Ozone concentration prediction method and device based on space-time diagram convolutional network model, equipment and medium

Abstract

The invention relates to the technical field of ozone concentration prediction, in particular to an ozone concentration prediction method, device, equipment and medium based on a space-time diagram convolutional network model.The method comprises the steps that historical ozone concentration and meteorological data of a target area are obtained, and training and prediction samples are constructed after standardization; a space diagram structure is established, nodes represent monitoring points, and edges represent the monitoring points with correlation exceeding a threshold value; extracting a spatial dependency relationship by adopting GCN and GAT, and generating time sequence data; tCN or RNN is adopted to extract time-dependent features, and spatial-temporal feature comprehensive representation is obtained; training a model by utilizing the training sample and the spatial-temporal characteristics, adopting MSE (mean square error) as a loss function, adjusting parameters by an Adam optimizer, and optimizing hyper-parameters to minimize errors; newest meteorological and historical data are input, future ozone concentration is predicted, GNN extracts spatial features, RNN processes a time sequence, time sequence data can be processed in parallel through convolution operation, and high calculation efficiency is achieved.

Description

Technical Field

[0001] The present invention relates to the technical field of ozone concentration prediction, and in particular to an ozone concentration prediction method, device, equipment and medium based on a spatiotemporal graph convolutional network model. Background Art

[0002] Ozone (O3) is one of the important components in the Earth's atmosphere, which is divided into tropospheric ozone and stratospheric ozone. Tropospheric ozone is a major component of air pollution and has a significant impact on human health and the ecological environment. Changes in ozone concentration are affected by many factors, including meteorological conditions, pollutant emissions, and chemical reactions. Accurately predicting changes in ozone concentration is of great significance for air pollution prevention and control and public health protection.

[0003] In recent years, with the development of machine learning and deep learning technologies, more and more researchers have begun to apply these technologies to the field of environmental science, especially air pollution prediction. However, traditional machine learning methods have limitations when dealing with complex spatiotemporal data. For example, although traditional methods such as linear regression, support vector machines, and random forests can predict ozone concentrations to a certain extent, they are mainly based on statistical characteristics and simple nonlinear mappings and cannot effectively capture the complex spatiotemporal dependencies of ozone concentrations. These methods perform poorly when dealing with high-dimensional, nonlinear, and time-varying features and large amounts of data, resulting in limited prediction accuracy.

[0004] The information disclosed in this background technology section is only intended to deepen the understanding of the overall background technology of the present invention, and should not be regarded as acknowledging or suggesting in any form that the information constitutes the prior art already known to those skilled in the art. Summary of the invention

[0005] The present invention provides an ozone concentration prediction method, device, equipment and medium based on a spatiotemporal graph convolutional network model, thereby effectively solving the problems in the background technology.

[0006] In order to achieve the above object, the technical solution adopted by the present invention is: a method for predicting ozone concentration based on a spatiotemporal graph convolutional network model, comprising the following steps:

[0007] S10: Obtain historical ozone concentration data and corresponding meteorological data of the target area, perform standardization preprocessing on the historical ozone concentration data and the meteorological data, and construct training samples and prediction samples based on a sliding window method;

[0008] S20: Based on the geographical locations of the monitoring points in the training samples and the correlation coefficients therebetween, a spatial graph structure is established, wherein nodes in the spatial graph structure represent monitoring points, and edges represent spatial correlations between monitoring points, and when the correlation coefficients between monitoring points exceed a preset threshold, an edge connection is established;

[0009] S30: using GCN and GAT to perform convolution operation on the spatial graph structure, extracting the spatial dependency relationship between monitoring points, and obtaining time series data containing spatial features;

[0010] S40: Based on the time series data including spatial features, using a TCN or RNN model, extracting time-dependent features in the time series data to obtain a comprehensive representation including spatiotemporal features;

[0011] S50: using the training sample and the comprehensive representation containing spatiotemporal features to train the spatiotemporal graph convolutional network model, wherein the training process uses MSE as a loss function, uses an Adam optimizer to adjust model parameters, and adjusts hyperparameter configuration according to the performance of the validation set to minimize the prediction error;

[0012] S60: Based on the trained spatiotemporal graph convolutional network model, the latest meteorological data of the target area and the historical ozone concentration data generated by the sliding window are input, and the future ozone concentration forecast value of the area is output.

[0013] Furthermore, in step S10, the historical ozone concentration data and the meteorological data are subjected to standardization preprocessing, the steps comprising:

[0014] S11: Acquire and organize historical monitoring data in a plurality of data files, wherein the data files are stored in a designated folder as data sources, wherein the historical ozone concentration data and the meteorological data contain relevant information in different time periods;

[0015] S12: Based on the data source, all data files are loaded, the required historical ozone concentration data and the meteorological data are extracted and cleaned, missing value filling, outlier detection and correction are performed on the loaded data, and the data is standardized and preprocessed so that the input data of different dimensions are within the same dimensional range;

[0016] S13: Merging the preprocessed data into a complete time series data set in a time-aligned manner, wherein the data set includes historical ozone concentration data and corresponding meteorological data, and ensuring the consistency of data timestamps for subsequent model training.

[0017] Further, in step S20, based on the geographical location of each monitoring point in the training sample and the correlation coefficient therebetween, a spatial graph structure is established, in which nodes represent monitoring points and edges represent spatial correlations between monitoring points. When the correlation coefficient between monitoring points exceeds a preset threshold, an edge connection is established. The steps include:

[0018] S21: determining the geographical scope of the target area according to the periodic variation characteristics of the historical ozone concentration, and extracting the ozone concentration data in the target area;

[0019] S22: Based on the extracted ozone concentration data in the target area, each latitude and longitude grid point in the target area is regarded as a node in a graph structure, wherein the node represents the ozone concentration value at a specific location and includes additional node features, including historical ozone concentration data and meteorological data at the corresponding location;

[0020] S23: Based on the historical concentration data and the meteorological data, calculating the correlation coefficient between the nodes, wherein the correlation coefficient adopts the Pearson correlation coefficient as a metric and is calculated based on the data in the historical time window corresponding to the node, and the length of the time window is determined according to the periodic variation characteristics of the ozone concentration;

[0021] S24: according to the correlation coefficient and a preset correlation coefficient threshold, when the correlation coefficient between two nodes exceeds the threshold, establishing an edge connection between the two nodes, wherein the threshold is selected through historical data analysis or cross-validation optimization to construct a graph structure reflecting spatial correlation;

[0022] S25: Based on the graph structure, a graph data object is created, wherein the graph data object includes a node feature matrix and an edge index, wherein the node feature matrix contains historical ozone concentration data and meteorological data features, and the edge index is used to represent the spatial connection relationship between nodes for subsequent graph neural network model training and prediction.

