Urban influenza incidence trend prediction method, system, terminal and storage medium

Abstract

The present application relates to a method, system, terminal and storage medium for predicting the trend of influenza incidence in a city. The method includes: obtaining influenza case data in the city, and collecting individual movement trajectory data in the city; processing individual movement trajectory data through a data-driven method to obtain population movement relations; based on population movement relations, using a graph neural network to extract spatial scale information of influenza case data, and using a long short-term memory network to extract time series relations of influenza case data; obtaining the forecast results of influenza incidence trends in influenza cities based on the spatial scale information and time series relations of influenza case data. The embodiment of the present application realizes the prediction of higher spatial resolution of influenza development trends in cities, completes the refined analysis of influenza, helps the government and public health departments to timely and accurately understand the development trends of influenza in cities, and conducts targeted epidemic prevention and control interventions, which can maximize the protection of people's lives, health and safety.

Description

Technical Field

[0001] The present application belongs to the field of influenza computing technology, and in particular relates to a method, system, terminal and storage medium for predicting influenza incidence trends in influenza cities. Background Art

[0002] From the end of 2019 to the beginning of 2021, COVID-19 has developed from a local outbreak to a global pandemic, infecting more than 100 million people and causing more than 2.52 million deaths worldwide. The outbreak of this epidemic reminds us that public health risks are still one of the main social risks facing mankind. Take influenza (flu) as an example. It is an acute respiratory influenza caused by influenza virus. The influenza virus can break out in a short period of time, seriously threatening people's health.

[0003] At present, the influenza prevention and control and early warning methods in various countries around the world mainly rely on traditional influenza monitoring. The influenza monitoring system can provide a large amount of information and data useful for influenza prevention and control, including the number of infected cases, the number of deaths, clinical symptoms and information on hospitalized persons. The government and public health departments evaluate the recent trend of influenza activity based on the information reported by the influenza monitoring system, and make corresponding public health decisions and emergency response plans. However, since the verification and reporting of grassroots monitoring data often takes a lot of time (weeks or even months), the monitoring data will lag behind the actual situation, and most countries can only provide early warnings at the provincial / state or city level due to the lack of complete information collection technology. Therefore, it is urgent to make real-time, accurate and rapid predictions of the influenza incidence trend within the city to help the government and public health departments timely and accurately understand the development trend of influenza within the city and maximize the protection of people's life, health and safety.

[0004] Human interaction and movement can lead to the spread and diffusion of influenza. In order to better realize the real-time, accurate and rapid prediction of the influenza incidence trend within the city, it is necessary to conduct in-depth research on the spatial interaction of the population movement process within the city and build a spatiotemporal prediction model of the urban influenza incidence trend that couples the spatial interaction. Existing influenza prediction modeling methods include those based on Gaussian process models and LSTM neural network models. However, most of the existing influenza incidence trend prediction methods often only rely on the time series characteristics of influenza-like illness statistical data for prediction, without considering the spatial dependence between different regions and the spatial interaction effect between population movement and influenza transmission. Such methods are difficult to accurately predict the influenza incidence trend within the city due to the lack of information. Although other influenza prediction methods make comprehensive use of the spatiotemporal feature information of specific locations, the spatial feature extraction method is not perfect, and the accuracy of the extracted spatial feature information needs to be improved.

[0005] As the complexity of human mobility networks within cities continues to deepen, traditional theoretical model methods are unable to simulate population mobility and interaction within cities in a refined manner, and have become difficult to adapt to the current needs of modeling spatial interactions within cities. In addition, spatial interaction modeling methods based on OD (traffic travel volume) flow are often studied at a lower spatial resolution, such as at the national, provincial / state or city level. At the city scale, the graph structure flow constructed by this method is too dispersed, resulting in deviations in the prediction of the spatiotemporal pattern of influenza outbreaks. Summary of the invention

[0006] The present application provides a method, system, terminal and storage medium for predicting influenza incidence trends in cities, aiming to solve at least one of the above-mentioned technical problems in the prior art to a certain extent.

