Animal epidemic disease cross-border transmission risk analysis method based on big data

Abstract

The invention discloses an animal epidemic disease cross-border transmission risk analysis method based on big data, and particularly relates to the technical field of data analysis. Through a multi-modal data fusion algorithm of deep learning, multi-source heterogeneous data is integrated in real time to form a unified data set, a cross-border propagation path model based on an infectious disease propagation model and an epidemiological theory is constructed, a propagation path of epidemic diseases is accurately simulated in combination with spatio-temporal information and propagation characteristics, and on the basis, the propagation path of the epidemic diseases is accurately simulated. A machine learning algorithm is utilized to analyze the fusion degree of cross-border propagation data between different regions, and based on a multi-source data fusion evaluation result, a high-risk region and a propagation speed are predicted in real time, so that potential cross-border propagation risks are evaluated, and targeted prevention and control intervention measures are proposed; according to the method, the problems of data isomerism, timeliness difference and fusion processing in the prior art are effectively solved, the accuracy, the real-time performance and the scientificity of epidemic disease cross-border transmission risk analysis are improved, and accurate support is provided for epidemic prevention and control decisions.

Description

Technical Field

[0001] The present invention relates to the field of data analysis technology, and in particular to a method for analyzing the risk of cross-border transmission of animal diseases based on big data. Background Art

[0002] With the continuous advancement of globalization, the cross-border spread of animal diseases has become an important issue in the field of global public health. Large-scale animal trade, personnel mobility, climate change and other factors have led to the rapid spread of animal diseases across national borders, bringing serious impacts on global agricultural production and ecosystems. Traditional animal disease monitoring systems focus on single monitoring points and local transmission chains, and are unable to comprehensively and dynamically predict and evaluate the risks of cross-border transmission.

[0003] In recent years, with the rapid development of big data technology, especially the widespread application of multi-dimensional data such as sensor technology, satellite remote sensing, social media information, and meteorological data, related fields have begun to try to use big data analysis methods to monitor and predict animal diseases. By integrating multi-source heterogeneous data, it is possible to obtain the epidemic trends, transmission paths, and potential transmission risks of animal diseases in real time. However, the existing cross-border transmission risk analysis methods have many shortcomings, mainly reflected in the complexity of data processing, data quality assurance, and risk prediction accuracy.

[0004] The prior art has the following deficiencies:

[0005] The cross-border spread of animal diseases involves a wide range of data types, including real-time epidemiological data from sensors and monitoring equipment, unstructured data from social media, news reports, and data from external environments such as meteorological satellites and shipping logistics. These data have significant differences in format, granularity, and timeliness. How to effectively integrate and process these multi-source heterogeneous data is a very challenging technical problem. Especially in cross-border transmission risk analysis, the timeliness and accuracy of data directly affect the effectiveness of risk prediction. If these data cannot be processed in real time, or if data from different sources cannot be effectively cleaned and connected, the analysis results may produce large errors, affecting the implementation of epidemic prevention and control measures. Summary of the invention

[0006] The purpose of the present invention is to provide a method for analyzing the risk of cross-border transmission of animal diseases based on big data to address the shortcomings of the background technology.

[0007] In order to achieve the above-mentioned purpose, the present invention provides the following technical solution: a method for analyzing the risk of cross-border transmission of animal diseases based on big data, comprising:

[0008] Real-time acquisition of heterogeneous data including epidemiological monitoring data, social media information, meteorological data, and shipping logistics information. Using a multimodal data fusion algorithm based on deep learning, the heterogeneous data from different sources are integrated to form a unified data set.

[0009] Construct a cross-border transmission path model based on infectious disease transmission models and epidemiological theories, combine spatiotemporal information and transmission characteristics in big data, and simulate the potential transmission path of animal diseases;

[0010] In the simulation process, machine learning algorithms are used to perform fusion processing and analysis on the integrated heterogeneous data to evaluate the degree of fusion of cross-border heterogeneous data between different regions;

[0011] Based on the fusion evaluation results of multi-source data, real-time predictions are made on high-risk areas and transmission speeds of cross-border transmission of animal diseases. Based on the prediction results, the potential cross-border transmission risks are judged and corresponding prevention and control intervention measures are proposed.

[0012] Preferably, the multimodal data fusion algorithm adopts a deep learning model, including but not limited to a convolutional neural network, a recurrent neural network, a multi-layer perceptron or a Transformer model, to fuse multi-source heterogeneous data through feature extraction and an attention mechanism.