[0023] Furthermore, in step S30, GCN and GAT are used to perform convolution operations on the spatial graph structure to extract the spatial dependency between monitoring points and obtain time series data containing spatial features, wherein:

[0024] The computational model of the GCN includes:

[0025] H (l+1) =σ(AH (l) W (l) );

[0026] In the formula, σ is the activation function, H (l) is the node feature representation of the lth layer, A is the adjacency matrix of the graph, W (l) is the weight matrix of the lth layer;

[0027] The computational model of the GAT comprises:

[0028] H i (i+1) =ω(∑ j∈N (i)a ijH j( l) W (l) );

[0029] Where N(i) is the set of neighbor nodes of node i, a ij is the attention weight between node i and node j, H j (l) is the feature representation of node j in layer l.

[0030] Further, in step S40, based on the time series data containing spatial features, a TCN or RNN model is used to extract the time-dependent features in the time series data to obtain a comprehensive representation containing spatiotemporal features, wherein:

[0031] When the TCN model is used, the long-range dependencies in the time series are captured through multi-layer convolution operations, and the convolution operation of TCN is expressed as:

[0032] Y = Conv1D(X,K,d);

[0033] In the formula, X is the input feature matrix, K is the convolution kernel size, and d is the expansion coefficient;

[0034] Alternatively, when the RNN model is used, the time features are further extracted through LSTM or GRU, and the update model of the RNN is:

[0035] h t =RNN(X t ,h t-1 );

[0036] Where, X t is the node feature representation at time step t, h t is the hidden state of RNN, h t-1 is the hidden state at the previous time step t-1.

[0037] Furthermore, in step S50, the training sample and the comprehensive representation including the spatiotemporal features are used to train the spatiotemporal graph convolutional network model, and the training process includes:

[0038] S51: According to the deviation between the predicted value and the actual value, MSE is used as a loss function to measure the deviation between the predicted value and the actual value. The calculation model of MSE includes:

[0039]

[0040] In the formula, y i is the actual value, is the predicted value, N is the number of samples;

[0041] S52: According to the calculation result of the loss function, the model parameters are adjusted using the Adam optimizer, and the optimization process is accelerated by an adaptive learning rate to ensure that the model converges quickly during the training process;

[0042] S53: Based on the optimized model parameters and the training performance of the validation set, dynamically adjust the model's hyperparameter configuration, including learning rate, regularization parameter, network depth, etc., to minimize the prediction error and avoid overfitting or underfitting.

[0043] Further, in step S60, based on the trained spatiotemporal graph convolutional network model, the latest meteorological data of the target area and the historical ozone concentration data generated by the sliding window are input, and the future ozone concentration prediction value of the area is output, the steps include:

[0044] S61: Based on the trained spatiotemporal graph convolutional network model, the latest meteorological data of the target area and the historical ozone concentration data generated by the sliding window are used as input to construct an input feature matrix of the model;

[0045] S62: outputting a predicted value of ozone concentration in the target area in the future through calculation of the spatiotemporal graph convolutional network model according to the input data of the target area, for short-term and long-term air quality monitoring and early warning;

[0046] S63: Based on the predicted value of ozone concentration, relevant departments can take corresponding prevention and control measures in advance to reduce the harm of ozone pollution to public health and the ecological environment.

[0047] The present invention also includes an ozone concentration prediction device based on a spatiotemporal graph convolutional network model, using the above method, including:

[0048] A data acquisition unit, used to acquire historical ozone concentration data and corresponding meteorological data of a target area, perform standardized preprocessing on the historical ozone concentration data and the meteorological data, and construct training samples and prediction samples based on a sliding window method;

[0049] A spatial graph structure construction unit, used to establish a spatial graph structure based on the geographical locations of the monitoring points in the training samples and the correlation coefficients therebetween, wherein the nodes in the spatial graph structure represent the monitoring points and the edges represent the spatial correlations between the monitoring points, and when the correlation coefficients between the monitoring points exceed a preset threshold, an edge connection is established;

[0050] A spatial feature extraction unit, used to perform a convolution operation on the spatial graph structure using GCN and GAT, extract the spatial dependency relationship between monitoring points, and obtain time series data containing spatial features;

[0051] A time feature extraction unit, used to extract time-dependent features in the time series data based on the time series data containing spatial features by using a TCN or RNN model to obtain a comprehensive representation containing spatiotemporal features;

[0052] A model training unit, used to train the spatiotemporal graph convolutional network model using the training samples and the comprehensive representation containing spatiotemporal features, wherein the training process uses MSE as a loss function, uses an Adam optimizer to adjust model parameters, and adjusts hyperparameter configurations according to validation set performance to minimize prediction errors;

[0053] The prediction output unit is used to output the future ozone concentration prediction value of the area based on the trained spatiotemporal graph convolutional network model, the latest meteorological data of the target area and the historical ozone concentration data generated by the sliding window.

[0054] The present invention also includes a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, the method described above is implemented.

[0055] The present invention also includes a storage medium on which a computer program is stored. When the computer program is executed by a processor, the method described above is implemented.