[0007] In order to solve the above problems, this application provides the following technical solutions:

[0008] A method for predicting the trend of influenza incidence in a city, comprising:

[0009] Obtain influenza case data within the city and collect individual movement trajectory data within the city;

[0010] The individual movement trajectory data is processed by a data-driven method to obtain the home address of each individual, set the home address as the starting point of each individual movement trajectory, and extract the flow data of each individual from the starting point to other areas to obtain the population movement relationship between various areas;

[0011] Based on the population movement relationship, a graph neural network is used to extract the spatial scale information of the influenza case data, and a long short-term memory network is used to extract the time series relationship of the influenza case data;

[0012] The prediction result of the urban influenza incidence trend is obtained based on the spatial scale information and time series relationship of the influenza case data.

[0013] The technical solution adopted by the embodiment of the present application also includes: the collection of individual movement trajectory data in the city includes:

[0014] The individual movement trajectory data in the city is collected by using mobile devices; the mobile devices include mobile phones or smart watches; the individual movement trajectory data includes each individual's mobile phone number, signaling timestamp, and base station longitude and latitude.

[0015] The technical solution adopted by the embodiment of the present application also includes: the processing of individual movement trajectory data by the data-driven method is specifically as follows:

[0016] The original Thiessen polygons established by the urban base stations in the grid area of ​​the set time period and the set distance are merged; the individual movement trajectory data is divided according to the residence time, the location where each individual stays the longest at night is set as the individual's home address, and the OD flow and home center flow at the set interval time in the grid area of ​​the set distance are extracted.

[0017] The technical solution adopted by the embodiment of the present application also includes: the processing of the individual movement trajectory data by the data-driven method also includes:

[0018] According to the population distribution information and geographic location information of each street in the city in the census data, network flow extraction is performed on the individual movement trajectory data through a theoretical model method; the theoretical model method includes a gravity model, a radiation model or a spatial proximity relationship model.

[0019] The technical solution adopted by the embodiment of the present application also includes: the spatial scale information of the influenza case data extracted by using the graph neural network is specifically:

[0020] The framework of the graph neural network is a message propagation neural network, which extracts spatial scale information based on spatial domain graph convolution; define an undirected graph G, the feature vector of node v is x v , the edge feature is e vw , connect nodes v and w, N(v) represents the neighbor nodes of node v in graph G, t is the running time step, and the feature x of node v is v As the initial state of its hidden state After that, the update of the hidden state by the spatial domain graph convolution is expressed as:

[0021]

[0022]

[0023] The message propagation neural network decomposes the spatial graph convolution into two parts: message passing and state update operations, respectively represented by the message function M l and the node update function U l Completed; the message function M l It is used to aggregate the features of neighbor nodes to form a message vector, which is ready to be transmitted to the central node; the node update function U l It is used to update the node representation at the current moment, combine the node representation at the current moment and the message obtained from the message function, and obtain the spatial scale information.

[0024] The technical solution adopted by the embodiment of the present application also includes: the spatial scale information of the influenza case data extracted by using the graph neural network is specifically:

[0025] The individual movement trajectory data of a set time range is converted into a weighted directed graph, where vertices represent street-level areas and edges are used to capture movement patterns; at time t, the flow between areas u and v forms an edge, which is multiplied by the number of cases in area u at time t represents how many infected people may move from area u to area v; let is a vector of node attributes containing the number of cases in region u in each of the past w weeks; the message passed through the message propagation neural network uses the combined scores from all regions to calculate the feature vector for each region:

[0026]

[0027] Where A represents the adjacency matrix of regional population mobility flow, X t is a matrix whose rows contain the attributes of different regions; x u ∈R ww is a vector combining the number of influenza cases in region u and those moving toward region u, x u The expression formula is:

[0028] x u =(x u w j，u +x u w i，u +…x u w v，u )+x u w u，u

[0029] where x u ∈R w represents the estimate of the number of new potential cases in region u.

