Air quality evaluation method, device and equipment based on graph neural network and storage medium

Abstract

The invention relates to an air quality evaluation method, device and equipment based on a graph neural network and a storage medium, and belongs to the technical field of environment monitoring, and the method comprises the following steps: after obtaining concentration data of various air in a target area in a target time period, processing the concentration data into graph data, and inputting the graph data into the trained air quality evaluation model to obtain the air quality grade of the target area in the target time period. According to the method, the data relationship in the model is processed by using the graph neural network model, so that the accuracy of estimating the air quality grade can be improved on the basis of quickly evaluating the air quality grade of the specified area.

Description

Technical Field

[0001] The present application relates to the field of environmental monitoring technology, and in particular to an air quality assessment method, device, equipment and storage medium based on a graph neural network. Background Art

[0002] In recent years, environmental issues have become a growing concern. Many places have established air quality monitoring platforms to measure and evaluate the air quality in their areas.

[0003] In related technologies, models based on artificial intelligence are used to evaluate air quality. For example, air quality evaluation models are constructed based on recurrent neural networks (RNN), long short-term memory (LSTM) or gated recurrent units (GRU). However, these methods usually only give indicators of the concentration of harmful substances in the air, and cannot intuitively give the level of air quality. Summary of the invention

[0004] The present application provides an air quality assessment method, device, electronic device and storage medium based on graph neural network, which can accurately and quickly assess the air quality level of a specified area. The technical solution of the present application is as follows:

[0005] In a first aspect, the present application provides an air quality assessment method based on a graph neural network, the method comprising:

[0006] Acquire air concentration data collected by IoT devices in a target area within a target time period, wherein the air concentration data includes time series concentration data of carbon monoxide CO, time series concentration data of nitrogen dioxide NO2, time series concentration data of ozone O3, and time series concentration data of PM 2.5;

[0007] Performing data preprocessing on the air concentration data to obtain first graph data and second graph data, wherein the first graph data is used to represent the concentration value of each element in the air corresponding to each time point in the target time period, and the second graph data is used to represent multiple time points in the target time period;

[0008] Inputting the first graph data and the second graph data into an air quality assessment model, wherein the air quality assessment model is a graph neural network model obtained through deep learning training;

[0009] Through the graph data transmission layer in the air quality assessment model, data aggregation is performed on the first graph data and the second graph data respectively to obtain a first aggregation feature corresponding to the first graph data and a second aggregation feature corresponding to the second graph data, wherein the first aggregation feature is used to characterize the aggregated air concentration corresponding to each time point in the target time period, and the second aggregation feature is used to characterize the correlation relationship between multiple time points in the target time period;

[0010] Performing feature expansion on the first aggregated feature through the graph convolution layer in the air quality assessment model to obtain an extended feature of the first aggregated feature, and performing feature extraction on the second aggregated feature based on the extended feature to obtain a feature vector of the second aggregated feature;

[0011] The feature vector is processed through the artificial neural network layer in the air quality assessment model to output the air quality level of the target area within the target time period.

[0012] In some embodiments, a node in the first graph data is used to represent the concentration value of an element, and an edge is used to represent that the probability of obtaining the same air quality level based on the concentration values ​​corresponding to the two nodes connected by the edge meets the condition;

[0013] In the second graph data, a node is used to represent a time point, and an edge is used to represent the time relationship between two time points.

[0014] In some embodiments, the performing data aggregation on the first graph data and the second graph data respectively through the graph data transmission layer in the air quality assessment model to obtain a first aggregation feature corresponding to the first graph data and a second aggregation feature corresponding to the second graph data includes:

[0015] The first transmission unit in the graph data transmission layer aggregates the first graph data based on the following formulas (1) to (3) to obtain the first aggregated feature:

[0016]

[0017]

[0018]

[0019] Among them, A 影响因素 represents the adjacency matrix corresponding to the first graph data, I 影响因素 represents the unit matrix corresponding to the first graph data, D 影响因素 represents the degree matrix corresponding to the first graph data, L 影响因素is a normalized matrix, used to represent the first aggregated feature;

[0020] The second graph data is aggregated by the second transmission unit in the graph data transmission layer based on the following formulas (4) to (6) to obtain the second aggregation feature:

[0021]

[0022]

[0023]

[0024] Among them, A 时间序列 represents the adjacency matrix corresponding to the second graph data, I 时间序列 Denotes the identity matrix corresponding to the second graph data, D 时间序列 represents the degree matrix corresponding to the second graph data, L 时间序列 is a normalized matrix used to represent the second aggregated feature.