[0056] The beneficial effects of the present invention are:

[0057] The time and space information are integrated into a unified framework, which fully combines the spatial feature extraction capabilities of graph neural networks (GCN, GAT) and the temporal feature extraction capabilities of models such as recurrent neural networks (LSTM, GRU) and temporal convolutional networks (TCN). This comprehensive method can better capture the spatiotemporal dependency of ozone concentration, thereby significantly improving the prediction accuracy. Traditional prediction methods are often limited to a single type of neural network structure, while the present invention achieves efficient processing of complex spatiotemporal data by integrating multiple network structures, such as GCN, GAT, LSTM, GRU and TCN. The benefit of integrating time and space information into one model is that it integrates feature learning, allowing the model to understand data features more comprehensively, thereby improving the representation ability of features. In addition, by comprehensively processing time and space information, it is helpful to capture the complex relationship between spatiotemporal structures, improve the representation ability and generalization performance of the model, and at the same time, this comprehensive method can also reduce information loss, improve the model's understanding of data relevance, and thus improve prediction accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0058] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0059] Figure 1 This is a flow chart of the ozone concentration prediction method based on the spatiotemporal graph convolutional network model;

[0060] Figure 2 This is the overall flow chart of the ozone concentration prediction method based on the spatiotemporal graph convolutional network model;

[0061] Figure 3 This is the architecture diagram of the spatiotemporal graph convolutional network based on RNN and GNN;

[0062] Figure 4 This is the architecture diagram of the spatiotemporal graph convolutional network based on TCN and GNN;

[0063] Figure 5 It is a schematic diagram of the structure of an ozone concentration prediction device based on a spatiotemporal graph convolutional network model;

[0064] Figure 6 A schematic diagram of the structure of a computer device. DETAILED DESCRIPTION

[0065] The technical solutions in the embodiments of the present invention will be described clearly and completely below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments.

[0066] like Figure 1 , 2 As shown: A method for predicting ozone concentration based on a spatiotemporal graph convolutional network model, comprising the following steps:

[0067] S10: Obtain historical ozone concentration data and corresponding meteorological data of the target area, perform standardized preprocessing on the historical ozone concentration data and meteorological data, and construct training samples and prediction samples based on the sliding window method;

[0068] In the data preprocessing stage, ozone concentration data files are loaded from the specified folder, and the xarray library is used to read and merge these data files. The longitude and latitude range of East China is selected to extract the ozone concentration data of the region. In order to better utilize the spatiotemporal characteristics of the data, we use the sliding window method to construct features and labels. Specifically, for each time point, we select the data of the previous N days as features and the ozone concentration at the current time point as a label. Through standardization, the data is converted into a form suitable for model training.

[0069] S20: Based on the geographical location of each monitoring point in the training sample and the correlation coefficient therebetween, a spatial graph structure is established, in which nodes represent monitoring points and edges represent spatial correlations between monitoring points. When the correlation coefficients between monitoring points exceed a preset threshold, an edge connection is established;

[0070] In the graph structure construction stage, we regard the longitude and latitude grid points in East China as nodes of the graph, and use the spatial correlation between nodes to construct the edges of the graph. Specifically, we calculate the correlation coefficient between the features of each node. If the correlation coefficient exceeds the set threshold, an edge is established between the two nodes. In this way, we get a graph structure that reflects spatial correlation. Assuming that the number of nodes in the graph is N, the set of edges is E, and the node feature matrix is ​​X, the graph can be expressed as G = (V, E), where V is the node set and E is the edge set.

[0071] S30: GCN (graph convolutional network) and GAT (graph attention network) are used to perform convolution operations on the spatial graph structure to extract the spatial dependency between monitoring points and obtain time series data containing spatial features;

[0072] S40: Based on the time series data containing spatial features, a TCN (temporal convolutional network) or RNN (recurrent neural network) model is used to extract the time-dependent features in the time series data to obtain a comprehensive representation containing spatiotemporal features;

[0073] S50: The spatiotemporal graph convolutional network model is trained using training samples and a comprehensive representation containing spatiotemporal features. The mean square error (MSE) is used as the loss function in the training process, the Adam optimizer is used to adjust the model parameters, and the hyperparameter configuration is adjusted according to the performance of the validation set to minimize the prediction error.

[0074] S60: Based on the trained spatiotemporal graph convolutional network model, the latest meteorological data of the target area and the historical ozone concentration data generated by the sliding window are input, and the future ozone concentration forecast value of the area is output.

[0075] A spatio-temporal graph convolutional network (STGCN) combining a graph neural network (GNN), a recurrent neural network (RNN) and a temporal convolutional network (TCN) is proposed. The graph neural network can effectively process data with a graph structure and capture the relationship between nodes through graph convolution operations to extract spatial features. Recurrent neural networks, especially long short-term memory networks (LSTM) and gated recurrent units (GRU), perform well in processing time series data and can effectively solve the problems of gradient vanishing and gradient exploding in long series data processing by introducing memory units and gating mechanisms. The temporal convolutional network can process time series data in parallel through convolution operations and has high computational efficiency.

[0076] This innovative ozone concentration prediction method effectively captures the spatiotemporal characteristics of ozone concentration changes by constructing a spatiotemporal graph convolutional network model, thereby improving prediction accuracy and generalization ability. Moreover, in this method, the spatial dimension and the temporal dimension are integrated into the same model, which brings multiple advantages: first, comprehensive feature learning enables the model to understand the data characteristics more comprehensively; second, information interaction helps capture the relationship between spatial structure and temporal dynamics, while reducing information loss and improving prediction accuracy. This comprehensive processing method simplifies the model architecture and improves generalization ability, making it more suitable for processing complex spatiotemporal data, such as meteorological and traffic data, and increasing the potential and possibility of spatiotemporal data analysis.

[0077] The model architecture of the invention processes spatial features including multiple layers of GCN (graph convolutional network) and GAT (graph attention network). Spatial features are extracted through these network layers. Each layer of the graph convolutional network aggregates node features with the features of its neighboring nodes to obtain a new node feature representation. Through multiple layers of graph convolution operations, deeper spatial features can be captured. In order to process time series features, the invention proposes two methods: one is to combine RNN (recurrent neural network), and the other is to combine TCN (temporal convolutional network). RNN includes two variants, LSTM and GRU. These networks can effectively capture long-range dependencies in time series data, while TCN can process time series data in parallel through multiple layers of dilated convolutions and residual connections, with high computational efficiency.

