Method, device, equipment and storage medium for analyzing the transmission path of infectious diseases

Abstract

The present application relates to the fields of artificial intelligence and digital medicine, and specifically discloses a method, device, equipment and storage medium for analyzing the transmission path of infectious diseases. The method includes: obtaining user information in the target area, the user information includes: identity information, infectious disease status information, social relationship information and action trajectory information; generating a network diagram and a feature matrix based on the identity information, the infectious disease status information, the social relationship information and the action trajectory information; using a pre-trained prediction neural network model to analyze the network diagram and the feature matrix to generate prediction data of the transmission path of infectious diseases; generating health warning information based on the prediction data, and outputting the health warning information. Based on this method, it is possible to predict the infection status of users in an infectious environment and avoid the loss of health of people caused by large-scale infection.

Description

Technical Field

[0001] This application relates to the fields of artificial intelligence and digital healthcare, and in particular to a method, apparatus, device, and storage medium for analyzing the transmission path of an infectious disease. Background Technology

[0002] Currently, there is a lack of mechanisms and systems for automatically detecting the transmission routes and scope of infectious diseases. Information obtained through existing methods cannot accurately reflect the spread of viruses, leading to delays in disease early warning and prevention deployment. Furthermore, the inability to link and share disease prevention systems across regions prevents local authorities from understanding the disease's progression and development, hindering user-level prevention and management of infectious diseases. Once a regional infectious disease outbreak occurs, it will seriously threaten public health and safety. Summary of the Invention

[0003] This application provides a method, apparatus, equipment, and storage medium for analyzing the transmission path of infectious diseases, used to predict the transmission path of infectious diseases, realize user-level prevention and management of infectious diseases, and avoid regional infectious disease epidemics.

[0004] In a first aspect, this application provides a method for analyzing the transmission path of an infectious disease, the method comprising:

[0005] Obtain user information in the target area, including: identity information, infectious disease status information, social relationship information, and movement trajectory information;

[0006] A network graph and a feature matrix are generated based on the identity information, the infectious disease status information, the social relationship information, and the movement trajectory information;

[0007] By analyzing the network graph and the feature matrix using a pre-trained predictive neural network model, predictive data on the transmission path of infectious diseases is generated.

[0008] Based on the predicted data, a health warning message is generated and output.

[0009] Secondly, this application provides a device for analyzing the transmission path of infectious diseases, the device comprising: a data acquisition module, a data processing module, a result prediction module, and a data transmission module;

[0010] The data acquisition module is used to acquire user information in the target area, including: identity information, infectious disease status information, social relationship information, and movement trajectory information;

[0011] The data processing module is used to generate a network graph and a feature matrix based on the identity information, the infectious disease status information, the social relationship information, and the movement trajectory information;

[0012] The result prediction module is used to analyze the network graph and the feature matrix using a pre-trained prediction neural network model to generate prediction data on the transmission path of infectious diseases.

[0013] The data transmission module is used to generate health warning information based on the predicted data and output the health warning information.

[0014] Thirdly, this application provides a computer device, which includes a memory and a processor;

[0015] The memory is used to store computer programs;

[0016] The processor is configured to execute the computer program and, when executing the computer program, implement any of the infectious disease transmission path analysis methods provided in the embodiments of this application.

[0017] Fourthly, this application provides a computer-readable storage medium, characterized in that the computer-readable storage medium stores a computer program, which, when executed by a processor, causes the processor to implement a method for analyzing the transmission path of an infectious disease as provided in any of the embodiments of this application.

[0018] This application discloses a method, apparatus, device, and storage medium for analyzing the transmission path of infectious diseases. The method includes: acquiring user information from a target area, including identity information, infectious disease status information, social relationship information, and movement trajectory information; generating a network graph and feature matrix based on the identity information, infectious disease status information, social relationship information, and movement trajectory information; analyzing the network graph and feature matrix using a pre-trained predictive neural network model to generate predicted data on the transmission path of the infectious disease; and generating and outputting health warning information based on the predicted data. The technical solution provided in this application constructs a network graph and feature matrix from user information within the target area, and processes the network graph and feature matrix using a neural network model to generate user-level health warning information. Based on this, the spread of infectious diseases can be reflected in a timely manner, reducing the response time of disease control management agencies and preventing regional infectious disease pandemics. Attached Figure Description

[0019] To more clearly illustrate the technical solutions of the embodiments of this application, the drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0020] Figure 1 This is a schematic diagram illustrating the transmission process of an infectious disease, as provided in an embodiment of this application.