[0025] In some embodiments, the step of performing feature expansion on the first aggregate feature through the graph convolution layer in the air quality assessment model to obtain an extended feature of the first aggregate feature, and performing feature extraction on the second aggregate feature based on the extended feature to obtain a feature vector of the second aggregate feature includes:

[0026] Through the first convolution unit of the graph convolution layer, based on the following formula (7) and formula (8), the first aggregated feature is extended to obtain the extended feature:

[0027]

[0028] H i+2 =tanh(H i+1 W 影响因素聚合 +b 影响因素聚合 ) (8)

[0029] Among them, W 影响因素聚合 represents the sum of M random matrices, RS N×F represents a function that returns an N×F random matrix, where N represents the number of nodes in the first graph data, F represents the number of features of the first graph data after the feature expansion, H represents the input matrix corresponding to the first graph data, and H i+2 represents the extended feature, b 影响因素聚合 Used to fit the deviations that occur during training;

[0030] The second convolution unit of the graph convolution layer performs feature extraction on the second aggregated feature based on the extended feature and the following formula (9) and formula (10) to obtain the feature vector:

[0031]

[0032] K i+1 =tanh(K i W 时间序列聚合 +b 时间序列聚合 ) (10)

[0033] Among them, W 时间序列聚合 represents the sum of M random matrices, RS F×z represents a function that returns an F×Z random matrix, where Z represents the number of features of each node after the feature extraction, K represents a matrix containing the features of the second graph data, and K i is a matrix obtained based on the extended features and the Relu activation function, K i+1 represents the feature vector, b 时间序列聚合 Used to fit the deviations that occur during training.

[0034] In some embodiments, the feature vector is processed by an artificial neural network layer in the air quality assessment model to output the air quality level of the target area within the target time period, including:

[0035] The feature vector is processed through the artificial neural network layer based on the Relu activation function and the Softmax classifier to obtain the air quality level. The artificial neural network layer includes an input layer, a hidden layer and an output layer, and the Dropout value is 0.2.

[0036] In some embodiments, the method further comprises:

[0037] The air quality assessment model is trained based on any of the following loss functions: cross entropy loss function, mean square error loss function, Hinge loss function.

[0038] In a second aspect, the present application provides an air quality assessment device based on a graph neural network, the device comprising:

[0039] An acquisition module is used to acquire air concentration data collected by IoT devices in a target area within a target time period, wherein the air concentration data includes time series concentration data of carbon monoxide CO, time series concentration data of nitrogen dioxide NO2, time series concentration data of ozone O3, and time series concentration data of PM 2.5;

[0040] A preprocessing module, used to perform data preprocessing on the air concentration data to obtain first graph data and second graph data, wherein the first graph data is used to represent the concentration value of each element in the air corresponding to each time point in the target time period, and the second graph data is used to represent multiple time points in the target time period;

[0041] An input module, used to input the first graph data and the second graph data into an air quality assessment model, wherein the air quality assessment model is a graph neural network model obtained through deep learning training;

[0042] a transmission module, configured to respectively perform data aggregation on the first graph data and the second graph data through a graph data transmission layer in the air quality assessment model to obtain a first aggregation feature corresponding to the first graph data and a second aggregation feature corresponding to the second graph data, wherein the first aggregation feature is used to characterize the aggregated air concentration corresponding to each time point in the target time period, and the second aggregation feature is used to characterize the correlation relationship between multiple time points in the target time period;

[0043] A convolution module, configured to perform feature expansion on the first aggregate feature through a graph convolution layer in the air quality assessment model to obtain an extended feature of the first aggregate feature, and perform feature extraction on the second aggregate feature based on the extended feature to obtain a feature vector of the second aggregate feature;

[0044] The output module is used to process the feature vector through the artificial neural network layer in the air quality assessment model and output the air quality level of the target area within the target time period.