[0078] Finally, the model maps the extracted spatiotemporal features to the predicted ozone concentration through a linear layer. Through repeated model training and verification, the model parameters are continuously optimized to improve the accuracy and stability of the prediction.

[0079] The model proposed in the present invention integrates time and space information into a unified framework, fully combining the spatial feature extraction capabilities of graph neural networks (GCN, GAT) and the temporal feature extraction capabilities of models such as recurrent neural networks (LSTM, GRU) and TCN (temporal convolutional network). This comprehensive method can better capture the spatiotemporal dependency of ozone concentration, thereby significantly improving the prediction accuracy. Traditional prediction methods are often limited to a single type of neural network structure, while the present invention achieves efficient processing of complex spatiotemporal data by integrating multiple network structures, such as GCN, GAT, LSTM, GRU and TCN. The advantage of integrating time and space information into one model is that it integrates feature learning, enabling the model to understand data features more comprehensively, thereby improving the feature representation capability. In addition, by comprehensively processing time and space information, it is helpful to capture the complex relationship between spatiotemporal structures and enhance the representation capability and generalization performance of the model. At the same time, this comprehensive method can also reduce information loss, improve the model's understanding of data correlation, and thus improve prediction accuracy.

[0080] The method of the present invention is not only suitable for the prediction of ozone concentration, but also has wide applicability and can be extended to other environmental data prediction fields. Its flexibility and scalability make it a powerful tool for processing various complex spatiotemporal data. By making full use of graph structure and time series information, the method of the present invention can achieve efficient calculation on large-scale data sets and provide accurate prediction results, which can provide more accurate decision-making support for air pollution prevention and control. By accurately predicting the changing trend of ozone concentration, relevant departments can take corresponding prevention and control measures in advance to reduce the harm of ozone pollution to public health and the ecological environment. In addition, the method can also be extended to the prediction of other air pollutants, and has broad application prospects and important social value.

[0081] As a preferred embodiment of the above, in step S10, the historical ozone concentration data and meteorological data are subjected to standardized preprocessing, the steps comprising:

[0082] S11: Acquire and organize historical monitoring data in multiple data files, the data files are stored in a designated folder as data sources, wherein historical ozone concentration data and meteorological data contain relevant information in different time periods;

[0083] S12: Based on the data source, load all data files, extract and clean the required historical ozone concentration data and meteorological data, fill missing values, detect and correct outliers for the loaded data, and perform standardization preprocessing on the data so that the input data of different dimensions are within the same dimensional range;

[0084] S13: The preprocessed data are merged into a complete time series data set in a time-aligned manner. The data set contains historical ozone concentration data and corresponding meteorological data, and ensures the consistency of data timestamps for subsequent model training.

[0085] By loading and merging multiple historical monitoring data stored in different files into a complete data set, data missing or inconsistency problems caused by scattered files can be eliminated, improving data integrity and consistency.

[0086] Among them, in step S10, the window size of the sliding window method is dynamically adjusted according to the statistical characteristics of historical data. The selection of the window size is based on the periodicity and trend of the change in ozone concentration in the target area to improve the representativeness of the training samples. Specifically, the size of the sliding window is determined. For example, the data of the past 90 days are selected as a time window. For each time point, the data of the previous N days are selected as features, and the ozone concentration at the current time point is used as a label. The features include statistics such as the mean, standard deviation, minimum and maximum values ​​of the data in the window. These statistics can describe the changes in the data in the time window and provide rich feature information for the model.

[0087] In this embodiment, in step S20, based on the geographical location of each monitoring point in the training sample and the correlation coefficient therebetween, a spatial graph structure is established, in which nodes represent monitoring points and edges represent the spatial correlation between monitoring points. When the correlation coefficient between the monitoring points exceeds a preset threshold, an edge connection is established. The steps include:

[0088] S21: According to the periodic variation characteristics of historical ozone concentration, the geographical scope of the target area is determined, and the ozone concentration data in the target area is extracted; the target area is a geographical area with a set longitude and latitude range, and the geographical scope of the target area (such as East China) is determined, that is, the longitude and latitude range of the area, and the ozone concentration data in the area is extracted from the loaded data set to focus on data analysis in a specific area, avoid processing unnecessary global data, and improve calculation efficiency;

[0089] S22: Based on the extracted ozone concentration data in the target area, each latitude and longitude grid point in the target area is regarded as a node in the graph structure. The node represents the ozone concentration value at a specific location and includes additional node features, including historical ozone concentration data and meteorological data at the corresponding location; such as meteorological data (temperature, humidity, wind speed, etc.); the extracted features are standardized so that their mean is 0 and their standard deviation is 1. The standardization process can eliminate the dimensional differences between different features, prevent certain features from excessively affecting the results during model training, and improve the stability of the model and the prediction accuracy;

[0090] Each longitude and latitude grid point in the target area is regarded as a node of the graph. Each node represents the ozone concentration value at a specific location. This method structures the spatial data into graph data, which is convenient for processing using graph neural networks.

[0091] S23: Based on the historical concentration data and meteorological data, the correlation coefficient between each node is calculated. The correlation coefficient uses the Pearson correlation coefficient as a metric and is calculated based on the data in the historical time window corresponding to the node. The length of the time window is determined according to the periodic variation characteristics of the ozone concentration.

[0092] S24: According to the correlation coefficient and the preset correlation coefficient threshold, when the correlation coefficient between two nodes exceeds the threshold, an edge connection is established between the two nodes, and the threshold is selected through historical data analysis or cross-validation optimization to construct a graph structure reflecting spatial correlation; specifically, the correlation coefficient between each node feature is calculated, and according to the set threshold, if the correlation coefficient between the two nodes exceeds the threshold, an edge is established between the two nodes to form an adjacency matrix of the graph, which can construct a graph structure reflecting spatial correlation and ensure that nodes in similar positions are connected by edges.