[0030] The technical solution adopted by the embodiment of the present application also includes: the time series relationship of extracting the influenza case data using the long short-term memory network is specifically:

[0031] The long short-term memory network is based on the input x at the current moment t and the output h of the hidden layer in the previous period t-1 To calculate the output h of the hidden layer at the current moment t , the calculation formula of the long short-term memory network is:

[0032] y i，t =LSTM(h i，t-n ,h i，t-n , ..., h i，t-1 )

[0033] Among them, h i，t-1represents the influenza case data of the ith region in the t-1th time period, y i，t Represents the influenza case data predicted for the ith region in the tth time period.

[0034] Another technical solution adopted by the embodiment of the present application is: a system for predicting the incidence trend of influenza in cities, comprising:

[0035] Data collection module: used to obtain influenza case data in the city and collect individual movement trajectory data in the city;

[0036] Data processing module: used to process the individual movement trajectory data through a data-driven method, obtain the home address of each individual, set the home address as the starting point of each individual movement trajectory, and extract the flow data of each individual from the starting point to other areas, and obtain the population movement relationship between various areas;

[0037] A spatiotemporal feature extraction module is used to extract the spatial scale information of the influenza case data based on the population movement relationship using a graph neural network, and to extract the time series relationship of the influenza case data using a long short-term memory network;

[0038] Flu prediction module: used to obtain the prediction result of the trend of flu incidence in the city according to the spatial scale information and time series relationship of the flu case data.

[0039] Another technical solution adopted by the embodiment of the present application is: a terminal, the terminal includes a processor and a memory coupled to the processor, wherein:

[0040] The memory stores program instructions for implementing the urban influenza incidence trend prediction method;

[0041] The processor is used to execute the program instructions stored in the memory to control the prediction of urban influenza incidence trends.

[0042] Another technical solution adopted by the embodiment of the present application is: a storage medium storing program instructions executable by a processor, wherein the program instructions are used to execute the urban influenza incidence trend prediction method.

[0043] Compared with the prior art, the beneficial effects produced by the embodiments of the present application are: the urban influenza incidence trend prediction method, system, terminal and storage medium of the embodiments of the present application extract population movement relationships through individual movement trajectory data in the city, and based on the population movement relationships, a graph convolutional neural network is used to obtain the spatial scale information of the influenza case data, and a long short-term memory network is used to extract the time series relationship of the influenza case data, and a refined prediction of the influenza trend is made based on the spatial scale information and the time series relationship. The embodiments of the present application achieve a higher spatial resolution prediction of the development trend of influenza within the city, complete a refined analysis of influenza, help the government and public health departments to timely and accurately understand the development trend of influenza within the city, and carry out targeted epidemic prevention and control interventions, which can maximize the protection of people's lives, health and safety. Compared with the prior art, the present application has at least the following beneficial effects:

[0044] (1) After incorporating large-scale mobile phone location data, the traditional theoretical model method has improved its defect of being unable to simulate the population movement and interaction within the city in a refined manner, and the mobility of individuals has been restored more realistically.

[0045] (2) Compared with other data-driven modeling methods such as the spatial interaction modeling method based on OD flow, the spatial interaction modeling method of household center flow can construct a more concentrated graph structure flow at a higher spatial resolution at the urban scale, and can clearly capture multiple strong commuting flows, thereby effectively improving the prediction accuracy of the deep learning model of influenza within the city.

[0046] (3) The present invention only needs to obtain the location information of the mobile phone, and does not need to obtain information of multiple dimensions such as weekly average temperature, air pressure, rainfall, relative humidity, maximum temperature difference and sunshine time, thereby reducing the difficulty of data acquisition and processing. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] Figure 1 is a flow chart of a method for predicting the incidence trend of influenza in a city according to an embodiment of the present application;

[0048] Figure 2 This is a schematic diagram of the structure of a system for predicting the incidence trend of urban influenza according to an embodiment of the present application;

[0049] Figure 3 This is a schematic diagram of the terminal structure of an embodiment of the present application;

[0050] Figure 4 A schematic diagram of the structure of a storage medium according to an embodiment of the present application. DETAILED DESCRIPTION

[0051] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0052] See also Figure 1 , is a flow chart of the method for predicting the incidence trend of urban influenza in an embodiment of the present application. The method for predicting the incidence trend of urban influenza in an embodiment of the present application comprises the following steps:

[0053] S1: Obtain influenza case data within the city and use mobile devices to collect individual movement trajectory data within the city;

[0054] In this step, mobile devices include but are not limited to mobile phones or smart watches. In the embodiment of the present application, individual mobile trajectory data includes but is not limited to each individual's mobile phone number, signaling timestamp, identified base station latitude and longitude, and other information. The individual mobile trajectory data set is obtained in the following manner: Based on the Unicom HDFS+Hive+Spark big data platform, the location information carried by the mobile phone signaling of 5.6 million users is obtained, the location information is deduplicated, and the records of missing key information in the location information are removed to obtain the individual mobile trajectory data.

[0055] S2: Process the individual mobility trajectory data through data-driven methods and theoretical model methods, obtain the home address of each individual, and set the home address as the starting point of all mobility trajectories of each individual, extract the flow data of each individual from the starting point to other areas, and obtain the population mobility relationship between various areas;

[0056] In this step, the individual movement trajectory data are divided according to the residence time through a data-driven method, the location with the longest nighttime residence time is set as the individual's home address, and the home address is set as the starting point of all the individual's movement trajectories; then the flow data of each individual from the starting point to other areas (such as work and entertainment places) are extracted to obtain the population movement relationship between different areas.

[0057] Furthermore, the individual mobility trajectory data are processed through a data-driven method as follows: the original Thiessen polygons established by the urban base stations in the 500-meter grid area within a certain period of time are merged, the location where each user stays longest at night is set as the user's home address, and the OD flow and home center flow every half hour in the 500-meter grid area of ​​the city are extracted; then they are mapped to each street unit to obtain the OD flow and home center flow every half hour in the 500-meter grid area within a certain period of time.

[0058] Secondly, the individual mobile trajectory data is processed through theoretical model methods. Specifically, according to the population distribution information and geographical location information of each street in the city in the census data, the network flow of the individual mobile trajectory dataset is extracted through theoretical model methods such as gravity model, radiation model and spatial proximity relationship model, among which:

[0059] The gravity model assumes that the intensity of population mobility between two regions is proportional to their respective population distributions and inversely proportional to the distance between them. The gravity model formula between two urban areas can usually be expressed as:

[0060]

[0061] In the above formula, T ij is the population mobility intensity between area i and area j in the city, m i and n j are the population distribution of region i and region j, r ij is the distance between region i and region j, a is a constant and can be 1.

[0062] The radiation model considers population mobility as a random process governed by joint probability, which depends on the population distribution of the departure point, destination and influence area. The radiation model formula is:

[0063]

[0064] T i =m i (N c / N) (3)

[0065] In the above formula, s ij represents the total population within the area with area i as the center and the distance between area i and area j as the radius. N is the total population of area i, N c is the commuter population of region i.

[0066] The spatial proximity relationship model is based on the first law of geography: all things are related to other things, but things that are close are more related than things that are far away. Therefore, the present invention considers the spatial relationship between things that are close, that is, the proximity relationship. Based on the proximity relationship of the regions, the present invention determines whether there is an edge between two regions (nodes) according to whether the two regions are in contact with each other, and obtains the adjacency matrix corresponding to the proximity relationship.

[0067] S3: Input the influenza case data and population mobility relationship into the Graph Neural Network (GNN), which extracts spatial features of the influenza case data based on the population mobility relationship to obtain the spatial scale information of the influenza case data;

[0068] In this step, graph neural networks are used for spatial dependency modeling, which has a wider range of applications and better generalization capabilities. The core idea of ​​graph neural networks is to learn a function that enables each node to aggregate its own features and its neighbor features to generate a new feature representation of the node. A neural network model containing multiple graph convolutional layers is constructed. Through continuous iteration and learning, the topological relationship on the graph can eventually be used to learn the input features of each node and make predictions. Graph neural networks are a neighbor aggregation strategy. The representation vector of a node is calculated by cyclically aggregating and transferring the representation vectors of its neighbor nodes. The framework of graph neural networks is message passing neural networks (MPNN). Message passing neural networks are a formal framework for spatial domain graph convolution. First, define an undirected graph G, where the feature vector of node v is x v , the edge feature is e vw , connect nodes v and w, N(v) represents the neighbor nodes of node v in graph G, t is the running time step, and the feature x of node v is v As the initial state of its hidden state After that, the update of the hidden state by the spatial domain graph convolution is expressed by the following formula:

[0069]

[0070]

[0071] The message propagation neural network decomposes the spatial graph convolution into two parts: message passing and state update operations, which are represented by the message function M l and the node update function U l Completed. Message function M l The function of is to aggregate the features of neighbor nodes to form a message vector, ready to be transmitted to the central node. Node update function U l The function of is to update the node representation at the current moment, combine the node representation at the current moment and the message obtained from the message function, so as to obtain the spatial scale information of the influenza case data.

[0072] Specifically, the graph neural network extracts spatial features of influenza case data based on population mobility relationships to obtain spatial scale information of influenza case data. Specifically, the individual mobility trajectory data of a set time range is converted into a weighted directed graph, whose vertices represent street-level areas and edges are used to capture mobility patterns. For example, the weight w of the edge (v, u) from vertex v to vertex u is v，w represents the total number of individuals with region v as their home address who move to region u at time t. At time t, the flow between regions u and v forms an edge that is multiplied by the number of cases in region u at time t provides a relative distribution, indicating how many infected people are likely to move from region u to region v. is a vector of node attributes containing the number of cases in region u in each of the past w weeks. The messages passed through the network use the combined scores from all regions to calculate the feature vector for each region, calculated as follows:

[0073]

[0074] Where A represents the adjacency matrix of regional population mobility flow, X t is a matrix whose rows contain the attributes of different regions. u ∈R w is a vector combining the number of influenza cases in region u and those moving toward region u. u The expression formula is:

[0075] x u =(x u w j，u +x u w i，u +…x u w v，u )+x u w u，u (7)

[0076] Where region u receives people from different regions, X t A vector containing past cases in the region. u ∈R w Represents an estimate of the number of new potential cases in region u, and is subdivided into cases received from other regions and new cases caused by mobility within region u. In order to update the vertex representation of each input graph, the embodiment of the present application uses the following neighborhood aggregation scheme:

[0077]

[0078] In the above formula, H i is a matrix containing the node representations of the previous layer, H 0 =X,W 0 is the matrix of the first layer of trainable parameters, and f is a nonlinear activation function, such as ReLU. By normalizing the adjacency matrix A, the sum of the weights of the incoming edges of each node is equal to 1. Based on a model with a K-order neighborhood aggregation layer, as the order of the neighborhood aggregation layer increases, the final node will capture more and more global information.

[0079] S4: Input the influenza case data and spatial scale information into the long short-term memory network (LSTM), and extract the time series relationship of the influenza case data through the LSTM;

[0080] In this step, the long short-term memory network is a recurrent neural network (RNN) that can be used to process sequence data. It can solve the gradient vanishing and gradient exploding problems in the long sequence training process. Compared with ordinary RNN, LSTM can achieve better performance in longer sequences. LSTM includes three stages: forgetting, selective memory, and output. The forgetting stage is used to selectively forget the input transmitted by the previous node. The selective memory stage is used to selectively "memorize" the input of the current stage. The output stage is used to determine which ones will be regarded as the output of the current state, and scale the output obtained in the previous stage through a tanh activation function. The LSTM calculation formula can be expressed as:

[0081] y i，t =LSTM(h i，t-n ,h i，t-n , ..., h i，t-1 ) (9)

[0082] Among them, h i，t-1 represents the influenza case data of the ith region in the t-1th time period, y i，t Represents the influenza case data predicted for the ith region in the tth time period.