[0045] In some embodiments, a node in the first graph data is used to represent the concentration value of an element, and an edge is used to represent that the probability of obtaining the same air quality level based on the concentration values ​​corresponding to the two nodes connected by the edge meets the condition;

[0046] In the second graph data, a node is used to represent a time point, and an edge is used to represent the time relationship between two time points.

[0047] In a third aspect, the present application provides an electronic device comprising a processor and a memory, wherein the memory stores program code, and the processor is used to execute the program code to implement the above-mentioned air quality assessment method based on graph neural network.

[0048] In a fourth aspect, the present application provides a computer-readable storage medium, which includes: when the program code in the computer-readable storage medium is executed by a processor of an electronic device, the electronic device is enabled to perform the above-mentioned air quality assessment method based on graph neural network.

[0049] It should be understood that the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the present application. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] Figure 1 It is a schematic diagram of the implementation environment of the air quality assessment method based on graph neural network;

[0051] Figure 2 It is a flow chart of an air quality assessment method based on graph neural network;

[0052] Figure 3 is a schematic diagram of the structure of first graph data;

[0053] Figure 4 is a schematic diagram of the structure of the second graph data;

[0054] Figure 5 It is a structural schematic diagram of an air quality assessment device based on a graph neural network;

[0055] Figure 6 It is a structural diagram of an electronic device. DETAILED DESCRIPTION

[0056] In order to enable ordinary persons in the art to better understand the technical solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the accompanying drawings.

[0057] The data involved in this application may be data authorized by the user or fully authorized by all parties.

[0058] Figure 1 is a schematic diagram of the implementation environment of the air quality assessment method based on graph neural network, see Figure 1 The implementation environment includes: an Internet of Things device 101 and a server 102, and the Internet of Things device 101 and the server 102 can be directly or indirectly connected through wired or wireless communication.

[0059] The IoT device 101 is equipped with a sensor that can collect concentration data of various elements in the air in a specified area. After collecting the air concentration data, the IoT device 101 sends the collected air concentration data to the server 102. The server 102 is equipped with a trained air quality assessment model. The server 102 can pre-process the received air concentration data and use the trained air quality assessment model to output the air quality level of the specified area based on the pre-processed data. For example, the server 102 is an independent physical server, or a server cluster composed of multiple physical servers, or a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, and big data and artificial intelligence platforms, but the present application is not limited thereto. In addition, the server 102 can also train the initial neural network based on the training data set to obtain a trained air quality assessment model.

[0060] Figure 2 It is a flow chart of an air quality assessment method based on graph neural network, such as Figure 2 As shown, the method is executed by the server and includes the following steps 201 to 206.

[0061] 201. Obtain air concentration data collected by IoT devices in a target area within a target time period.

[0062] Among them, the air concentration data includes the time series concentrations of multiple elements in the air. Schematically, the air concentration data includes the time series concentration data of carbon monoxide CO, the time series concentration data of nitrogen dioxide NO2, the time series concentration data of ozone O3 and the time series concentration data of PM 2.5. The target time period is a preset time period. For example, the target time period is 8 hours, which is not limited. The target area refers to a specified area. For example, the target area refers to a factory, an apartment building, a business district area, etc. The present application does not limit the basis for the division of the target area. At least one Internet of Things device is deployed in the target area, and the Internet of Things device is equipped with a sensor that can collect concentration data of various air in the target area and send the collected air concentration data to the server.

[0063] 202. Perform data preprocessing on the air concentration data to obtain first graph data and second graph data, wherein the first graph data is used to represent the concentration value of each element in the air corresponding to each time point in the target time period, and the second graph data is used to represent multiple time points in the target time period.

[0064] Among them, by preprocessing the air concentration data, the air concentration data is expressed in the form of graph data, which is convenient for subsequent air quality assessment through an air quality assessment model.

[0065] In some embodiments, a node in the first graph data is used to represent the concentration value of an element, and an edge is used to represent the probability of obtaining the same air quality level based on the concentration values ​​corresponding to the two nodes connected by the edge. Figure 3 As shown, Figure 3 It is a structural diagram of the first graph data. For any time point in the target time period, four nodes are used to represent the concentration values ​​of four elements, CO, NO2, O3 and PM 2.5. If two concentration values ​​represented by two nodes may lead to the same prediction result, there will be an edge connecting them, that is, the edge between the two nodes represents that the same air quality level may be obtained based on the concentration values ​​corresponding to the two nodes, that is, the aforementioned probability meets the conditions. For example, the condition is set to a probability of 90%, which is not limited. It should be understood that since the concentration values ​​of multiple elements will comprehensively affect the final air quality level, the first graph data can also be understood as an influencing factor for the air quality level.