[0093] S25: Based on the graph structure, a graph data object is created, where the graph data object includes a node feature matrix and an edge index, wherein the node feature matrix contains historical ozone concentration data and meteorological data features, and the edge index is used to represent the spatial connection relationship between nodes for subsequent graph neural network model training and prediction.

[0094] By defining the longitude and latitude grids in the target area as nodes and calculating the correlation between monitoring points based on the Pearson correlation coefficient, the spatial dependency between monitoring points can be accurately captured, providing a more accurate spatial feature representation for subsequent model training; using historical data analysis or cross-validation methods to optimize the correlation coefficient threshold can ensure that the threshold adapts to the characteristics of different data sets and regions, and improves the graph structure's ability to characterize spatial correlation; according to the periodic variation characteristics of ozone concentration, the time window length is reasonably set to ensure that the calculation of the correlation coefficient is more scientific and can reflect the periodic dynamics of ozone concentration when capturing spatial features; by constructing a complete graph data object including a node feature matrix and edge index, high-quality spatial features are input into the graph neural network model, providing a solid foundation for subsequent spatiotemporal feature extraction and prediction; by setting the correlation coefficient threshold, node connections with low spatial correlation are filtered out, significantly reducing the computational complexity while retaining key spatial dependency information.

[0095] In step S30, GCN (graph convolutional network) and GAT (graph attention network) are used to perform convolution operation on the spatial graph structure to extract the spatial dependency relationship between monitoring points and obtain time series data containing spatial features, where:

[0096] The computational model of GCN (graph convolutional network) includes:

[0097] H (l+1) =σ(AH (l) W (l) );

[0098] In the formula, σ is the activation function, H (l) is the node feature representation of the lth layer, A is the adjacency matrix of the graph, W (l) is the weight matrix of the lth layer;

[0099] The computational model of GAT (Graph Attention Network) includes:

[0100] H i (i+1) =ω(∑ j∈N(i) a ij H j (l) W (l) );

[0101] Where N(i) is the set of neighbor nodes of node i, a ij is the attention weight between node i and node j, H j (l) is the feature representation of node j in layer l.

[0102] A multi-layer graph convolutional network and graph attention network layer are constructed to extract the spatial features of the nodes. The graph convolutional network performs convolution operations on the node features through the adjacency matrix and weight matrix to capture the relationship between the node and its neighbor nodes. The graph attention network further introduces the attention mechanism to assign different weights to each neighbor node to enhance the representation ability of the node features.

[0103] By utilizing the efficient neighbor aggregation characteristics of GCN and the adaptive attention mechanism of GAT, the spatial dependencies between monitoring points can be accurately captured, and the feature extraction and prediction performance of the model can be improved. Combining the advantages of both, not only the computational efficiency of the model is improved, but also the adaptability and interpretability to heterogeneous spatial data are enhanced, providing reliable technical support for high-precision ozone concentration prediction.

[0104] In the model building stage, we proposed two spatiotemporal models combined with graph neural networks, namely the STGCN model of TCN+GNN and the STGCN model of RNN+GNN.

[0105] As a preferred embodiment of the above embodiment, in step S40, based on the time series data containing spatial features, a TCN (temporal convolutional network) or RNN (recurrent neural network) model is used to extract the time-dependent features in the time series data to obtain a comprehensive representation containing spatiotemporal features, wherein:

[0106] When using the TCN (temporal convolutional network) model, long-range dependencies in time series are captured through multi-layer convolution operations, such as Figure 4 As shown in the figure, the spatiotemporal graph convolutional network architecture diagram based on TCN and GNN, the convolution operation of TCN is expressed as:

[0107] Y = Conv1D(X,K,d);

[0108] In the formula, X is the input feature matrix, K is the convolution kernel size, and d is the expansion coefficient;

[0109] Alternatively, when the RNN (recurrent neural network) model is used, the temporal features are further extracted through LSTM (long short-term memory network) or GRU (gated recurrent unit), such as Figure 3 As shown in the figure, the spatiotemporal graph convolutional network architecture diagram based on RNN and GNN, the update model of RNN is:

[0110] h t =RNN(X t ,h t-1 );

[0111] In the formula, X t is the node feature representation at time step t, h t is the hidden state of RNN, h t-1is the hidden state at the previous time step t-1.

[0112] By adopting TCN or RNN models to extract time-dependent features, a high-quality comprehensive representation including spatiotemporal features is provided for subsequent predictions.

[0113] The hidden state is mapped to the predicted ozone concentration value through a linear layer to form a comprehensive representation containing spatiotemporal features, converting the high-dimensional feature space into the target prediction value, ensuring that the model output can be directly used in practical applications.

[0114] TCN expands the receptive field by dilated convolution and can efficiently capture long-range dependencies in time series, while RNN effectively retains historical information through memory mechanisms (such as LSTM and GRU) to accurately extract complex time-dependent features. It provides two options, TCN and RNN, and flexibly selects the time feature extraction method according to the characteristics of the task (such as computational efficiency priority or long-range dependency priority), adapts to the needs of different scenarios, and improves the versatility of the method.

[0115] Two methods are used: recurrent neural network (RNN) and temporal convolutional network (TCN). The recurrent neural network processes time series data through hidden states and is suitable for capturing long-range time dependencies. The temporal convolutional network processes time features through multiple layers of dilated convolutions and has parallel computing capabilities and high efficiency. Through the operations of these layers, dynamic features in time series are extracted.

[0116] As a preferred embodiment of the above, in step S50, in the model training and verification stage, we divide the data into a training set and a test set, and use the training samples and the comprehensive representation containing the spatiotemporal features to train the spatiotemporal graph convolutional network model. The training process includes:

[0117] S51: According to the deviation between the predicted value and the actual value, MSE (mean square error) is used as the loss function to measure the deviation between the predicted value and the actual value. The calculation model of MSE includes:

[0118]

[0119] In the formula, y i is the actual value, is the predicted value, N is the number of samples;

[0120] S52: According to the calculation results of the loss function, the Adam optimizer is used to adjust the model parameters, and the optimization process is accelerated by adaptive learning rate to ensure that the model converges quickly during training;

[0121] S53: Based on the optimized model parameters and the training performance of the validation set, dynamically adjust the model's hyperparameter configuration, including learning rate, regularization parameters, and network depth, to minimize the prediction error and avoid overfitting or underfitting.