[0083] In the embodiment of the present application, the time series relationship of the influenza case data extracted by the long short-term memory network is specifically as follows: the long short-term memory network is based on the input x at the current moment t and the output h of the hidden layer in the previous period t-1 To calculate the output h of the hidden layer at the current moment t , and a new input gate i is added t 、Forget gate t , output gate o t and memory unit c t . Among them, by inputting x t And the output h of the previous hidden layer t-1 A linear transformation is performed, and then the input gate is calculated through the activation function to control the extent to which the new state of the current calculation is updated to the memory unit. The calculation method of the forget gate and the output gate is similar to that of the input gate, which are used to control the extent to which the information in the previous memory unit is forgotten and the extent to which the current output depends on the current memory unit. The input gate, forget gate, and output gate all have their own parameters W and b, and the memory unit is mainly controlled by the input gate and the forget gate: the input gate controls the information that needs to be remembered in the current input sequence, and the forget gate controls the information that needs to be forgotten in the previous historical memory. The output h of the hidden layer of the long short-term memory network at time t tFinally, it is determined by the output gate and the memory unit. The update formula of each computing unit of the long short-term memory network is as follows:

[0084] i t =σ(W i [h t-1 , x t ]+b i ) (10)

[0085] f t =σ(W f [h t-1 , x t ]+b f ) (11)

[0086] o t =σ(W o [h t-1 , x t ]+b o ) (12)

[0087]

[0088]

[0089] h t =o t *tanh(C t ) (15)

[0090] Based on the above, for each input moment information X t , first use the graph neural network to extract the spatial scale information of each region, and then use the newly obtained input sequence of each region as the input of the long short-term memory network, and obtain the implicit features corresponding to each region through training and learning. At this time, each region shares the same long short-term memory network model, which helps to improve the generalization ability of the model, while reducing the model parameters and thus reducing the complexity of the model. While learning the original features in the time series relationship, the introduced spatial context information is also fully utilized to make the prediction model more effective. As the number of network layers increases, the final node function will capture more and more global information. However, in order to retain local intermediate information, the present invention splices the hidden states H1, H2 of the last time step of the two LSTM layers with the input historical information. The formula is as follows:

[0091] H=CONCAT(X t-n , X t-n+1 , ..., X t-1 ,H1,H2) (16)

[0092] The rows of the matrix H can be viewed as vertex representations encoding multi-scale structural information (including the initial features of the nodes), which are then passed to the output layer consisting of a two-layer fully connected network.

[0093] S5: The prediction results of the urban influenza incidence trend are obtained based on the spatial scale information and time series relationship of influenza case data;

[0094] Among them, since the embodiment of the present application adopts artificial intelligence deep learning to predict the trend of influenza incidence, the model parameters can be updated by learning new data after multiple predictions, making the model more intelligent and efficient.

[0095] Based on the above, the urban influenza incidence trend prediction method of the embodiment of the present application extracts population movement relationships through individual movement trajectory data in the city, and based on the population movement relationship, uses a graph convolutional neural network to obtain the spatial scale information of the influenza case data, and uses a long short-term memory network to extract the time series relationship of the influenza case data, and makes a refined prediction of the influenza trend based on the spatial scale information and time series relationship. The embodiment of the present application realizes the prediction of the development trend of influenza in the city with higher spatial resolution, completes the refined analysis of influenza, helps the government and public health departments to timely and accurately understand the development trend of influenza in the city, and carry out targeted epidemic prevention and control interventions, which can maximize the protection of people's lives, health and safety. Compared with the prior art, the present application has at least the following beneficial effects:

[0096] (1) After incorporating large-scale mobile phone location data, the traditional theoretical model method has improved its defect of being unable to simulate the population movement and interaction within the city in a refined manner, and the mobility of individuals has been restored more realistically.

[0097] (2) Compared with other data-driven modeling methods such as the spatial interaction modeling method based on OD flow, the spatial interaction modeling method of household center flow can construct a more concentrated graph structure flow at a higher spatial resolution at the urban scale, and can clearly capture multiple strong commuting flows, thereby effectively improving the prediction accuracy of the deep learning model of influenza within the city.

[0098] (3) The present invention only needs to obtain the location information of the mobile phone, and does not need to obtain information of multiple dimensions such as weekly average temperature, air pressure, rainfall, relative humidity, maximum temperature difference and sunshine time, thereby reducing the difficulty of data acquisition and processing.