[0066] In some implementations, a node in the second graph data is used to represent a time point, and an edge is used to represent the time relationship between two time points. Figure 4 As shown, Figure 4 It is a structural diagram of the second graph data. For example, this application uses 10 consecutive time points within the target time period for prediction. The interval between each time point can be set according to demand, and this application does not limit this. Among them, a node corresponds to a time point, and the edge between two nodes represents the time relationship. That is, the second graph data can also be understood as a feature relationship data based on time series.

[0067] 203. Input the first graph data and the second graph data into an air quality assessment model, where the air quality assessment model is a graph neural network model trained based on deep learning.

[0068] Among them, the air quality assessment model is a graph neural network model GNN, including a graph transmission layer, a graph convolution layer and an artificial neural network ANN. The input of the air quality assessment model is the aforementioned first graph data and the second graph data, and the output is the air quality level. Schematically, the air quality assessment model is trained based on a supervised training method, wherein the air quality level is pre-set, for example, the air quality level is divided into 1-10 levels, a total of 10 levels, and the lower the level, the worse the air quality. By inputting the first graph data and the second graph data into the air quality assessment model, the corresponding air quality level can be directly obtained, which is more efficient.

[0069] It should be noted that this application does not limit the loss function used in the training process of the air quality assessment model. For example, the air quality assessment model is trained based on any of the following loss functions: cross entropy loss function, mean square error loss function, Hinge loss function, and so on.

[0070] The detailed processing of the air quality assessment model is introduced below through steps 204 to 206.

[0071] 204. Through the graph data transmission layer in the air quality assessment model, the first graph data and the second graph data are respectively aggregated to obtain a first aggregation feature corresponding to the first graph data and a second aggregation feature corresponding to the second graph data. The first aggregation feature is used to characterize the aggregated air concentration corresponding to each time point in the target time period, and the second aggregation feature is used to characterize the correlation between multiple time points in the target time period.

[0072] The graph data transmission layer includes a first transmission unit and a second transmission unit. The first graph data and the second graph data are respectively aggregated by the two transmission units, that is, matrix normalization. Schematically, the following steps are included:

[0073] Step 1: Through the first transmission unit in the graph data transmission layer, based on the following formulas (1) to (3), the first graph data is aggregated to obtain a first aggregation feature:

[0074]

[0075]

[0076]

[0077] Among them, A 影响因素 Represents the adjacency matrix corresponding to the first graph data, I 影响因素 Denotes the identity matrix corresponding to the first graph data, D 影响因素 Represents the degree matrix corresponding to the first graph data, L 影响因素 is a normalized matrix used to represent the first aggregated feature.

[0078] The above step 1 is the stage of aggregating the air concentration data corresponding to each time point.

[0079] Step 2: Through the second transmission unit in the graph data transmission layer, based on the following formulas (4) to (6), the second graph data is aggregated to obtain a second aggregate feature:

[0080]

[0081]

[0082]

[0083] Among them, A 时间序列 Represents the adjacency matrix corresponding to the second graph data, I 时间序列 Denotes the identity matrix corresponding to the data in the second figure, D 时间序列 Represents the degree matrix corresponding to the second graph data, L 时间序列 is a normalized matrix used to represent the second aggregated feature.

[0084] The above step 2 is also the stage of aggregating information at multiple time points within the target time period.

[0085] 205. Perform feature expansion on the first aggregate feature through the graph convolution layer in the air quality assessment model to obtain the extended feature of the first aggregate feature, perform feature extraction on the second aggregate feature to obtain the feature vector of the second aggregate feature.