[0122] By introducing the Adam optimizer and using adaptive learning rate to dynamically adjust the parameter update amplitude, the convergence speed of the model is significantly accelerated and the training efficiency is improved; using MSE as the loss function can effectively quantify the error between the predicted value and the actual value, providing an accurate target direction for parameter optimization, thereby improving the prediction accuracy of the model; dynamically adjusting the hyperparameter configuration (including learning rate, regularization parameter, network depth, etc.) through the performance of the validation set, effectively reducing overfitting and underfitting phenomena and improving the generalization ability of the model.

[0123] By adjusting the learning rate and the number of training rounds, we can find the optimal model parameters. During the training process, we adjust the model's hyperparameters based on the performance of the validation set to prevent overfitting.

[0124] Through the above model training, we can obtain a spatiotemporal graph convolutional network model that can accurately predict ozone concentration. In practical applications, we input the latest ozone concentration data, and the model can predict future ozone concentrations based on historical data and spatial correlation. It can not only be used for ozone concentration prediction, but also can be extended to the prediction of other atmospheric pollutants, such as PM2.5 concentration prediction, sulfur dioxide concentration prediction, etc.

[0125] The data set is divided into a training set and a test set according to a certain ratio (90% training set, 10% test set).

[0126] The model is trained iteratively. Each iterative step includes forward propagation, loss calculation, back propagation, and weight update. Through repeated iterations, the model parameters are gradually optimized, and the loss value is continuously reduced. The hyperparameters of the model are adjusted based on the performance of the validation set to prevent overfitting.

[0127] The prediction performance of the model was evaluated on the test set, and indicators such as mean square error (MSE), root mean square error (RMSE) and mean absolute error (MAE) were calculated to comprehensively evaluate the accuracy and stability of the model.

[0128] The latest ozone concentration data is input, and the model predicts future ozone concentrations based on historical data and spatial correlation. The prediction results can be used for short-term and long-term air quality monitoring and early warning.

[0129] In this embodiment, in step S60, based on the trained spatiotemporal graph convolutional network model, the latest meteorological data of the target area and the historical ozone concentration data generated by the sliding window are input, and the future ozone concentration prediction value of the area is output. The steps include:

[0130] S61: Based on the trained spatiotemporal graph convolutional network model, the latest meteorological data of the target area and the historical ozone concentration data generated by the sliding window are used as input to construct the input feature matrix of the model;

[0131] S62: Based on the input data of the target area, the future ozone concentration forecast value of the target area is output through the calculation of the spatiotemporal graph convolutional network model, which is used for short-term and long-term air quality monitoring and early warning;

[0132] S63: Based on the predicted ozone concentration, relevant departments can take corresponding prevention and control measures in advance, including but not limited to pollution source control, public health protection and emergency plan implementation, to reduce the harm of ozone pollution to public health and the ecological environment. The prediction results can be used to formulate environmental protection policies and emergency plans to improve the scientificity and effectiveness of environmental management.

[0133] By using the trained spatiotemporal graph convolutional network model, taking the latest meteorological data and historical ozone concentration data as input and comprehensively considering the temporal and spatial characteristics, the accuracy and reliability of the prediction results can be significantly improved; by outputting future ozone concentration prediction values, data support is provided for short-term pollution event warning and long-term air quality management, facilitating relevant departments to formulate scientific environmental governance strategies; the predicted values ​​are directly applied to monitoring, early warning and prevention and control measures, which significantly improves the practicality of the technical solution and provides an operational tool for environmental management departments.

[0134] The present invention also includes an ozone concentration prediction device based on a spatiotemporal graph convolutional network model, using the above method, such as Figure 5 As shown, including:

[0135] A data acquisition unit, used to acquire historical ozone concentration data and corresponding meteorological data of the target area, perform standardized preprocessing on the historical ozone concentration data and meteorological data, and construct training samples and prediction samples based on a sliding window method;

[0136] A spatial graph structure building unit is used to build a spatial graph structure based on the geographical location of each monitoring point in the training sample and the correlation coefficient between them. The nodes in the spatial graph structure represent the monitoring points, and the edges represent the spatial correlation between the monitoring points. When the correlation coefficient between the monitoring points exceeds a preset threshold, an edge connection is established.

[0137] The spatial feature extraction unit is used to perform convolution operations on the spatial graph structure using GCN (graph convolutional network) and GAT (graph attention network), extract the spatial dependency between monitoring points, and obtain time series data containing spatial features;

[0138] A time feature extraction unit is used to extract time-dependent features in the time series data based on the time series data containing spatial features, using a TCN (temporal convolutional network) or RNN (recurrent neural network) model to obtain a comprehensive representation containing temporal and spatial features;

[0139] The model training unit is used to train the spatiotemporal graph convolutional network model using training samples and a comprehensive representation containing spatiotemporal features. The training process uses MSE (mean square error) as the loss function, uses the Adam optimizer to adjust the model parameters, and adjusts the hyperparameter configuration according to the performance of the validation set to minimize the prediction error;

[0140] The prediction output unit is used to output the future ozone concentration prediction value of the region based on the trained spatiotemporal graph convolutional network model, the latest meteorological data of the target area and the historical ozone concentration data generated by the sliding window.

[0141] See also Figure 6 A computer device 400 provided in an embodiment of the present application includes: a processor 410 and a memory 420, wherein the memory 420 stores a computer program executable by the processor 410, and when the computer program is executed by the processor 410, the above method is executed.

[0142] The embodiment of the present application further provides a storage medium 430 on which a computer program is stored. When the computer program is run by the processor 410, the above method is executed.

[0143] Among them, the storage medium 430 can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable red-only memory (PROM), read-only memory (ROM), magnetic storage, flash memory, disk or optical disk.

[0144] In the description of the present invention, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined as "first" and "second" may explicitly or implicitly include one or more of the features. "Multiple" means two or more, unless otherwise clearly and specifically defined.