[0013] Preferably, the constructed cross-border transmission path model is based on the improved SIR infectious disease transmission model, and combined with the GIS geographic information system to dynamically simulate the disease transmission path under different time and space conditions.

[0014] Preferably, the migration paths and migration times of different animal species are analyzed to generate a species migration anomaly index, and the method for obtaining the species migration anomaly index is:

[0015] Suppose there is a directed graph G=(V,E), where V represents the set of all habitats, are all habitat nodes, E represents the set of migration paths between habitats, From the habitat To the habitat migration paths, Represents edge The weight of arrive the intensity of migration;

[0016] Set the PageRank values ​​of all nodes to be equal initially, that is, , where n is the number of habitats; the iterative formula for the PageRank value is: ;in: It is a habitat The PageRank value of the page; α is the damping factor, which indicates the probability of the browser jumping to a random node. Pointing to the habitat The collection of all habitats; From the habitat The set of all migration paths sent out; From the habitat arrive The migration intensity of each habitat is initialized, and the PageRank value of each habitat is updated according to the iterative formula of the PageRank value. The update is repeated until the PageRank value converges. Once the PageRank value of each habitat is calculated, the PageRank mean μ and standard deviation σ of all habitats are calculated, and the species migration anomaly index is calculated. The expression is: ; Where AE is the species migration anomaly index.

[0017] Preferably, the differences in logistics flow in different regions are analyzed to generate a logistics flow difference index, and the method for obtaining the logistics flow difference index is:

[0018] Obtain the cargo inflow and outflow volume Lij, transportation times Li and total transportation volume Q of a region within different time ranges; calculate the mean of the logistics flow of all regions, that is, the average level of logistics flow of all regions, expressed as: ;in: is the mean of the logistics flow in all regions, m is the total number of regions, and the standard deviation of the logistics flow measures the fluctuation of the logistics flow, which is expressed as: ;in: is the standard deviation of logistics flow, is the total logistics flow in region i, is the mean of the logistics flow in all regions; it is obtained by calculating the difference between the logistics flow in each region and the mean of the overall flow, and the expression is: ; Where: DI is the logistics flow difference index.

[0019] Preferably, the species migration anomaly index and the logistics flow difference index are converted into comprehensive feature vectors, and the comprehensive feature vectors are used as inputs of the machine learning model. The machine learning model uses each group of comprehensive feature vectors to predict the fusion degree value label of cross-border heterogeneous data between different regions as the prediction target, and takes minimizing the sum of prediction errors of the fusion degree value labels of all cross-border heterogeneous data between different regions as the training target. The machine learning model is trained until the sum of prediction errors converges and the model training is stopped. The fusion degree value of cross-border heterogeneous data between different regions is determined according to the model output results, wherein the machine learning model is a polynomial regression model.

[0020] Preferably, the obtained fusion degree value of cross-border heterogeneous data between different regions is compared with a preset threshold. If the fusion degree value of the cross-border heterogeneous data is greater than or equal to the preset threshold, it means that the fusion degree of the cross-border heterogeneous data is high, and a normal data fusion signal is generated. If the fusion degree value of the cross-border heterogeneous data is less than the preset threshold, it means that the fusion degree of the cross-border heterogeneous data is low, and a data fusion abnormal signal is generated.

[0021] Preferably, the speed at which an animal disease spreads from one area to another is defined as the transmission speed S, the high-risk area is marked as Ahigh, indicating an area with a high risk of transmission, and the infection probability of the high-risk area is defined as R;

[0022] By integrating multi-source data, we obtain the fusion degree value LR of cross-border transmission heterogeneous data between different regions. We will use a mathematical formula based on the infectious disease transmission model to predict the transmission speed and risk.

[0023] The formula for calculating the propagation speed is: ;in: Indicates the infectious capacity of the source area. Indicates the susceptibility of the receiving area;

[0024] The calculation expression of cross-border transmission risk R is: ;in: represents the propagation probability of the source region, represents the input probability of receiving the area;

[0025] The judgment of high-risk areas is based on the predicted value of risk R. If the R value of a region exceeds the preset threshold Trisk, the region is considered a high-risk area: Once the high-risk area Ahigh and the transmission speed S are obtained, the potential cross-border transmission risk is further judged. The potential cross-border transmission risk is evaluated by the formula: ;in: Indicates potential cross-border transmission risk, which is the risk of transmission in all high-risk areas The accumulation of potential risks , and proposed different prevention and control intervention measures.