[0086] The graph convolution layer includes a first convolution unit and a second convolution unit, through which the first aggregate feature is expanded and the second aggregate feature is extracted based on the expanded feature. Schematically, the following steps are included:

[0087] Step 1: Through the first convolution unit of the graph convolution layer, based on the following formulas (7) and (8), the first aggregated feature is expanded to obtain the extended feature:

[0088]

[0089] H i+2 =tanh(H i+1 W 影响因素聚合 +b 影响因素聚合 ) (8)

[0090] Among them, W 影响因素聚合 represents the sum of M random matrices (M refers to the number of nodes in the second graph data), RS N×F represents a function that returns an N×F random matrix, where N represents the number of nodes in the first graph data, F represents the number of features of the first graph data after feature expansion, H represents the input matrix corresponding to the first graph data, and H i+2 represents the extended feature, b 影响因素聚合 Used to fit the deviations that occur during training.

[0091] Step 2: Through the second convolution unit of the graph convolution layer, based on the extended features and the following formulas (9) and (10), the second aggregated features are extracted to obtain a feature vector:

[0092]

[0093] K i+1 =tanh(K i W 时间序列聚合 +b 时间序列聚合 ) (10)

[0094] Among them, W 时间序列聚合 represents the sum of M random matrices, RS F×Z represents a function that returns a F×Z random matrix, where Z represents the number of features of each node after feature extraction, K represents the matrix containing the features of the second graph data, and K i is based on the extended feature (i.e., H calculated by the previous formula (8) i+2 ) and the matrix obtained by the Relu activation function, K i+1 represents the feature vector, b 时间序列聚合 Used to fit the deviations that occur during training.

[0095] 206. The feature vector is processed through the artificial neural network layer in the air quality assessment model to output the air quality level of the target area within the target time period.

[0096] The artificial neural network layer includes an input layer, a hidden layer, and an output layer, and the Dropout value is 0.2. The feature vector is input into the artificial neural network layer, and the feature vector is processed by the artificial neural network layer based on the Relu activation function and the Softmax classifier to obtain the air quality level.

[0097] In the method provided in the embodiment of the present application, after obtaining the concentration data of various air in the target area within the target time period, these concentration data are processed into graph data, and the graph data is input into the trained air quality assessment model to obtain the air quality level of the target area within the target time period. In the above method, by using a graph neural network model to process the data relationship in the model, the accuracy of the air quality level can be improved on the basis of quickly evaluating the air quality level of the specified area.

[0098] Figure 5 This is a schematic diagram of the structure of an air quality assessment device based on a graph neural network. Figure 5 The device includes an acquisition module 501, a preprocessing module 502, an input module 503, a transmission module 504, a convolution module 505 and an output module 506.

[0099] An acquisition module 501 is used to acquire air concentration data collected by the IoT devices in the target area within a target time period, wherein the air concentration data includes time series concentration data of carbon monoxide CO, time series concentration data of nitrogen dioxide NO2, time series concentration data of ozone O3, and time series concentration data of PM 2.5;

[0100] A preprocessing module 502 is used to perform data preprocessing on the air concentration data to obtain first graph data and second graph data, wherein the first graph data is used to represent the concentration value of each element in the air corresponding to each time point in the target time period, and the second graph data is used to represent multiple time points in the target time period;

[0101] An input module 503, configured to input the first graph data and the second graph data into an air quality assessment model, wherein the air quality assessment model is a graph neural network model obtained through deep learning training;

[0102] A transmission module 504 is used to perform data aggregation on the first graph data and the second graph data respectively through the graph data transmission layer in the air quality assessment model to obtain a first aggregation feature corresponding to the first graph data and a second aggregation feature corresponding to the second graph data, wherein the first aggregation feature is used to characterize the aggregated air concentration corresponding to each time point in the target time period, and the second aggregation feature is used to characterize the correlation relationship between multiple time points in the target time period;

[0103] A convolution module 505 is used to perform feature expansion on the first aggregate feature through a graph convolution layer in the air quality assessment model to obtain an extended feature of the first aggregate feature, and perform feature extraction on the second aggregate feature based on the extended feature to obtain a feature vector of the second aggregate feature;

[0104] The output module 506 is used to process the feature vector through the artificial neural network layer in the air quality assessment model, and output the air quality level of the target area within the target time period.

[0105] In some embodiments, the above device is also used to implement the above Figure 2 Other optional implementations in the method embodiment shown are not described in detail here.