[0145] In the present invention, unless otherwise clearly specified and limited, the terms "installed", "connected", "connected", "fixed" and the like should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection or an indirect connection through an intermediate medium, it can be the internal connection of two elements or the interaction relationship between two elements. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0146] In the description of this specification, the description with reference to the terms "one embodiment", "some embodiments", "example", "specific example", or "some examples" etc. means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or more embodiments or examples in a suitable manner. In addition, those skilled in the art may combine and combine the different embodiments or examples described in this specification and the features of the different embodiments or examples, unless they are contradictory.

[0147] Any process or method description in a flowchart or otherwise described herein may be understood to represent a module, segment or portion of code that includes one or more executable instructions for implementing the steps of a specific logical function or process, and the scope of the preferred embodiments of the present invention includes alternative implementations in which functions may not be performed in the order shown or discussed, including performing functions in a substantially simultaneous manner or in the reverse order depending on the functions involved, which should be understood by those skilled in the art to which the embodiments of the present invention belong.

[0148] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as an ordered list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by an instruction execution system, device or apparatus (such as a computer-based system, a system including a processor, or other system that can fetch instructions from an instruction execution system, device or apparatus and execute instructions), or in combination with these instruction execution systems, devices or apparatuses. For the purposes of this specification, a "computer-readable medium" can be any device that can contain, store, communicate, propagate or transmit a program for use by an instruction execution system, device or apparatus, or in combination with these instruction execution systems, devices or apparatuses. More specific examples of computer-readable media (a non-exhaustive list) include the following: an electrical connection with one or more wires (electronic devices), a portable computer disk box (magnetic device), a random access memory (RAM), a read-only memory (ROM), an erasable and programmable read-only memory (EPROM or flash memory), a fiber optic device, and a portable compact disk read-only memory (CDROM). In addition, the computer-readable medium may even be paper or other suitable medium on which the program is printed, since the program may be obtained electronically, for example, by optically scanning the paper or other medium and then editing, interpreting or processing in other suitable ways if necessary, and then stored in a computer memory.

[0149] It should be understood that the various parts of the present invention can be implemented by hardware, software, firmware or a combination thereof. In the above-mentioned embodiments, a plurality of steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented by hardware, as in another embodiment, it can be implemented by any one of the following technologies known in the art or their combination: a discrete logic circuit having a logic gate circuit for implementing a logic function for a data signal, a dedicated integrated circuit having a suitable combination of logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc.

[0150] A person skilled in the art may understand that all or part of the steps in the method for implementing the above-mentioned embodiment may be completed by instructing related hardware through a program, and the program may be stored in a computer-readable storage medium, which, when executed, includes one or a combination of the steps of the method embodiment.

[0151] The storage medium mentioned above may be a read-only memory, a magnetic disk or an optical disk, etc. Although the embodiments of the present invention have been shown and described above, it can be understood that the above embodiments are exemplary and cannot be understood as limiting the present invention. A person of ordinary skill in the art may change, modify, replace and modify the above embodiments within the scope of the present invention.

Claims

1. The ozone concentration prediction method based on the spatiotemporal graph convolutional network model is characterized by: The steps include: S10: Obtain historical ozone concentration data and corresponding meteorological data of the target area, perform standardization preprocessing on the historical ozone concentration data and the meteorological data, and construct training samples and prediction samples based on a sliding window method; S20: Based on the geographical locations of the monitoring points in the training samples and the correlation coefficients therebetween, a spatial graph structure is established, wherein nodes in the spatial graph structure represent monitoring points, and edges represent spatial correlations between monitoring points, and when the correlation coefficients between monitoring points exceed a preset threshold, an edge connection is established; S30: performing a convolution operation on the spatial graph structure using GAT to extract the spatial dependency relationship between monitoring points, and obtaining time series data containing spatial features; S40: Based on the time series data including spatial features, using a TCN or RNN model, extracting time-dependent features in the time series data to obtain a comprehensive representation including spatiotemporal features; S50: using the training sample and the comprehensive representation containing spatiotemporal features to train the spatiotemporal graph convolutional network model, wherein the training process uses MSE as a loss function, uses an Adam optimizer to adjust model parameters, and adjusts hyperparameter configuration according to the performance of the validation set to minimize the prediction error; S60: Based on the trained spatiotemporal graph convolutional network model, the latest meteorological data of the target area and the historical ozone concentration data generated by the sliding window are input, and the future ozone concentration forecast value of the area is output.

2. The ozone concentration prediction method based on the spatiotemporal graph convolutional network model according to claim 1 is characterized in that: In step S10, the historical ozone concentration data and the meteorological data are subjected to standardization preprocessing, the steps comprising: S11: Acquire and organize historical monitoring data in a plurality of data files, wherein the data files are stored in a designated folder as data sources, wherein the historical ozone concentration data and the meteorological data contain relevant information in different time periods; S12: Based on the data source, all data files are loaded, the required historical ozone concentration data and the meteorological data are extracted and cleaned, missing value filling, outlier detection and correction are performed on the loaded data, and the data is standardized and preprocessed so that the input data of different dimensions are within the same dimensional range; S13: Merging the preprocessed data into a complete time series data set in a time-aligned manner, wherein the data set includes historical ozone concentration data and corresponding meteorological data, and ensuring the consistency of data timestamps for subsequent model training.