[0026] In the above technical solution, the technical effects and advantages provided by the present invention are:

[0027] 1. The present invention effectively solves the problems of data heterogeneity, timeliness and accurate prediction in the risk analysis of cross-border transmission of animal diseases through multi-source data fusion, deep learning modeling and infectious disease transmission analysis. A multimodal data fusion algorithm is used to integrate heterogeneous data such as epidemiological monitoring data, social media information, meteorological data, shipping logistics information, etc. to form a high-quality data set to ensure the comprehensiveness and timeliness of the information. By combining the improved SIR infectious disease transmission model with GIS technology, the cross-border transmission path of the disease is dynamically simulated to improve the accuracy of the prediction of the epidemic transmission trend. Furthermore, based on the machine learning algorithm, the species migration anomaly index and the logistics flow difference index are calculated to quantify the degree of fusion of data between different regions. Finally, the polynomial regression model is used to realize the intelligent assessment of cross-border transmission risks, and corresponding prevention and control measures are proposed according to the risk level.

[0028] 2. The present invention calculates the disease transmission speed and cross-border transmission risk index in real time. The method can actively generate data fusion normal or abnormal signals, quickly discover potential risks in the early stage of the epidemic, and take precise prevention and control measures, such as regional blockade, quarantine enhancement, vaccination, etc., to reduce the risk of epidemic spread. Overall, the present invention significantly improves the intelligence, automation and scientific level of cross-border animal disease transmission risk assessment, and provides a strong guarantee for public health security. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present invention. For ordinary technicians in this field, other drawings can also be obtained based on these drawings.

[0030] Figure 1 The present invention is a flow chart of the method. DETAILED DESCRIPTION

[0031] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0032] For examples, see Figure 1 As shown, the method for analyzing the risk of cross-border transmission of animal diseases based on big data in this embodiment includes:

[0033] Real-time acquisition of heterogeneous data including epidemiological monitoring data, social media information, meteorological data, and shipping logistics information. Using a multimodal data fusion algorithm based on deep learning, the heterogeneous data from different sources are integrated to form a unified data set.

[0034] Construct a cross-border transmission path model based on infectious disease transmission models and epidemiological theories, combine spatiotemporal information and transmission characteristics in big data, and simulate the potential transmission path of animal diseases;

[0035] In the simulation process, machine learning algorithms are used to perform fusion processing and analysis on the integrated heterogeneous data to evaluate the degree of fusion of cross-border heterogeneous data between different regions;

[0036] Based on the fusion evaluation results of multi-source data, real-time predictions are made on high-risk areas and transmission speeds of cross-border transmission of animal diseases. Based on the prediction results, the potential cross-border transmission risks are judged and corresponding prevention and control intervention measures are proposed.

[0037] The specific steps for real-time acquisition and integration of heterogeneous data include:

[0038] Through various sensors, monitoring equipment and public reports from health organizations, epidemiological data such as animal disease case data, epidemic area information, transmission speed, infection status, etc. are obtained in real time.

[0039] Natural language processing (NLP) technology is used to capture public discussions, posts, news reports, etc. related to animal diseases from social media platforms (such as Twitter, Facebook, Weibo, etc.) to extract epidemic-related information (such as source of infection, epidemic area, transmission route, etc.).

[0040] Obtain meteorological data related to the spread of animal diseases, including temperature, humidity, precipitation, wind speed and other meteorological parameters. These meteorological factors may affect the spread of diseases or the occurrence of emergencies.

[0041] Obtaining information on cross-border animal trade flows, transport routes, cargo types, transport timeliness, etc. from shipping companies, logistics platforms, customs, border monitoring and other channels is crucial for predicting the spread of animal diseases.

[0042] Structured data (such as sensor data, meteorological data, logistics information, etc.) usually has a fixed format, and the accuracy of the data is ensured through methods such as data cleaning, denoising, outlier processing, and data standardization. Unstructured data (such as social media information, news reports, etc.) needs to be processed using text mining technology, sentiment analysis, keyword extraction, etc. to convert the information into structured data and extract useful information related to animal diseases.

[0043] According to the timeliness requirements of the data, real-time data is timestamped to ensure that all data sources are synchronized within a unified time window to avoid analysis errors caused by timeliness differences.