[0021] Figure 2 This is a schematic flowchart of a method for analyzing the transmission path of an infectious disease provided in an embodiment of this application;

[0022] Figure 3 This is a schematic block diagram of a network diagram provided in an embodiment of this application;

[0023] Figure 4 This is a schematic diagram of a predictive data generation process provided in an embodiment of this application;

[0024] Figure 5 This is a schematic block diagram of an infectious disease transmission path analysis device provided in an embodiment of this application;

[0025] Figure 6 This is a schematic block diagram of the structure of a computer device provided in an embodiment of this application. Detailed Implementation

[0026] 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, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0027] The flowchart shown in the attached diagram is for illustrative purposes only and does not necessarily include all content and operations / steps, nor does it necessarily have to be performed in the order described. For example, some operations / steps can be broken down, combined, or partially merged, so the actual execution order may change depending on the actual situation.

[0028] It should be understood that the terminology used in this specification is for the purpose of describing particular embodiments only and is not intended to limit the scope of the application. As used in this specification and the appended claims, the singular forms “a,” “an,” and “the” are intended to include the plural forms unless the context clearly indicates otherwise.

[0029] It should also be understood that the term “and / or” as used in this application specification and the appended claims means any combination of one or more of the associated listed items and all possible combinations, and includes such combinations.

[0030] In order to predict the infection status of users in infectious environments and avoid health losses caused by large-scale infections, this application provides a method, apparatus, equipment and storage medium for analyzing the transmission path of infectious diseases.

[0031] The following detailed description of some embodiments of this application is provided in conjunction with the accompanying drawings. Unless otherwise specified, the following embodiments and features can be combined with each other.

[0032] Please see Figure 1 , Figure 1 This illustration shows a schematic diagram of the transmission process of an infectious disease according to an embodiment of this application. Figure 1 As shown, the activities of an infectious disease patient A1 in society may lead to the spread of the infectious disease. The targets of the infection may be other users who have close social relationships or spatial contact, such as close contact groups S1 in schools, close contact groups K1 in public transportation places, close contact groups M1 in public catering places, and close contact groups J1 in families. Among them, users who enter the transmission network have characteristics such as social association and spatial aggregation. Therefore, these characteristics can be used to accurately identify users who have entered the transmission path of the infectious disease.

[0033] It should be noted that the infectious diseases in this application embodiment are a class of diseases caused by various pathogens that can be transmitted between people, animals, or between people and animals. In particular, the epidemic prevention department must keep abreast of the incidence of these diseases and take timely countermeasures. Therefore, once discovered, these diseases should be reported to the local epidemic prevention department in a timely manner according to the prescribed time.

[0034] It should also be noted that the embodiments of this application can acquire and process relevant data based on artificial intelligence technology, such as identifying users in susceptible environments based on social connections and spatial clustering. Artificial intelligence (AI) is the theory, method, technology, and application system that uses digital computers or machines controlled by digital computers to simulate, extend, and expand human intelligence, perceive the environment, acquire knowledge, and use that knowledge to obtain optimal results.

[0035] Foundational technologies for artificial intelligence generally include sensors, dedicated AI chips, cloud computing, distributed storage, big data processing, operating / interactive systems, and mechatronics. AI software technologies mainly encompass computer vision, robotics, biometrics, speech processing, natural language processing, and machine learning / deep learning.

[0036] Please see Figure 2 , Figure 2 This is a schematic flowchart illustrating a method for analyzing the transmission path of an infectious disease, as provided in an embodiment of this application. This method is used to quickly and accurately identify users in the transmission chain of an infectious disease. The method utilizes a predictive neural network model and a webpage ranking algorithm (PageRank power iteration formula).

[0037] like Figure 2 As shown, the propagation path analysis method specifically includes steps S101 to S106.

[0038] S101. Obtain user information in the target area. The user information includes: infectious disease status information, social relationship information, and movement trajectory information.

[0039] Specifically, the target area is determined, and user information of all users within the target area is obtained. The user information to be obtained includes at least: identity information, infectious disease status information, social relationship information, and movement trajectory information.

[0040] It should be noted that a user's movement trajectory information is time-stamped location information generated by the user during social activities. This movement trajectory information can be obtained through the user's terminal device, which can be an electronic device such as a mobile phone, tablet, laptop, desktop computer, personal digital assistant, and wearable device.

[0041] It should also be noted that in this application, infectious disease status information includes: infected persons, healthy persons, asymptomatic infected persons, recovered persons, and deceased persons; social relationship information includes blood relations, friend relations, and colleague relations, which can be further subdivided, for example, blood relations include parent-child relations, grandparent-grandchild relations, uncle-nephew relations, and nephew-uncle relations.