[0106] Figure 6It is a structural diagram of an electronic device. The electronic device can be configured as the above-mentioned server, wherein the electronic device 600 may have relatively large differences due to different configurations or performances, and may include one or more processors (Central Processing Units, CPU) 601 and one or more memories 602, wherein the one or more memories 602 store at least one program code, and the at least one program code is loaded and executed by the one or more processors 601 to implement the process executed by the server in the air quality assessment method based on the graph neural network provided by the above-mentioned various method embodiments. Of course, the electronic device 600 may also have components such as a wired or wireless network interface, a keyboard, and an input and output interface for input and output. The electronic device 600 may also include other components for realizing the functions of the device, which will not be repeated here.

[0107] Those skilled in the art will readily appreciate other embodiments of the present application after considering the specification and practicing the invention disclosed herein. The present application is intended to cover any modification, use or adaptation of the present application, which follows the general principles of the present application and includes common knowledge or customary techniques in the art that are not disclosed in the present application. The specification and examples are intended to be exemplary only, and the true scope and spirit of the present application are indicated by the following claims.

[0108] It should be understood that the present application is not limited to the precise structures that have been described above and shown in the drawings, and that various modifications and changes may be made without departing from the scope thereof. The scope of the present application is limited only by the appended claims.

Claims

1. An air quality assessment method based on graph neural network, characterized in that: The method comprises: Acquire air concentration data collected by IoT devices in a target area within a target time period, wherein the air concentration data includes time series concentration data of carbon monoxide CO, time series concentration data of nitrogen dioxide NO2, time series concentration data of ozone O3, and time series concentration data of PM 2.5; Performing data preprocessing on the air concentration data to obtain first graph data and second graph data, wherein the first graph data is used to represent the concentration value of each element in the air corresponding to each time point in the target time period, and the second graph data is used to represent multiple time points in the target time period; Inputting the first graph data and the second graph data into an air quality assessment model, wherein the air quality assessment model is a graph neural network model obtained through deep learning training; Through the graph data transmission layer in the air quality assessment model, data aggregation is performed on the first graph data and the second graph data respectively to obtain a first aggregation feature corresponding to the first graph data and a second aggregation feature corresponding to the second graph data, wherein the first aggregation feature is used to characterize the aggregated air concentration corresponding to each time point in the target time period, and the second aggregation feature is used to characterize the correlation relationship between multiple time points in the target time period; Performing feature expansion on the first aggregated feature through the graph convolution layer in the air quality assessment model to obtain an extended feature of the first aggregated feature, and performing feature extraction on the second aggregated feature based on the extended feature to obtain a feature vector of the second aggregated feature; The feature vector is processed through the artificial neural network layer in the air quality assessment model to output the air quality level of the target area within the target time period.

2. The method according to claim 1, characterized in that A node in the first graph data is used to represent the concentration value of an element, and an edge is used to represent that the probability of obtaining the same air quality level based on the concentration values ​​corresponding to the two nodes connected by the edge meets the condition; In the second graph data, a node is used to represent a time point, and an edge is used to represent the time relationship between two time points.

3. The method according to claim 1, characterized in that The step of respectively aggregating the first graph data and the second graph data through the graph data transmission layer in the air quality assessment model to obtain a first aggregation feature corresponding to the first graph data and a second aggregation feature corresponding to the second graph data includes: The first transmission unit in the graph data transmission layer aggregates the first graph data based on the following formulas (1) to (3) to obtain the first aggregated feature: Among them, A 影响因素 represents the adjacency matrix corresponding to the first graph data, I 影响因素 represents the unit matrix corresponding to the first graph data, D 影响因素 represents the degree matrix corresponding to the first graph data, L 影响因素 is a normalized matrix, used to represent the first aggregated feature; The second graph data is aggregated by the second transmission unit in the graph data transmission layer based on the following formulas (4) to (6) to obtain the second aggregation feature: Among them, A 时间序列 represents the adjacency matrix corresponding to the second graph data, I 时间序列 Denotes the identity matrix corresponding to the second graph data, D 时间序列 represents the degree matrix corresponding to the second graph data, L 时间序列 is a normalized matrix used to represent the second aggregated feature.