[0044] Choose the right deep learning algorithm: Use a multimodal data fusion algorithm based on deep learning (such as multi-layer perceptron (MLP), convolutional neural network (CNN) or recurrent neural network (RNN)) to design a fusion model. Convolutional neural network (CNN): Suitable for processing data with spatial relationships, such as meteorological data or geographic location-related epidemic data. Recurrent neural network (RNN): Suitable for processing time series data, such as real-time epidemic spread data, shipping logistics data, meteorological data, etc. Transformer model: Suitable for parallel processing of multimodal data, capable of processing multiple time series data and complex multi-source heterogeneous data sets.

[0045] By encoding data from different data sources, different types of heterogeneous data such as epidemiological data, social media information, meteorological data, and logistics information are converted into a unified embedding space, ensuring that various types of data can be effectively compared and integrated in the same feature space.

[0046] Use the feature extraction layer in the deep learning model to extract key features (such as the outbreak area, transmission speed, weather changes, etc.) from each data source and align them to ensure that data from different sources can effectively correspond.

[0047] Through deep learning methods such as weighting, splicing, and attention mechanisms, the features extracted from different data sources are fused. Especially when facing heterogeneous data sources such as meteorology, social media, and logistics, multi-head attention mechanisms or weighted fusion algorithms are used to automatically adjust the importance of different data sources.

[0048] The fused multimodal features are integrated into a unified dataset (usually in tensor or vector format), where each data item contains comprehensive information from different sources, providing input for subsequent model training and analysis.

[0049] The fused data set is evaluated for quality, including data accuracy, completeness, consistency, etc. The high quality of the data set is ensured by calculating indicators such as the correlation between data sources and information redundancy.

[0050] The effect of the fusion algorithm is verified through cross-validation and test sets, and the improvement of model performance by data fusion is evaluated to ensure that the fused data can accurately reflect the transmission pattern of animal diseases.

[0051] To ensure the real-time and accuracy of the data set, a dynamic update mechanism is used to obtain the latest epidemiological monitoring data, social media information, meteorological data and shipping logistics information in real time, and to adjust and optimize the deep learning model in real time through online learning methods (such as incremental learning or online training). Through feedback and monitoring of prediction results, the deep learning model and data fusion algorithm are continuously improved to make them more adaptable to changes in the epidemic and new data sources.

[0052] Construct a cross-border transmission path model based on infectious disease transmission model and epidemiological theory, specifically:

[0053] SIR model: The classic SIR (susceptible-infectious-recovered) model is suitable for describing the spread of infectious diseases in a population, where: S (susceptible): refers to uninfected individuals who may be infected by contact with infected people. I (infected): an infected individual who can spread the disease. R (recovered): refers to individuals who have recovered and are immune or dead.

[0054] In the cross-border transmission path model, the improved SIR model or its variants (such as SEIR, SIS models, etc.) are used to consider more complex transmission processes, such as incubation period, loss of immunity and other factors.

[0055] Taking into account the geographical transmission characteristics of the disease, diffusion models (such as Laplacian model, random walk model, contagion model, etc.) are used to analyze the transmission paths between different regions.

[0056] Based on the theory of transmission dynamics in epidemiology, analyze how diseases spread across borders through different transmission pathways (such as animal trade, traffic flow, human activities, etc.).

[0057] Basic reproduction number R0: The concept of basic reproduction number (R0) is introduced to calculate the disease transmission potential of each region. An R0 value greater than 1 indicates that the disease is likely to spread in the region, and vice versa. Based on the changes in the R0 value of cross-border transmission paths, high-risk areas can be identified.

[0058] Consider the geographic, climatic, socioeconomic and cultural characteristics of different regions, and the characteristics that affect the spread of animal diseases. For example, climate change may affect the seasonality of disease transmission, and trade flows and population gatherings may affect the speed and scope of disease transmission.

[0059] Use spatiotemporal information in big data, such as meteorological data, logistics flow, animal movement, and epidemic history, to enhance the spatiotemporal dimension of the model. Use big data to model epidemic trends in different regions, time, and space.

[0060] GIS technology is used to analyze the geographical transmission paths of animal diseases, and disease information is combined with geographic coordinates, boundary data, trade routes, etc. to visualize and dynamically simulate cross-border transmission paths.

[0061] Construct a network model of disease transmission, and predict the cross-border transmission path of the disease by analyzing the transmission relationship between various nodes (such as countries, regions, and animal populations). Analyze the transmission risk and key transmission nodes of each region through a weighted graph structure.