3. The ozone concentration prediction method based on the spatiotemporal graph convolutional network model according to claim 1 is characterized in that: In step S20, based on the geographical location of each monitoring point in the training sample and the correlation coefficient therebetween, a spatial graph structure is established, in which nodes represent monitoring points and edges represent the spatial correlation between monitoring points. When the correlation coefficient between monitoring points exceeds a preset threshold, an edge connection is established. The steps include: S21: determining the geographical scope of the target area according to the periodic variation characteristics of the historical ozone concentration, and extracting the ozone concentration data in the target area; S22: Based on the extracted ozone concentration data in the target area, each latitude and longitude grid point in the target area is regarded as a node in a graph structure, wherein the node represents the ozone concentration value at a specific location and includes additional node features, including historical ozone concentration data and meteorological data at the corresponding location; S23: Based on the historical concentration data and the meteorological data, calculating the correlation coefficient between the nodes, wherein the correlation coefficient adopts the Pearson correlation coefficient as a metric and is calculated based on the data in the historical time window corresponding to the node, and the length of the time window is determined according to the periodic variation characteristics of the ozone concentration; S24: according to the correlation coefficient and a preset correlation coefficient threshold, when the correlation coefficient between two nodes exceeds the threshold, establishing an edge connection between the two nodes, wherein the threshold is selected through historical data analysis or cross-validation optimization to construct a graph structure reflecting spatial correlation; S25: Based on the graph structure, a graph data object is created, wherein the graph data object includes a node feature matrix and an edge index, wherein the node feature matrix contains historical ozone concentration data and meteorological data features, and the edge index is used to represent the spatial connection relationship between nodes for subsequent graph neural network model training and prediction.

4. The ozone concentration prediction method based on the spatiotemporal graph convolutional network model according to claim 1 is characterized in that: In step S30, GCN and GAT are used to perform convolution operations on the spatial graph structure to extract the spatial dependency between monitoring points and obtain time series data containing spatial features, where: The computational model of the GCN includes: H (l+1) =σ(AH (l) W (l) ); In the formula, σ is the activation function, H( l ) is the node feature representation of the lth layer, A is the adjacency matrix of the graph, W( l ) is the weight matrix of the lth layer; The computational model of the GAT comprises: H i (l+1) =σ(∑ j∈N(i) a ij H j (l) W (l) ); Where N(i) is the set of neighbor nodes of node i, a ij is the attention weight between node i and node j, H j (l) is the feature representation of node j in layer l.

5. The ozone concentration prediction method based on the spatiotemporal graph convolutional network model according to claim 1 is characterized in that: In step S40, based on the time series data containing spatial features, a TCN or RNN model is used to extract the time-dependent features in the time series data to obtain a comprehensive representation containing spatiotemporal features, wherein: When the TCN model is used, the long-range dependencies in the time series are captured through multi-layer convolution operations, and the convolution operation of TCN is expressed as: Y = Conv1D(X,K,d); In the formula, X is the input feature matrix, K is the convolution kernel size, and d is the expansion coefficient; Alternatively, when the RNN model is used, the time features are further extracted through LSTM or GRU, and the update model of the RNN is: h t =RNN(X t ,h t-1 ); In the formula, X t is the node feature representation at time step t, h t is the hidden state of RNN, h t-1 is the hidden state at the previous time step t-1.

6. The ozone concentration prediction method based on the spatiotemporal graph convolutional network model according to claim 1 is characterized in that: In step S50, the spatiotemporal graph convolutional network model is trained using the training sample and the comprehensive representation including spatiotemporal features, and the training process includes: S51: According to the deviation between the predicted value and the actual value, MSE is used as a loss function to measure the deviation between the predicted value and the actual value. The calculation model of MSE includes: In the formula, y i is the actual value, is the predicted value, N is the number of samples; S52: According to the calculation result of the loss function, the model parameters are adjusted using the Adam optimizer, and the optimization process is accelerated by an adaptive learning rate to ensure that the model converges quickly during the training process; S53: Based on the optimized model parameters and the training performance of the validation set, dynamically adjust the model's hyperparameter configuration, including learning rate, regularization parameter, network depth, etc., to minimize the prediction error and avoid overfitting or underfitting.

7. The ozone concentration prediction method based on the spatiotemporal graph convolutional network model according to claim 1 is characterized in that: In step S60, based on the trained spatiotemporal graph convolutional network model, the latest meteorological data of the target area and the historical ozone concentration data generated by the sliding window are input, and the future ozone concentration forecast value of the area is output. The steps include: S61: Based on the trained spatiotemporal graph convolutional network model, the latest meteorological data of the target area and the historical ozone concentration data generated by the sliding window are used as input to construct an input feature matrix of the model; S62: outputting a predicted value of ozone concentration in the target area in the future through calculation of the spatiotemporal graph convolutional network model according to the input data of the target area, for short-term and long-term air quality monitoring and early warning; S63: Based on the predicted value of ozone concentration, relevant departments can take corresponding prevention and control measures in advance to reduce the harm of ozone pollution to public health and the ecological environment.

8. An ozone concentration prediction device based on a spatiotemporal graph convolutional network model, characterized in that: Use of the method according to any one of claims 1 to 7, comprising: A data acquisition unit, used to acquire historical ozone concentration data and corresponding meteorological data of a target area, perform standardized preprocessing on the historical ozone concentration data and the meteorological data, and construct training samples and prediction samples based on a sliding window method; A spatial graph structure construction unit, used to establish a spatial graph structure based on the geographical locations of the monitoring points in the training samples and the correlation coefficients therebetween, wherein the nodes in the spatial graph structure represent the monitoring points and the edges represent the spatial correlations between the monitoring points, and when the correlation coefficients between the monitoring points exceed a preset threshold, an edge connection is established; A spatial feature extraction unit, used to perform a convolution operation on the spatial graph structure using GCN and GAT, extract the spatial dependency relationship between monitoring points, and obtain time series data containing spatial features; A time feature extraction unit, used to extract time-dependent features in the time series data based on the time series data containing spatial features by using a TCN or RNN model to obtain a comprehensive representation containing spatiotemporal features; A model training unit, used to train the spatiotemporal graph convolutional network model using the training samples and the comprehensive representation containing spatiotemporal features, wherein the training process uses MSE as a loss function, uses an Adam optimizer to adjust model parameters, and adjusts hyperparameter configurations according to validation set performance to minimize prediction errors; The prediction output unit is used to output the future ozone concentration prediction value of the area based on the trained spatiotemporal graph convolutional network model, the latest meteorological data of the target area and the historical ozone concentration data generated by the sliding window.

9. A computer device comprising a memory, a processor and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the method according to any one of claims 1 to 7 is implemented.

10. A storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method according to any one of claims 1 to 7 is implemented.

Images