[0042] For example, Futian District of Shenzhen is identified as the target area, and information on all users who have passed through Futian District of Shenzhen within seven days is obtained, such as the user's identity, whether they are infected, who the other users with whom they have social relationships are, and their movement trajectory in the past 14 days.

[0043] In some embodiments, the scope of information acquisition can be expanded based on the user's social relationship information. For example, if Zhang San and Li Si meet at a party, when analyzing Zhang San, other users in Li Si's social network can be included in the scope of the analysis.

[0044] S102. Generate a network graph and feature matrix based on identity information, infectious disease status information, social relationship information, and movement trajectory information.

[0045] Specifically, the identity information in the user information is mapped into a network graph, and multiple network nodes are generated in the network graph. Each network node corresponds to a specific user. Edges between network nodes are generated based on social relationship information and movement trajectory information. At the same time, the infectious disease status information, social relationship information, and movement trajectory information are encoded to generate three feature matrices, namely the infectious disease status matrix, the social relationship matrix, and the movement trajectory matrix.

[0046] For example, let G(V,E) represent the network graph (topology) between people. Based on the user information of each infected person and asymptomatic infected person, a corresponding network node is generated in the network graph, where V={v1,v2,…,v…} N} is the set of network nodes of infected individuals, N is the total number of nodes, and E represents the set of "edges" between individuals, which are composed of arrows and routes in the network graph.

[0047] Please see Figure 3 , Figure 3 A schematic block diagram of a network graph is shown. For example... Figure 3 As shown, the diagram contains eight network nodes, U1 to U8. Each network node represents a specific user. There are one or two edges between different network nodes. Solid edges can indicate that there is a social relationship between two network nodes, while dashed edges can indicate that there is an overlap in the movement trajectories between two network nodes.

[0048] In some embodiments, before generating the feature matrix, it is also necessary to binarize the social relationship information and the movement trajectory information.

[0049] For example, advancements in transportation and shipping technologies have enabled cities to transcend geographical distances. Even though permanent residences may be far apart, people can quickly shorten these distances using transportation. Therefore, it is necessary to consider movement trajectories when controlling the spread of infectious diseases. In one implementation, a distance threshold is set. Movement trajectory relationships greater than the distance threshold are classified as "other," while those less than or equal to the distance threshold are classified as "adjacent." Therefore, the characteristic values ​​of distance relationships include "other" and "adjacent," which can be represented by "0" and "1," respectively. The following movement trajectory matrix can be constructed:

[0050]

[0051] The user's movement trajectory information can be collected through the navigation and positioning module of the user's terminal device, specifically including longitude and latitude information. When applying the information, it is also necessary to compare the time when the longitude and latitude information were generated to determine whether the patient had close contact with others.

[0052] In other embodiments, to indicate whether person i and person j constitute an adjacency relationship in terms of social relations, a social relationship matrix can be constructed as follows:

[0053]

[0054] In other embodiments, the infectious disease states that users in a susceptible environment may have include: infected, healthy, asymptomatic infected, recovered, and deceased. Each infectious disease state is represented by a specific numerical value in a matrix, where 0 represents an infected person, 1 represents a healthy person, 2 represents an asymptomatic infected person, 3 represents a recovered person, and 4 represents a deceased person; column vector x (t) Let x represent the infectious disease status of a user on day t. The infectious disease status of each user at the previous T time points forms an infectious disease status matrix based on time distribution, which is [x]. (t-T) ,x (t-T-1) ,…,x (T) The infectious disease state matrix is ​​the input feature vector matrix.

[0055] S103. Analyze the network graph and feature matrix using a pre-trained predictive neural network model to generate predictive data on the transmission path of infectious diseases.

[0056] Specifically, the network graph, infectious disease status matrix, social relationship matrix, and action trajectory matrix are input into the trained predictive neural network model. The model outputs the influence of network nodes in the virus infection state on other network nodes and the prediction results of the infectious disease status of other network nodes. The prediction result of each network node corresponds to the infectious disease status of a user. The prediction results are then converted into prediction data of the spread path of infectious diseases in the target area.

[0057] For example, please refer to Figure 4 , Figure 4 This illustrates a schematic diagram of a predictive data generation process. For example... Figure 4As shown, in a network graph consisting of 10 users, there is one infected person Z (represented by black fill). The user information of the 10 users is collected, and the social relationship information, movement trajectory information of the 10 users, and the infectious disease status of infected person Z are output to the trained predictive neural network model. The model outputs the predicted infectious disease status information of the other 9 users. For example, black fill indicates that the user has been infected or is suspected of being infected, and white fill indicates that the user has not been infected.