[0062] Combine meteorological data (such as temperature, humidity, precipitation, etc.) and external environment (such as information on key logistics hubs such as seaports and airports) to build a cross-border transmission model. External environmental data helps identify climatic conditions that are prone to transmission, animal transportation routes, and key transmission nodes.

[0063] By establishing a cross-border transmission path model, combined with spatiotemporal data and transmission characteristics, the transmission path of animal diseases in different time periods and different geographical regions is simulated. Multi-objective optimization methods (such as particle swarm optimization and genetic algorithm) are used to optimize the transmission path, and the possible routes of disease transmission are predicted by combining actual transportation flow, trade flow, population flow and other factors.

[0064] Using infectious disease transmission models and big data, we simulate the speed and direction of the epidemic spread. Based on the interconnections between various nodes, we predict the spread of the epidemic and continuously adjust the prediction results of the transmission path based on real-time data.

[0065] Animal migration behavior (such as migration routes, cycles, habitat changes, etc.) has an important impact on the spread of animal diseases, especially in cross-border transmission, where migrating animals may become new sources of transmission. Use global animal migration databases (such as animal GPS tracking data, ecological monitoring data, etc.) to obtain the migration routes and migration times of different animal species. Integrate animal activity data and disease transmission rate information to establish a correlation model between animal migration and disease transmission. Analyze the migration patterns of different regions and species to assess their transmission risks in different seasons and climate changes.

[0066] Use global animal migration databases (such as animal GPS tracking data, ecological monitoring data, etc.) to obtain the migration paths and migration times of different animal species. Integrate animal activity data and disease transmission rate information to establish a correlation model between animal migration and disease transmission. Analyze the migration patterns of different regions and species to assess their transmission risks in different seasons and climate changes.

[0067] The species migration anomaly index is generated by analyzing the migration paths and migration times of different animal species. The method for obtaining the species migration anomaly index is as follows:

[0068] Define the graph structure, nodes: habitats, migration points, or ecological blocks. Edges: the migration paths of species from one node (habitat) to another node (habitat). Edge weights: The weights of edges can be defined based on factors such as migration flow, number of animals, and length of migration time. The larger the weight, the more frequent the migration activity, and vice versa. Suppose there is a directed graph G=(V,E), where: V represents the set of all habitats, are all habitat nodes, E represents the set of migration paths between habitats, From the habitat To the habitat migration paths, Represents edge The weight of arrive intensity of migration.

[0069] The calculation of PageRank value is based on the random walk model, assuming that a random "browser" jumps from one habitat to another with a certain probability. The core idea of ​​the algorithm is that the importance of each habitat comes from the "recommendation" of its neighboring nodes (migration paths). The PageRank values ​​of all nodes are set to be equal at the beginning, that is, , where n is the number of habitats; the iterative formula for the PageRank value is: ;in: It is a habitat PageRank value; α is the damping factor (usually set to 0.15), which indicates the probability of a browser jumping to a random node. Pointing to the habitat The collection of all habitats (i.e. habitats); From the habitat The set of all migration paths sent out (i.e. from Departure habitat); From the habitat arrive intensity of migration.

[0070] Initialize the PageRank value of each habitat.

[0071] The PageRank value of each habitat is updated according to the iterative formula of PageRank value.

[0072] Repeat the update until the PageRank value converges (that is, the change is less than the set threshold).

[0073] Once the PageRank values ​​of each habitat are calculated, migration anomalies can be evaluated based on these values. The PageRank mean μ and standard deviation σ of all habitats are calculated, and the species migration anomaly index is calculated as follows: ; In the formula, AE is the species migration abnormality index. The larger the value, the more abnormal the migration behavior of the habitat is and the more it deviates from the normal migration pattern.

[0074] The flow patterns and frequencies of cross-border logistics flow data (including animal trade, veterinary drugs, feed transportation, etc.) are not only closely related to the speed and path of the spread of the epidemic, but also reflect the risk of disease transmission during trade and transportation. There are significant differences in logistics flows in different regions, especially the heterogeneity of factors such as different trade agreements, customs control measures, and cross-border transportation methods, which also need to be carefully captured and analyzed.

[0075] By collecting trade and logistics flow data from different countries (such as customs data, international trade data, transportation route information, etc.), we can analyze the disease transmission paths during cross-border transportation. We can obtain different types of cross-border logistics data (such as animal trade, veterinary product transportation, etc.), analyze their spatiotemporal patterns, and assess which logistics routes become high-risk channels for disease transmission. We can also consider the impact of transportation timeliness, transportation density, and heterogeneity of transportation routes on the spread of the epidemic.

