability ≤ 0.1 < 0.3 are considered low-risk areas (1500 units); geographical units with an incidence probability ≤ 0.3 < 0.5 are considered medium-risk areas (1000 units); and geographical units with an incidence probability ≥ 0.5 are considered high-risk areas (400 units). Based on the above division results, a risk level planning map of echinococcosis in the study area is generated using GIS software, clearly presenting the risk distribution in different areas.

[0042] This invention also discloses a spatial distribution prediction system for alveolar echinococcosis based on multiple environmental factors, used to apply the above-mentioned spatial distribution prediction method for alveolar echinococcosis based on multiple environmental factors, comprising: The data acquisition module is configured to acquire multi-source environmental data of echinococcosis cases within the study area and preprocess the multi-source environmental data, including data cleaning, outlier handling, and data standardization. The environmental factor screening module is configured to use correlation analysis or principal component analysis to determine the correlation strength between each environmental factor and the incidence of alveolar echinococcosis, and to screen out key environmental factors based on the correlation strength. The sample data acquisition module is configured to set sample points, mark vesicular echinococcosis case points as positive sample points, mark several non-vesicular echinococcosis case points in the same study area as negative sample points, acquire the dataset of sample points based on the key environmental factors, and divide the dataset into training set and test set. The prediction model building module is configured to select a spatial prediction model, train and test the spatial prediction model based on the training set and the test set, and obtain a spatial prediction model for echinococcosis. The spatial distribution prediction module is configured to divide the study area into several geographical units, input the key environmental factor data of each geographical unit into the spatial prediction model of echinococcosis, obtain the incidence probability of echinococcosis in each geographical unit, divide each geographical unit into different risk levels based on the incidence probability of echinococcosis, and generate a echinococcosis risk level planning map of the study area.

[0043] This invention enables accurate prediction of the spatial distribution of alveolar echinococcosis by comprehensively considering multiple environmental factors. The data acquisition module collects and preprocesses multi-source environmental data, ensuring data quality and usability and providing a reliable foundation for subsequent analysis. The environmental factor screening module identifies key environmental factors, avoiding interference from irrelevant or redundant information, allowing the model to focus more on factors significantly impacting disease incidence. The sample data collection module rationally divides positive and negative sample points and constructs a dataset, providing effective data support for model training and testing. The prediction model building module, after training and testing, produces a spatial prediction model for alveolar echinococcosis with high accuracy and reliability. The spatial distribution prediction module divides the study area into geographical units and determines the probability of disease incidence and risk level; the generated risk level planning map visually displays the distribution of alveolar echinococcosis within the study area.

[0044] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program goods. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program goods embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0045] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program goods according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0046] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0047] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0048] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the specific implementation of the present invention. Any modifications or equivalent substitutions that do not depart from the spirit and scope of the present invention should be covered within the scope of protection of the claims of the present invention.

Claims

1. A method for predicting the spatial distribution of alveolar echinococcosis based on multiple environmental factors, characterized in that, include: Multi-source environmental data of echinococcosis cases within the study area were acquired, and the multi-source environmental data were preprocessed, including data cleaning, outlier handling, and data standardization. Correlation analysis or principal component analysis was used to determine the correlation strength between each environmental factor and the incidence of alveolar echinococcosis, and key environmental factors were screened based on the correlation strength. Set up sample points, mark vesicular echinococcosis case points as positive sample points, and mark several non-vesicular echinococcosis case points in the same study area as negative sample points. Obtain the sample point dataset based on the key environmental factors, and divide the dataset into training set and test set. A spatial prediction model is selected, and the spatial prediction model is trained and tested based on the training set and the test set to obtain a spatial prediction model for echinococcosis. The study area is divided into several geographical units. The key environmental factor data of each geographical unit are input into the spatial prediction model of echinococcosis to obtain the incidence probability of echinococcosis in each geographical unit. Based on the incidence probability of echinococcosis, each geographical unit is divided into different risk levels, and a echinococcosis risk level planning map of the study area is generated.

2. The method for predicting the spatial distribution of alveolar echinococcosis based on multiple environmental factors according to claim 1, characterized in that, The multi-source environmental data includes: climate factor data, topographic factor data, vegetation factor data, soil factor data, and land use factor data; The climate factor data include annual average temperature, monthly average temperature range, annual precipitation, relative humidity, and sunshine duration. The topographic factor data includes: elevation, slope, aspect, and topographic relief. The vegetation factor data include: normalized vegetation index, enhanced vegetation index, and leaf area index; The soil factor data includes: soil type, soil pH value, and soil clay content; The land use factor data includes: land use type.

3. The method for predicting the spatial distribution of alveolar echinococcosis based on multiple environmental factors according to claim 2, characterized in that, The preprocessing also includes: imputing missing values ​​in the climate factor data using the average value of data from adjacent stations; performing One-Hot encoding on land use types in the land use factor data; and resampling the multi-source environmental data at the same spatial resolution.