[0058] In some embodiments, the prediction data can also be dynamically changing. Therefore, the prediction neural network model should also include a real-time data refresh function to update user information and prediction data, such as upgrading health warning information for some users or lifting health alert status for some users.

[0059] In one embodiment, before using the trained predictive neural network model, a training method for the predictive neural network model is also included. It should be noted that the predictive neural network model includes a propagation layer and a prediction layer. The propagation layer is generated based on a webpage ranking algorithm, and the prediction layer is generated based on a graph convolutional neural network model. The specific steps of this training method include:

[0060] Obtain a training sample set, which includes user information of multiple infected individuals;

[0061] Generate a training network graph and a training feature matrix based on user information from multiple infected individuals;

[0062] The training network graph and training feature matrix are input into the predictive neural network model for training, resulting in the predictive neural network model.

[0063] Therefore, the trained predictive neural network model can predict the infectious disease status of other network nodes based on the known infectious disease status of network nodes in the network graph.

[0064] Specifically, the main calculation formula for the Graph Convolutional Network (GCN) model used to build the prediction layer is as follows:

[0065]

[0066]

[0067]

[0068] Where Agg represents aggregation operation, l = 0, 1, 2, ..., is the current layer number (i.e., passing through the GCN model l times), X (l) It is the input feature vector matrix (i.e., the infectious disease state matrix). It is a learnable parameter matrix. It is an adjacency matrix with self-loops. It is a social relationship adjacency matrix. It is the adjacency matrix of movement trajectories. yes The degree matrix, It is the Laplace matrix The regularized form of the expression, where k represents the neighboring nodes of the given node (someone), i.e., nodes with social relationships, or nodes connected by location information. The Concat function concatenates the hidden vectors obtained from social relationships and distance. It is a feature vector of social relationships. It is the feature vector of the movement trajectory, H (l+1) It is the prediction vector matrix.

[0069] The main calculation formula for the webpage ranking algorithm used to build the propagation layer is:

[0070] H (0) =H l+1

[0071]

[0072]

[0073] y T =RELU(H T )

[0074] Among them, H (0) The input matrix for the propagation layer is directly taken here as the feature vector obtained after passing through GCN, i.e.

[0075] H (0) =H l+1

[0076] And the eigenvector H at time T+1 (T+1) y is the sum of its own eigenvector and the propagation result of the node representation from the previous time step on the graph structure. T It is a label for the predicted state of an infectious disease.

[0077] It's important to note that PageRank is a ranking algorithm that calculates page ranking based on the hyperlinks between web pages. A page's "votes" are determined by the importance of all pages linking to it; a hyperlink to a page is equivalent to casting a vote for that page. A page's PageRank is derived recursively from the importance of all pages linking to it ("inbound pages"). A page with more inbound links will have a higher ranking, while a page with no inbound links will have no ranking.

[0078] Based on the above construction process, a predictive neural network model is obtained.

[0079] It should be noted that the training process can be summarized as learning a function through a GCN (Graph Convolutional Network) model: The mapping. Input the states [x] at the previous T time steps in sequence. (t-T) ,x (t-T-1) ,…,x (t) Predict the state at the next time step, and adjust the learnable parameters to make the predicted value at time t+1 more accurate. Compared with the true value x (t+1) The error between them should be as small as possible.

[0080] Separating the neural network from the message-passing network solves the problem of excessive parameter count. K defines the number of iterations, and the transmission probability adjusts the size of the neighborhood affecting each node, allowing us to freely modify different types of networks to adjust the model, expand the analytical neighborhood, and improve prediction accuracy. The above formula can also address overly smooth linearity, achieving a steady-state distribution even with increased layer depth.

[0081] In some embodiments, a loss function is used to reduce training error, and parameters are updated via backpropagation to optimize the model. The loss function includes the mean squared error loss function, which is calculated using the following formula:

[0082] Where L is the loss factor, H lf To predict the vector Y corresponding to the f-th input feature vector X in the l-th layer of the vector matrix. lf This is the vector corresponding to the f-th input vector X in the l-th layer of the input feature vector matrix.

[0083] Once the predictive neural network is built, you can begin training it using the training data.

[0084] In some embodiments, a training sample set is obtained, which includes: user information of multiple infected individuals; a training network graph and a training feature matrix are generated based on the user information of multiple infected individuals.