[0076] After analyzing the differences in logistics flow in different regions, a logistics flow difference index is generated. The method for obtaining the logistics flow difference index is as follows:

[0077] Obtain the cargo inflow and outflow volume Lij, transportation times Li and total transportation volume Q of a region within different time ranges; calculate the mean of the logistics flow of all regions, that is, the average level of logistics flow of all regions, expressed as: ;in: is the mean of the logistics flow in all regions, and m is the total number of regions. The standard deviation of logistics flow measures the fluctuation of logistics flow. By calculating the standard deviation, we can understand the dispersion of logistics flow in each region relative to the average level. The expression is: ;in: is the standard deviation of logistics flow, is the total logistics flow in region i. It is the average of logistics flow in all regions.

[0078] The logistics flow difference index is an indicator used to indicate the degree of difference in logistics flow in different regions. It is obtained by calculating the difference in logistics flow in each region relative to the overall flow mean and standardizing it. The expression is: ; Where: DI is the logistics flow difference index.

[0079] High Difference Index Region: If DI>2 (or other thresholds), it can be considered that there are significant abnormal fluctuations in logistics flow in the region, which may be caused by logistics bottlenecks, demand surges, traffic disruptions, etc. Low Difference Index Region: If DI is close to zero, it means that the logistics flow in the region is similar to that in other regions and the flow is relatively stable.

[0080] The species migration anomaly index and the logistics flow difference index are converted into comprehensive feature vectors, and the comprehensive feature vectors are used as the input of the machine learning model. The machine learning model takes each group of comprehensive feature vectors to predict the fusion degree value label of cross-border heterogeneous data between different regions as the prediction target, and takes minimizing the sum of prediction errors of the fusion degree value labels of all cross-border heterogeneous data between different regions as the training target. The machine learning model is trained until the sum of prediction errors converges and the model training is stopped. The fusion degree value of cross-border heterogeneous data between different regions is determined according to the model output results, wherein the machine learning model is a polynomial regression model.

[0081] The method for obtaining the fusion degree value of heterogeneous data transmitted across different regions is to obtain the corresponding function expression from the comprehensive feature vector training data of the trained machine learning model: ; In the formula, is the output function of the model, AE is the species migration anomaly index, DI is the logistics flow difference index, It is the fusion degree value of heterogeneous data transmitted across borders between different regions.

[0082] The obtained fusion degree value of cross-border heterogeneous data between different regions is compared with the preset threshold. If the fusion degree value of cross-border heterogeneous data is greater than or equal to the preset threshold, it means that the fusion degree of cross-border heterogeneous data is high, and a normal data fusion signal is generated. If the fusion degree value of cross-border heterogeneous data is less than the preset threshold, it means that the fusion degree of cross-border heterogeneous data is low, and a data fusion abnormal signal is generated.

[0083] Based on the fusion evaluation results of multi-source data, the high-risk areas and transmission speed of cross-border transmission of animal diseases are predicted in real time. According to the prediction results, the potential cross-border transmission risks are judged and corresponding prevention and control intervention measures are proposed, specifically:

[0084] The speed at which an animal disease spreads from one area to another is defined as the spread speed S.

[0085] The high-risk area is marked as Ahigh, indicating an area with a high risk of transmission, which usually requires priority prevention and control. The infection probability of the high-risk area is defined as R.

[0086] The fusion degree value LR of cross-border transmission heterogeneous data between different regions is obtained through the fusion of multi-source data (including epidemiological data, social media information, meteorological data, logistics flow, etc.). A mathematical formula based on the infectious disease transmission model will be used to predict the transmission speed and risk.

[0087] The transmission speed S is mainly determined by two factors: the transmission ability of the source of infection and the susceptibility of the receiving area. The formula for calculating the transmission speed is: ;in: It indicates the infectious capacity of the source area of ​​transmission (such as the outbreak site), which usually depends on the infection rate, epidemic characteristics and animal density in the source area. Indicates the susceptibility of the receiving area, which usually depends on the sanitary conditions, immunization levels and animal quarantine measures in the area.

[0088] The calculation expression of cross-border transmission risk R is: ;in: Represents the probability of transmission in the source area, which usually depends on the severity of the epidemic in the source area (e.g., the number of infected people, transmission rate, etc.). Represents the input probability of the receiving area, which usually depends on the disease exposure probability of the receiving area and reflects the mobility between the area and the source area (e.g., transportation flow, personnel flow, etc.).