[0099] See also Figure 2 , is a schematic diagram of the structure of the urban influenza incidence trend prediction system of the embodiment of the present application. The urban influenza incidence trend prediction system 40 of the embodiment of the present application includes:

[0100] Data collection module 41: used to obtain influenza case data in the city and collect individual movement trajectory data in the city;

[0101] Data processing module 42: used to process the individual movement trajectory data set through a data-driven method, obtain the home address of each individual, set the home address as the starting point of all movement trajectories of each individual, extract the flow data of each individual from the starting point to other areas, and obtain the population movement relationship between various areas;

[0102] Spatiotemporal feature extraction module 43: used to extract spatial scale information of influenza case data based on population mobility relationship using graph neural network, and to extract time series relationship of influenza case data using long short-term memory network;

[0103] Flu prediction module 44: used to obtain the prediction result of the trend of flu incidence in the city according to the spatial scale information and time series relationship of the flu case data.

[0104] See also Figure 3 , is a schematic diagram of the terminal structure of an embodiment of the present application. The terminal 50 includes a processor 51 and a memory 52 coupled to the processor 51.

[0105] The memory 52 stores program instructions for implementing the above-mentioned urban influenza incidence trend prediction method.

[0106] The processor 51 is used to execute the program instructions stored in the memory 52 to control the prediction of the trend of influenza incidence in the city.

[0107] The processor 51 may also be referred to as a CPU (Central Processing Unit). The processor 51 may be an integrated circuit chip having the ability to process signals. The processor 51 may also be a general-purpose processor, a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components. The general-purpose processor may be a microprocessor or the processor may also be any conventional processor, etc.

[0108] See also Figure 4, which is a schematic diagram of the structure of the storage medium of the embodiment of the present application. The storage medium of the embodiment of the present application stores a program file 61 that can implement all the above methods, wherein the program file 61 can be stored in the above storage medium in the form of a software product, including a number of instructions to enable a computer device (which can be a personal computer, a server, or a network device, etc.) or a processor (processor) to perform all or part of the steps of the methods of each embodiment of the present invention. The aforementioned storage medium includes: various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, or terminal devices such as a computer, a server, a mobile phone, and a tablet.

[0109] The above description of the disclosed embodiments enables professionals and technicians in the field to implement or use the present application. Various modifications to these embodiments will be apparent to professionals and technicians in the field, and the general principles defined in this application can be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application will not be limited to the embodiments shown in the present application, but will conform to the widest range consistent with the principles and novel features disclosed in the present application.

Claims

1. A method for predicting the trend of influenza incidence in a city, characterized in that: include: Obtain influenza case data within the city and collect individual movement trajectory data within the city; The individual movement trajectory data is processed by a data-driven method to obtain the home address of each individual, set the home address as the starting point of each individual movement trajectory, and extract the flow data of each individual from the starting point to other areas to obtain the population movement relationship between various areas; Based on the population movement relationship, a graph neural network is used to extract the spatial scale information of the influenza case data, and a long short-term memory network is used to extract the time series relationship of the influenza case data; Obtaining a prediction result of the trend of urban influenza incidence based on the spatial scale information and time series relationship of the influenza case data; The method of extracting the spatial scale information of the influenza case data using a graph neural network is specifically as follows: The framework of the graph neural network is a message propagation neural network, which extracts spatial scale information based on spatial domain graph convolution; define an undirected graph G, the feature vector of node v is x v , the edge feature is e vw , connect nodes v and w, N(v) represents the neighbor nodes of node v in graph G, t is the running time step, and the feature x of node v is v As the initial state of its hidden state After that, the update of the hidden state by the spatial domain graph convolution is expressed as: The message propagation neural network decomposes the spatial graph convolution into two parts: message passing and state update operations, respectively represented by the message function M l and the node update function U l Completed; the message function M l It is used to aggregate the features of neighbor nodes to form a message vector, which is ready to be transmitted to the central node; the node update function U l It is used to update the node representation at the current moment, combine the node representation at the current moment and the message obtained from the message function, and obtain the spatial scale information.