[0085] Specifically, information on multiple infected individuals and asymptomatic carriers over W days is obtained from medical system databases or CDC databases. The user information is mapped into a training network graph, and multiple network nodes are generated in the network graph. Each network node corresponds to a specific infected individual. At the same time, infectious disease status information, social relationship information, and movement trajectory information are encoded to generate three training feature matrices, namely the training infectious disease status matrix, the training social relationship matrix, and the training movement trajectory matrix.

[0086] For example, user information from 327 infected individuals over 50 days was selected as the training sample set. The input feature of the prediction training network is the infectious disease state matrix. The infectious disease state inputs of the users in the training sample set are denoted as X1, X2, ..., X 50 , where X1 = [x (1) ,x (2) ,…,x (7) ], X2 = [x (2) ,x (3) ,…,x (8) ..., X 43 =[x (43) ,x (W-6) ,…,x (50) The training label is y. (1) ,y (2) ,…,y (43) , where [y (1) ,y (2) ,…,y (43) ] corresponds to [x (8) ,x (9) ,…,x (50) The training labels serve as a control group for each training result, correcting the training process. (1) The infectious disease status of users at 8:00 AM.

[0087] In some embodiments, the training network graph and the training feature matrix are input into the predictive neural network model to train the predictive neural network model and obtain the predictive neural network model.

[0088] Specifically, the training network graph and three training feature matrices are input into the prediction neural network model. The prediction neural network model is trained to obtain a mapping relationship between social relationships and movement trajectories and the state of infectious diseases, simulating the spread of the virus in network nodes.

[0089] For example, because infectious diseases are characterized by social connections and spatial clustering, different social relationships and travel trajectories with varying degrees of overlap have different transmission weights. For instance, if a certain influenza outbreak occurs in a region, during the learning process of a predictive neural network model, it is found that the proportion of patients with parent-child relationships is much higher than the proportion with friends. This is because parent-child relationships have a greater possibility of close contact. Therefore, for this influenza, the transmission weight obtained by the parent-child relationship in the predictive neural network model is greater than the transmission weight obtained by the friend relationship.

[0090] In other embodiments, the propagation weights of social relationships in the predictive neural network model can be adjusted based on travel trajectories. For example, a high school student, boarding at their school, did not see their family during a flu outbreak. Based on their travel trajectories, it can be determined that the student's close contacts are teachers and classmates. Therefore, if the student contracts the flu virus, the probability of them spreading the virus to other individuals with whom they have close relationships (classmates and teachers / students) is highest, while the probability of spreading it to their parents is lower. Thus, when analyzing the probability of the student spreading the flu, the propagation weight for other individuals with whom they had close contact along their travel trajectories is highest, while the propagation weight for their parents, with whom they had no contact along their travel trajectories, is lowest.

[0091] In other embodiments, travel history can be used as the primary basis for calculating transmission weights. For example, based on the travel history of a known infected person, public places along that travel history can be disinfected, and close contacts whose travel history completely overlaps can be identified. It should be noted that when using travel history as the primary basis, social relationships can be used as supplementary information to correct the predictive neural network model.

[0092] In other embodiments, information such as the user's age, occupation, medical history, and vaccination status can be used as auxiliary information, processed accordingly, and then input into the predictive neural network model to correct the predictive neural network model and obtain more accurate prediction results.

[0093] S104. Based on the predicted data, generate health warning information and send the health warning information to the target area.

[0094] Specifically, the system acquires predictive data, identifies infected or suspected infected individuals in the target area based on the predictive data, generates health warning information, and sends the health warning information to the infectious disease prevention and control agency in the target area. The health warning information includes the user identity of the suspected infected individual.

[0095] For example, when multiple infected individuals appear in a region, a predictive neural network model can be used to identify close contacts who may be affected by the infected individuals, such as colleagues, friends, and relatives. This could also include people who have had close contact in public places, such as staff at the restaurant where the infected individuals ate. Furthermore, the social relationships and travel trajectories of close contacts can be re-predicted, and suspected infected individuals who may be affected by the infected individuals can be provided to relevant institutions within a specified time. This allows relevant institutions to take timely warnings and preventive measures to avoid large-scale infections and the resulting loss of human health.

[0096] Please see Figure 5 , Figure 5 This application also provides a schematic block diagram of an infectious disease transmission path analysis device 300, which is used to execute the aforementioned infectious disease transmission path analysis method. The infectious disease transmission path analysis device can be configured in a server or terminal.