[0089] The judgment of high-risk areas is based on the predicted value of risk R. If the R value of a region exceeds the preset threshold Trisk, the region is considered a high-risk area:

[0090] Once the high-risk area Ahigh and the transmission speed S are obtained, the potential cross-border transmission risk can be further determined and corresponding prevention and control intervention measures can be proposed.

[0091] The potential risk of cross-border transmission is assessed using the formula: ;in: Indicates potential cross-border transmission risk, which is the risk of transmission in all high-risk areas The accumulation of.

[0092] Based on potential risk , different prevention and control intervention measures can be proposed, such as:

[0093] Blockade and isolation of high-risk areas: For high-risk areas with a higher transmission rate, prevention and control measures such as blockade and isolation can be taken to restrict the movement of animals and reduce the spread of the virus.

[0094] Strengthen quarantine and surveillance: Strengthen quarantine measures and surveillance in high-risk areas and potential transmission routes to ensure timely detection and handling of potential outbreaks.

[0095] Vaccination and Immunization: Carry out large-scale animal vaccination in high-risk areas to increase immunity levels and reduce the risk of animal infection.

[0096] The above formulas are all dimensionless and numerical calculations. The formula is a formula for the most recent real situation obtained by collecting a large amount of data and performing software simulation. The preset parameters in the formula are set by technicians in this field according to actual conditions.

[0097] The above embodiments can be implemented in whole or in part by software, hardware, firmware or any other combination. When implemented by software, the above embodiments can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, the process or function described in the embodiment of the present application is generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the computer instructions can be transmitted from one website site, computer, server or data center to another website site, computer, server or data center by wired or wireless (e.g., infrared, wireless, microwave, etc.). The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that contains one or more available media sets. The available medium can be a magnetic medium (e.g., a floppy disk, a hard disk, a tape), an optical medium (e.g., a DVD), or a semiconductor medium. The semiconductor medium can be a solid-state hard disk.

[0098] It should be understood that the term "and / or" in this article is only a description of the association relationship of associated objects, indicating that there may be three relationships. For example, A and / or B can represent: A exists alone, A and B exist at the same time, and B exists alone, where A and B can be singular or plural. In addition, the character " / " in this article generally indicates that the objects associated with each other are in an "or" relationship, but it may also indicate an "and / or" relationship, which can be understood by referring to the context. A person of ordinary skill in the art can appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed in this article can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to exceed the scope of this application.

[0099] The above description is only a specific implementation manner of the present application, but the protection scope of the present application is not limited thereto. Any technician familiar with the technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application.

Claims

1. A risk analysis method for cross-border transmission of animal diseases based on big data, characterized by: include: Real-time acquisition of heterogeneous data including epidemiological monitoring data, social media information, meteorological data, and shipping logistics information. Using a multimodal data fusion algorithm based on deep learning, the heterogeneous data from different sources are integrated to form a unified data set. Construct a cross-border transmission path model based on infectious disease transmission models and epidemiological theories, combine spatiotemporal information and transmission characteristics in big data, and simulate the potential transmission path of animal diseases; In the simulation process, machine learning algorithms are used to perform fusion processing and analysis on the integrated heterogeneous data to evaluate the degree of fusion of cross-border heterogeneous data between different regions; Based on the fusion evaluation results of multi-source data, real-time predictions are made on high-risk areas and transmission speeds of cross-border transmission of animal diseases. Based on the prediction results, the potential cross-border transmission risks are judged and corresponding prevention and control intervention measures are proposed.

2. The method for analyzing cross-border transmission risk of animal diseases based on big data according to claim 1 is characterized in that: The multimodal data fusion algorithm adopts a deep learning model, including but not limited to a convolutional neural network, a recurrent neural network, a multi-layer perceptron or a Transformer model, to fuse multi-source heterogeneous data through feature extraction and an attention mechanism.

3. The method for analyzing cross-border transmission risk of animal diseases based on big data according to claim 2 is characterized in that: The constructed cross-border transmission path model is based on the improved SIR infectious disease transmission model, and combined with the GIS geographic information system to dynamically simulate the disease transmission path under different time and space conditions.