4. The method for predicting the spatial distribution of alveolar echinococcosis based on multiple environmental factors according to claim 1, characterized in that, When using correlation analysis or principal component analysis to determine the correlation strength between various environmental factors and the incidence of alveolar echinococcosis, if correlation analysis is used, the correlation strength is determined according to the following formula: ； Where r represents the Pearson correlation coefficient, xi and yi represent the observed values ​​in the two datasets respectively, and xˉ and yˉ represent the means of the two datasets respectively; If principal component analysis is used, the correlation strength is determined according to the following formula: ； Where Loadingij represents the correlation strength, Zj represents the standardized score of the j-th principal component, sigmai represents the standard deviation of the i-th variable, and lambdaj represents the eigenvalue of the j-th principal component.

5. The method for predicting the spatial distribution of alveolar echinococcosis based on multiple environmental factors according to claim 1, characterized in that, Key environmental factors were selected based on the correlation strength, including: A correlation strength threshold is set, and the correlation strength is compared with the correlation strength threshold. If the correlation strength is greater than or equal to the correlation strength threshold, the multi-source environmental data corresponding to the correlation strength is determined to be a key environmental factor. If the correlation strength is less than the correlation strength threshold, the multi-source environmental data corresponding to the correlation strength will be excluded.

6. The method for predicting the spatial distribution of alveolar echinococcosis based on multiple environmental factors according to claim 1, characterized in that, The training set and the test set are divided in a ratio of 7:3 or 8:

2.

7. The method for predicting the spatial distribution of alveolar echinococcosis based on multiple environmental factors according to claim 1, characterized in that, Selecting a spatial prediction model, training and testing the spatial prediction model based on the training set and test set to obtain a spatial prediction model for echinococcosis, including: training the spatial prediction model based on the training set and obtaining the model parameters of the spatial prediction model; The model parameters are tested based on the test set to obtain the evaluation index of the model parameters; The evaluation indicators are judged. If all the evaluation indicators reach the specified threshold, the model parameters are fitted to obtain a spatial prediction model for echinococcosis. Otherwise, the model parameters are retrained and optimized until all the evaluation indicators reach the specified threshold. The evaluation metrics include the area under the receiver operating characteristic curve, accuracy, recall, precision, and F1 score.

8. The method for predicting the spatial distribution of alveolar echinococcosis based on multiple environmental factors according to claim 1, characterized in that, The spatial prediction model is a geographic weighted regression model, and the model formula is: ； Where y represents the incidence probability of alveolar echinococcosis, β0(μi,νi) represents the intercept term at position (μi,νi), βk(μi,νi) represents the regression coefficient of the k-th environmental factor at position (μi,νi), xk represents the value of the k-th environmental factor, and ϵi represents the error term.

9. The method for predicting the spatial distribution of alveolar echinococcosis based on multiple environmental factors according to claim 1, characterized in that, Based on the incidence rate of alveolar echinococcosis, each geographical unit is divided into different risk levels, including: Set a first incidence probability threshold, a second incidence probability threshold, and a third incidence probability threshold, with the first incidence probability threshold, the second incidence probability threshold, and the third incidence probability threshold increasing sequentially; The risk level of each geographical unit is determined based on the relationship between the incidence probability of alveolar echinococcosis and the first incidence probability threshold, the second incidence probability threshold and the third incidence probability threshold. If the incidence probability of alveolar echinococcosis is less than the first incidence probability threshold, then the corresponding geographical unit is determined to be a risk-free area. If the incidence probability of alveolar echinococcosis is greater than or equal to the first incidence probability threshold, and the incidence probability of alveolar echinococcosis is less than the first incidence probability threshold, then the corresponding geographical unit is determined to be a low-risk area. If the incidence probability of alveolar echinococcosis is greater than or equal to the second incidence probability threshold, and the incidence probability of alveolar echinococcosis is less than the third incidence probability threshold, then the corresponding geographical unit is determined to be a medium-risk area. If the incidence rate of alveolar echinococcosis is greater than or equal to the third incidence rate threshold, then the corresponding geographical unit is determined to be a high-risk area.

10. A spatial distribution prediction system for alveolar echinococcosis based on multiple environmental factors, used to apply the spatial distribution prediction method for alveolar echinococcosis based on multiple environmental factors as described in any one of claims 1-9, characterized in that, include: The data acquisition module is configured to acquire multi-source environmental data of echinococcosis cases within the study area and preprocess the multi-source environmental data, including data cleaning, outlier handling, and data standardization. The environmental factor screening module is configured to use correlation analysis or principal component analysis to determine the correlation strength between each environmental factor and the incidence of alveolar echinococcosis, and to screen out key environmental factors based on the correlation strength. The sample data acquisition module is configured to set sample points, mark vesicular echinococcosis case points as positive sample points, mark several non-vesicular echinococcosis case points in the same study area as negative sample points, acquire the dataset of sample points based on the key environmental factors, and divide the dataset into training set and test set. The prediction model building module is configured to select a spatial prediction model, train and test the spatial prediction model based on the training set and the test set, and obtain a spatial prediction model for echinococcosis. The spatial distribution prediction module is configured to divide the study area into several geographical units, input the key environmental factor data of each geographical unit into the spatial prediction model of echinococcosis, obtain the incidence probability of echinococcosis in each geographical unit, divide each geographical unit into different risk levels based on the incidence probability of echinococcosis, and generate a echinococcosis risk level planning map of the study area.

Images