[0097] The server can be a standalone server, a server cluster, or a cloud server providing basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, content delivery networks (CDNs), and big data and artificial intelligence platforms. The terminal can be an electronic device such as a mobile phone, tablet, laptop, desktop computer, user digital assistant, or wearable device.

[0098] like Figure 5 As shown, the infectious disease transmission path analysis device 300 includes: a data acquisition module 301, a data processing module 302, a result prediction module 303, and a data transmission module 304.

[0099] The data acquisition module 301 is used to acquire user information in the target area. The user information includes: identity information, infectious disease status information, social relationship information, and movement trajectory information.

[0100] The data processing module 302 is used to generate network graphs and feature matrices based on identity information, infectious disease status information, social relationship information, and movement trajectory information.

[0101] In some embodiments, the data processing module 302 is used to map identity information into a network graph, generate multiple network nodes in the network graph, and each network node corresponds to a user;

[0102] Based on social relationship information and movement trajectory information, generate edges between network nodes;

[0103] Information on infectious disease status, social relationships, and movement trajectories is encoded to generate three feature matrices: the infectious disease status matrix, the social relationship matrix, and the movement trajectory matrix.

[0104] In some embodiments, the data processing module 302 also performs binarization processing on social relationship information and movement trajectory information.

[0105] The result prediction module 303 is used to analyze the network graph and feature matrix using a pre-trained prediction neural network model to generate prediction data on the transmission path of infectious diseases.

[0106] In some embodiments, the predictive neural network model includes a propagation layer and a prediction layer, wherein the propagation layer is generated based on a webpage ranking algorithm and the prediction layer is generated based on a graph convolutional neural network model.

[0107] In some embodiments, the result prediction module 303 is further configured to:

[0108] Obtain a training sample set, which includes user information of multiple infected individuals;

[0109] Generate a training network graph and a training feature matrix based on user information from multiple infected individuals;

[0110] The training network graph and training feature matrix are input into the predictive neural network model for training, resulting in the predictive neural network model.

[0111] In some embodiments, the result prediction module 303 is further configured to create a correction function to correct the prediction neural network model;

[0112] The correction functions include: the mean squared error loss function, which is:

[0113]

[0114] Where L is the loss factor, H lf To predict the vector Y corresponding to the f-th input feature vector X in the l-th layer of the vector matrix. lf This is the vector corresponding to the f-th input vector X in the l-th layer of the input feature vector matrix.

[0115] In some embodiments, the result prediction module 303 is further configured to:

[0116] Input the network graph and feature matrix into the trained predictive neural network model;

[0117] Using a trained predictive neural network model, the influence of network nodes with viral infection status on other network nodes is analyzed, and the prediction results of the infectious disease status of other network nodes are output.

[0118] Based on the prediction results, predictive data on the transmission paths of infectious diseases are generated.

[0119] The data sending module 304 is used to generate and output health warning information based on the predicted data.

[0120] In some embodiments, the data sending module 304 is further configured to send health warning information to the disease management system to indicate the spread of infectious diseases in the target area.

[0121] It should be noted that those skilled in the art will understand that, for the sake of convenience and brevity, the specific working process of the infectious disease transmission path analysis device and its modules described above can be referred to the corresponding process in the aforementioned embodiments of the infectious disease transmission path analysis method, and will not be repeated here.

[0122] It should be noted that those skilled in the art will understand that, for the sake of convenience and brevity, the specific working process of the model training device and each module described above can be referred to the corresponding process in the aforementioned embodiment of the method for analyzing the transmission path of infectious diseases, and will not be repeated here.

[0123] The aforementioned infectious disease transmission path analysis device can be implemented as a computer program, which can, for example... Figure 6 It runs on the computer device shown.

[0124] Please see Figure 6 , Figure 6 This is a schematic block diagram illustrating the structure of a computer device according to an embodiment of this application. The computer device may be a server or a terminal.

[0125] See Figure 6 The computer device includes a processor, memory, and network interface connected via a system bus, wherein the memory may include storage media and internal memory.

[0126] The storage medium may store an operating system and a computer program. The computer program includes program instructions that, when executed, cause the processor to perform any of the infectious disease transmission path analysis methods provided in the embodiments of this application.

[0127] The processor provides computing and control capabilities, supporting the operation of the entire computer device.

[0128] Internal memory provides an environment for the execution of computer programs stored in the storage medium. When executed by a processor, these programs can enable the processor to perform methods for analyzing the transmission paths of any infectious disease or training predictive neural networks. The storage medium can be non-volatile or volatile.