2. The method for predicting the trend of urban influenza incidence according to claim 1, characterized in that: The collection of individual movement trajectory data in the city includes: The individual movement trajectory data in the city is collected by using mobile devices; the mobile devices include mobile phones or smart watches; the individual movement trajectory data includes each individual's mobile phone number, signaling timestamp, and base station longitude and latitude.

3. The method for predicting the trend of urban influenza incidence according to claim 2, characterized in that: The processing of the individual movement trajectory data by the data-driven method is specifically as follows: The original Thiessen polygons established by the urban base stations in the grid area of ​​the set time period and the set distance are merged; the individual movement trajectory data is divided according to the residence time, the location where each individual stays the longest at night is set as the individual's home address, and the OD flow and home center flow at the set interval time in the grid area of ​​the set distance are extracted.

4. The method for predicting the trend of urban influenza incidence according to claim 1, characterized in that: The processing of the individual movement trajectory data by a data-driven method further includes: According to the population distribution information and geographic location information of each street in the city in the census data, network flow extraction is performed on the individual movement trajectory data through a theoretical model method; the theoretical model method includes a gravity model, a radiation model or a spatial proximity relationship model.

5. The method for predicting the trend of urban influenza incidence according to claim 4, characterized in that: The spatial scale information of the influenza case data extracted by using a graph neural network is specifically: The individual movement trajectory data of a set time range is converted into a weighted directed graph, where vertices represent street-level areas and edges are used to capture movement patterns; at time t, the flow between areas u and v forms an edge, which is multiplied by the number of cases in area u at time t represents how many infected people may move from area u to area v; let is a vector of node attributes containing the number of cases in region u in each of the past w weeks; the message passed through the message propagation neural network uses the combined scores from all regions to calculate the feature vector for each region: Where A represents the adjacency matrix of regional population mobility flow, X t is a matrix whose rows contain the attributes of different regions; x u ∈R w is a vector combining the number of influenza cases in region u and those moving toward region u, x u The expression formula is: x u =(x u w j,u +x u w i,u +…x u w v,u )+x u w u,u where x u ∈R w represents the estimate of the number of new potential cases in region u.

6. The method for predicting the trend of urban influenza incidence according to claim 5, characterized in that: The time series relationship of extracting the influenza case data using the long short-term memory network is specifically as follows: The long short-term memory network is based on the input x at the current moment t and the output h of the hidden layer in the previous period t-1 To calculate the output h of the hidden layer at the current moment t , the calculation formula of the long short-term memory network is: y i,t =LSTM(h i,t-n ,h i,t-n ,...,h i,t-1 ) Among them, h i,t-1 represents the influenza case data of the ith region in the t-1th time period, y i,t Represents the influenza case data predicted for the ith region in the tth time period.

7. A system for predicting the incidence trend of urban influenza using the method for predicting the incidence trend of urban influenza according to claim 1, characterized in that: include: Data collection module: used to obtain influenza case data in the city and collect individual movement trajectory data in the city; Data processing module: used to process the individual movement trajectory data through a data-driven method, obtain the home address of each individual, set the home address as the starting point of each individual movement trajectory, and extract the flow data of each individual from the starting point to other areas, and obtain the population movement relationship between various areas; A spatiotemporal feature extraction module is used to extract the spatial scale information of the influenza case data based on the population movement relationship using a graph neural network, and to extract the time series relationship of the influenza case data using a long short-term memory network; Flu prediction module: used to obtain the prediction result of the trend of flu incidence in the city according to the spatial scale information and time series relationship of the flu case data.

8. A terminal, characterized in that: The terminal includes a processor and a memory coupled to the processor, wherein: The memory stores program instructions for implementing the urban influenza incidence trend prediction method according to any one of claims 1 to 6; The processor is used to execute the program instructions stored in the memory to control the influenza incidence trend prediction in influenza cities.

9. A storage medium, characterized in that: The method stores program instructions executable by a processor, wherein the program instructions are used to execute the method for predicting the trend of influenza incidence in a city as claimed in any one of claims 1 to 6.

Images