4. The method according to claim 1, characterized in that: The step of performing feature expansion on the first aggregate feature through the graph convolution layer in the air quality assessment model to obtain an extended feature of the first aggregate feature, and performing feature extraction on the second aggregate feature based on the extended feature to obtain a feature vector of the second aggregate feature includes: Through the first convolution unit of the graph convolution layer, based on the following formula (7) and formula (8), the first aggregated feature is extended to obtain the extended feature: H i+2 =tanh(H i+1 W 影响因素聚合 +b 影响因素聚合 ) (8), Among them, W 影响因素聚合 represents the sum of M random matrices, RS N×F represents a function that returns an N×F random matrix, where N represents the number of nodes in the first graph data, F represents the number of features of the first graph data after the feature expansion, H represents the input matrix corresponding to the first graph data, and H i+2 represents the extended feature, b 影响因素聚合 Used to fit the deviations that occur during training; The second convolution unit of the graph convolution layer performs feature extraction on the second aggregated feature based on the extended feature and the following formula (9) and formula (10) to obtain the feature vector: K i+1 =tanh(K i W 时间序列聚合 +b 时间序列聚合 ) (10) Among them, W 时间序列聚合 represents the sum of M random matrices, RS F×z represents a function that returns an F×Z random matrix, where Z represents the number of features of each node after the feature extraction, K represents a matrix containing the features of the second graph data, and K i is a matrix obtained based on the extended features and the Relu activation function, K i+1 represents the feature vector, b 时间序列聚合 Used to fit the deviations that occur during training.

5. The method according to claim 1, characterized in that The feature vector is processed by the artificial neural network layer in the air quality assessment model to output the air quality level of the target area within the target time period, including: The feature vector is processed through the artificial neural network layer based on the Relu activation function and the Softmax classifier to obtain the air quality level. The artificial neural network layer includes an input layer, a hidden layer and an output layer, and the Dropout value is 0.

2.

6. The method according to claim 1, characterized in that The method further comprises: The air quality assessment model is trained based on any of the following loss functions: cross entropy loss function, mean square error loss function, Hinge loss function.

7. An air quality assessment device based on graph neural network, characterized in that: The device comprises: An acquisition module is used to acquire air concentration data collected by IoT devices in a target area within a target time period, wherein the air concentration data includes time series concentration data of carbon monoxide CO, time series concentration data of nitrogen dioxide NO2, time series concentration data of ozone O3, and time series concentration data of PM 2.5; A preprocessing module, used to perform data preprocessing on the air concentration data to obtain first graph data and second graph data, wherein the first graph data is used to represent the concentration value of each element in the air corresponding to each time point in the target time period, and the second graph data is used to represent multiple time points in the target time period; An input module, used to input the first graph data and the second graph data into an air quality assessment model, wherein the air quality assessment model is a graph neural network model obtained through deep learning training; a transmission module, configured to respectively perform data aggregation on the first graph data and the second graph data through a graph data transmission layer in the air quality assessment model to obtain a first aggregation feature corresponding to the first graph data and a second aggregation feature corresponding to the second graph data, wherein the first aggregation feature is used to characterize the aggregated air concentration corresponding to each time point in the target time period, and the second aggregation feature is used to characterize the correlation relationship between multiple time points in the target time period; A convolution module, configured to perform feature expansion on the first aggregate feature through a graph convolution layer in the air quality assessment model to obtain an extended feature of the first aggregate feature, and perform feature extraction on the second aggregate feature based on the extended feature to obtain a feature vector of the second aggregate feature; The output module is used to process the feature vector through the artificial neural network layer in the air quality assessment model and output the air quality level of the target area within the target time period.

8. The device according to claim 7, characterized in that A node in the first graph data is used to represent the concentration value of an element, and an edge is used to represent that the probability of obtaining the same air quality level based on the concentration values ​​corresponding to the two nodes connected by the edge meets the condition; In the second graph data, a node is used to represent a time point, and an edge is used to represent the time relationship between two time points.

9. An electronic device, characterized in that: The electronic device includes a processor and a memory, the memory stores program code, and the processor is used to execute the program code to implement the air quality assessment method based on graph neural network as described in any one of claims 1 to 6.

10. A computer-readable storage medium, characterized in that: When the program code in the computer-readable storage medium is executed by a processor of an electronic device, the electronic device is enabled to execute the air quality assessment method based on a graph neural network as described in any one of claims 1 to 6.

Images