[0129] This network interface is used for network communication, such as sending assigned tasks. Those skilled in the art will understand that... Figure 6 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0130] It should be understood that the processor can be a Central Processing Unit (CPU), but it can also be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. Among these, a general-purpose processor can be a microprocessor or any conventional processor.

[0131] For example, in one embodiment, the processor is configured to run a computer program stored in memory to perform the following steps:

[0132] Obtain user information in the target area, including: identity information, infectious disease status information, social relationship information, and movement trajectory information;

[0133] Network graphs and feature matrices are generated based on identity information, infectious disease status information, social relationship information, and movement trajectory information;

[0134] By analyzing the network graph and feature matrix using a pre-trained predictive neural network model, predictive data on the transmission paths of infectious diseases can be generated.

[0135] Based on the predicted data, health warning information is generated and output.

[0136] In one embodiment, the processor is used to run a computer program stored in memory to perform the following steps:

[0137] Obtain a training sample set, which includes user information of multiple infected individuals;

[0138] Generate a training network graph and a training feature matrix based on user information from multiple infected individuals;

[0139] The training network graph and training feature matrix are input into the predictive neural network model for training, resulting in the predictive neural network model.

[0140] In some embodiments, when implementing training of a predictive neural network model, the processor is also specifically configured to implement:

[0141] Create a correction function to correct the predictive neural network model;

[0142] The correction functions include: the mean squared error loss function, which is:

[0143]

[0144] Where L is the loss factor, H lf To predict the vector Y corresponding to the f-th input feature vector X in the l-th layer of the vector matrix. lf This is the vector corresponding to the f-th input vector X in the l-th layer of the input feature vector matrix.

[0145] In some embodiments, when the processor generates the network graph and feature matrix based on identity information, infectious disease status information, social relationship information, and movement trajectory information, it is also specifically used to implement:

[0146] Identity information is mapped into a network graph, and multiple network nodes are generated in the network graph, with each network node corresponding to a user;

[0147] Based on social relationship information and movement trajectory information, generate edges between network nodes;

[0148] Information on infectious disease status, social relationships, and movement trajectories is encoded to generate three feature matrices: the infectious disease status matrix, the social relationship matrix, and the movement trajectory matrix.

[0149] In some embodiments, before generating the network graph and feature matrix based on identity information, infectious disease status information, social relationship information, and movement trajectory information, the processor is also specifically configured to implement:

[0150] Binarization is performed on social relationship information and movement trajectory information.

[0151] In some embodiments, when the processor analyzes the network graph and feature matrix using a pre-trained predictive neural network model to generate predictive data on the transmission path of infectious diseases, it is further specifically configured to implement:

[0152] Input the network graph and feature matrix into the trained predictive neural network model;

[0153] Using a trained predictive neural network model, the influence of network nodes with viral infection status on other network nodes is analyzed, and the prediction results of the infectious disease status of other network nodes are output.

[0154] Based on the prediction results, predictive data on the transmission paths of infectious diseases are generated.

[0155] In some embodiments, when implementing the output of health warning information, the processor is also specifically used to implement:

[0156] Health alerts are sent to the disease management system to indicate the spread of infectious diseases within the target area.

[0157] The computer-readable storage medium may be an internal storage unit of the computer device described in the foregoing embodiments, such as the hard disk or memory of the computer device. The computer-readable storage medium may also be an external storage device of the computer device, such as a plug-in hard disk, SmartMedia Card (SMC), Secure Digital (SD) card, or Flash Card equipped on the computer device.

[0158] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any person skilled in the art can easily conceive of various equivalent modifications or substitutions within the technical scope disclosed in this application, and these modifications or substitutions should all be covered within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A method for analyzing the transmission path of an infectious disease, characterized in that: The method comprises: Acquire user information in the target area, the user information including: identity information, infectious disease status information, social relationship information, and movement trajectory information; Generate a network graph and a feature matrix according to the identity information, the infectious disease status information, the social relationship information and the action trajectory information, including: mapping the identity information into a network graph, generating multiple network nodes in the network graph, each network node corresponding to a user; generating edges between network nodes according to the social relationship information and the action trajectory information; encoding the infectious disease status information, the social relationship information and the action trajectory information, and generating three feature matrices, namely, an infectious disease status matrix, a social relationship matrix and an action trajectory matrix; The network graph and the feature matrix are analyzed using a pre-trained prediction neural network model to generate prediction data of the transmission path of the infectious disease; wherein the prediction neural network model includes a propagation layer and a prediction layer, the propagation layer is generated based on a web page ranking algorithm, and the prediction layer is generated based on a graph convolutional neural network model; the method for obtaining the prediction neural network model also includes: obtaining a training sample set, the training sample set includes user information of multiple infected persons; generating a training network graph and a training feature matrix based on the user information of multiple infected persons; inputting the training network graph and the training feature matrix into the prediction neural network model for training to obtain a prediction neural network model; the main calculation formula for building the graph convolutional neural network model used in the prediction layer is: Among them, Agg represents the aggregation operation, l=0,1,2,…, is the current layer number, is the input eigenvector matrix, is a learnable parameter matrix, is an adjacency matrix with self-loops, is the social relationship adjacency matrix, is the action trajectory adjacency matrix, yes The degree matrix of is the Laplace matrix The regularized form of k represents the neighboring node corresponding to the network node; the Concat function connects the hidden vectors obtained by social interaction and distance; is the social relationship feature vector, is the action trajectory feature vector, is the prediction vector matrix; Health warning information is generated based on the prediction data, and the health warning information is output.