4. The method for analyzing cross-border transmission risk of animal diseases based on big data according to claim 3 is characterized in that: The species migration anomaly index is generated by analyzing the migration paths and migration times of different animal species. The method for obtaining the species migration anomaly index is as follows: Suppose there is a directed graph G=(V,E), where V represents the set of all habitats, are all habitat nodes, E represents the set of migration paths between habitats, From the habitat To the habitat migration paths, Represents edge The weight of arrive the intensity of migration; Set the PageRank values ​​of all nodes to be equal initially, that is, , where n is the number of habitats; the iterative formula for the PageRank value is: ;in: It is a habitat The PageRank value of the page; α is the damping factor, which indicates the probability of the browser jumping to a random node. Pointing to the habitat The collection of all habitats; From the habitat The set of all migration paths sent out; From the habitat arrive The migration intensity of each habitat is initialized, and the PageRank value of each habitat is updated according to the iterative formula of the PageRank value. The update is repeated until the PageRank value converges. Once the PageRank value of each habitat is calculated, the PageRank mean μ and standard deviation σ of all habitats are calculated, and the species migration anomaly index is calculated. The expression is: ; Where AE is the species migration anomaly index.

5. The method for analyzing cross-border transmission risk of animal diseases based on big data according to claim 4 is characterized in that: After analyzing the differences in logistics flow in different regions, a logistics flow difference index is generated. The method for obtaining the logistics flow difference index is as follows: Obtain the cargo inflow and outflow volume Lij, transportation times Li and total transportation volume Q of a region within different time ranges; calculate the mean of the logistics flow of all regions, that is, the average level of logistics flow of all regions, expressed as: ;in: is the mean of the logistics flow in all regions, m is the total number of regions, and the standard deviation of the logistics flow measures the fluctuation of the logistics flow, which is expressed as: ;in: is the standard deviation of logistics flow, is the total logistics flow in region i, is the mean of the logistics flow in all regions; it is obtained by calculating the difference between the logistics flow in each region and the mean of the overall flow, and the expression is: ; Where: DI is the logistics flow difference index.

6. The method for analyzing cross-border transmission risk of animal diseases based on big data according to claim 5 is characterized in that: The species migration anomaly index and the logistics flow difference index are converted into comprehensive feature vectors, and the comprehensive feature vectors are used as the input of the machine learning model. The machine learning model takes each group of comprehensive feature vectors to predict the fusion degree value label of cross-border heterogeneous data between different regions as the prediction target, and takes minimizing the sum of prediction errors of the fusion degree value labels of all cross-border heterogeneous data between different regions as the training target. The machine learning model is trained until the sum of prediction errors converges and the model training is stopped. The fusion degree value of cross-border heterogeneous data between different regions is determined according to the model output results, wherein the machine learning model is a polynomial regression model.

7. The method for analyzing cross-border transmission risk of animal diseases based on big data according to claim 6 is characterized in that: The obtained fusion degree value of the cross-border heterogeneous data between different regions is compared with the preset threshold. If the fusion degree value of the cross-border heterogeneous data is greater than or equal to the preset threshold, it means that the fusion degree of the cross-border heterogeneous data is high, and a data fusion normal signal is generated at this time; If the fusion degree value of cross-border heterogeneous data is less than the preset threshold, it means that the fusion degree of cross-border heterogeneous data is low, and a data fusion anomaly signal is generated.

8. The method for analyzing cross-border transmission risk of animal diseases based on big data according to claim 7 is characterized in that: The speed at which animal diseases spread from one area to another is defined as the transmission speed S, the high-risk area is marked as Ahigh, indicating an area with a high risk of transmission, and the infection probability of the high-risk area is defined as R; By integrating multi-source data, we obtain the fusion degree value LR of cross-border transmission heterogeneous data between different regions. We will use a mathematical formula based on the infectious disease transmission model to predict the transmission speed and risk. The formula for calculating the propagation speed is: ;in: Indicates the infectious capacity of the source area. Indicates the susceptibility of the receiving area; The calculation expression of cross-border transmission risk R is: ;in: represents the propagation probability of the source region, represents the input probability of receiving the area; The judgment of high-risk areas is based on the predicted value of risk R. If the R value of a region exceeds the preset threshold Trisk, the region is considered a high-risk area: Once the high-risk area Ahigh and the transmission speed S are obtained, the potential cross-border transmission risk is further judged. The potential cross-border transmission risk is evaluated by the formula: ;in: Indicates potential cross-border transmission risk, which is the risk of transmission in all high-risk areas The accumulation of potential risks , and proposed different prevention and control intervention measures.

Images