2. The method according to claim 1, characterized in that The training of the prediction neural network model also includes: Create a correction function and use it to correct the prediction neural network model; The correction function includes: a mean square error loss function, and the mean square error loss function is: Where L is the loss factor, is the vector corresponding to the f-th input feature vector X in the l-th layer in the prediction vector matrix, is the vector corresponding to the f-th input vector X in the l-th layer of the input feature vector matrix.

3. The method according to claim 1, characterized in that Before encoding the infectious disease status information, social relationship information and action trajectory information, the method further includes: The social relationship information and the action trajectory information are binarized.

4. The method according to claim 1, characterized in that The method of analyzing the network diagram and the feature matrix using a pre-trained prediction neural network model to generate prediction data of the transmission path of the infectious disease includes: Inputting the network graph and the feature matrix into a trained prediction neural network model; Using the trained prediction neural network model, analyze the impact of the network node in the virus infection state on other network nodes, and output the prediction results of the infectious disease state of other network nodes; The prediction data of the transmission path of the infectious disease is generated based on the prediction results.

5. The method according to claim 1, characterized in that The target area is provided with a disease management system, and the outputting of the health warning information includes: The health warning information is sent to the disease management system to prompt the spread of infectious diseases in the target area.

6. A device for analyzing the transmission path of an infectious disease, characterized in that: include: A data acquisition module is used to acquire user information in a target area, wherein the user information includes: identity information, infectious disease status information, social relationship information, and action trajectory information; A data processing module, for generating a network graph and a feature matrix according to the identity information, the infectious disease status information, the social relationship information and the action trajectory information, including: mapping the identity information into a network graph, generating multiple network nodes in the network graph, each network node corresponding to a user; generating edges between network nodes according to the social relationship information and the action trajectory information; encoding the infectious disease status information, the social relationship information and the action trajectory information, and generating three feature matrices, namely, an infectious disease status matrix, a social relationship matrix and an action trajectory matrix; The result prediction module is used to analyze the network graph and the feature matrix using a pre-trained prediction neural network model to generate prediction data of the transmission path of the infectious disease; wherein the prediction neural network model includes a propagation layer and a prediction layer, the propagation layer is generated based on a web page ranking algorithm, and the prediction layer is generated based on a graph convolutional neural network model; the method for obtaining the prediction neural network model also includes: obtaining a training sample set, the training sample set includes user information of multiple infected persons; generating a training network graph and a training feature matrix based on the user information of multiple infected persons; inputting the training network graph and the training feature matrix into the prediction neural network model for training to obtain a prediction neural network model; the main calculation formula for building the graph convolutional neural network model used in the prediction layer is: Among them, Agg represents the aggregation operation, l=0,1,2,…, is the current layer number, is the input eigenvector matrix, is a learnable parameter matrix, is an adjacency matrix with self-loops, is the social relationship adjacency matrix, is the action trajectory adjacency matrix, yes The degree matrix of is the Laplace matrix The regularized form of k represents the neighboring node corresponding to the network node; the Concat function connects the hidden vectors obtained by social interaction and distance; is the social relationship feature vector, is the action trajectory feature vector, is the prediction vector matrix; The data sending module is used to generate health warning information according to the predicted data and output the health warning information.

7. A computer device, characterized in that: The computer device includes a memory and a processor; The memory is used to store computer programs; The processor is used to execute the computer program and implement the infectious disease transmission path analysis method as described in any one of claims 1 to 5 when executing the computer program.

8. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, which, when executed by a processor, enables the processor to implement the method for analyzing the transmission path of an infectious disease as described in any one of claims 1 to 5.

Images