CN113705880B - Traffic speed prediction method and device based on space-time attention force diagram convolution network - Google Patents

Traffic speed prediction method and device based on space-time attention force diagram convolution network Download PDF

Info

Publication number
CN113705880B
CN113705880B CN202110980592.8A CN202110980592A CN113705880B CN 113705880 B CN113705880 B CN 113705880B CN 202110980592 A CN202110980592 A CN 202110980592A CN 113705880 B CN113705880 B CN 113705880B
Authority
CN
China
Prior art keywords
matrix
time
network
space
attention
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Active
Application number
CN202110980592.8A
Other languages
Chinese (zh)
Other versions
CN113705880A (en
Inventor
夏莹杰
田瑞
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Hangzhou Yuantiao Science And Technology Co ltd
Original Assignee
Hangzhou Yuantiao Science And Technology Co ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Hangzhou Yuantiao Science And Technology Co ltd filed Critical Hangzhou Yuantiao Science And Technology Co ltd
Priority to CN202110980592.8A priority Critical patent/CN113705880B/en
Publication of CN113705880A publication Critical patent/CN113705880A/en
Application granted granted Critical
Publication of CN113705880B publication Critical patent/CN113705880B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
    • G06F17/10—Complex mathematical operations
    • G06F17/16—Matrix or vector computation, e.g. matrix-matrix or matrix-vector multiplication, matrix factorization
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/045—Combinations of networks
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
    • G06Q50/10—Services
    • G06Q50/26—Government or public services
    • G—PHYSICS
    • G08—SIGNALLING
    • G08G—TRAFFIC CONTROL SYSTEMS
    • G08G1/00—Traffic control systems for road vehicles
    • G08G1/01—Detecting movement of traffic to be counted or controlled
    • G08G1/0104—Measuring and analyzing of parameters relative to traffic conditions
    • G08G1/0125—Traffic data processing
    • G08G1/0129—Traffic data processing for creating historical data or processing based on historical data
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02T—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO TRANSPORTATION
    • Y02T10/00—Road transport of goods or passengers
    • Y02T10/10—Internal combustion engine [ICE] based vehicles
    • Y02T10/40—Engine management systems

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Business, Economics & Management (AREA)
  • Mathematical Physics (AREA)
  • Data Mining & Analysis (AREA)
  • Human Resources & Organizations (AREA)
  • Tourism & Hospitality (AREA)
  • Economics (AREA)
  • General Health & Medical Sciences (AREA)
  • Software Systems (AREA)
  • Computing Systems (AREA)
  • General Engineering & Computer Science (AREA)
  • Strategic Management (AREA)
  • Health & Medical Sciences (AREA)
  • Mathematical Optimization (AREA)
  • Molecular Biology (AREA)
  • Computational Mathematics (AREA)
  • Mathematical Analysis (AREA)
  • Evolutionary Computation (AREA)
  • Pure & Applied Mathematics (AREA)
  • General Business, Economics & Management (AREA)
  • Marketing (AREA)
  • Computational Linguistics (AREA)
  • Development Economics (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Biophysics (AREA)
  • Biomedical Technology (AREA)
  • Artificial Intelligence (AREA)
  • Databases & Information Systems (AREA)
  • Primary Health Care (AREA)
  • Educational Administration (AREA)
  • Algebra (AREA)
  • Game Theory and Decision Science (AREA)
  • Entrepreneurship & Innovation (AREA)
  • Operations Research (AREA)
  • Quality & Reliability (AREA)
  • Chemical & Material Sciences (AREA)
  • Analytical Chemistry (AREA)

Abstract

The invention discloses a traffic speed prediction method and a traffic speed prediction device based on a space-time attention-seeking graph convolutional network, wherein the traffic speed prediction method comprises the following steps: sampling the acquired existing speed data set to obtain a component data matrix related to the time sequence; constructing a time attention network; inputting the component data matrix into a time attention network to obtain a time correlation matrix; constructing a spatial attention network; inputting the time correlation matrix into a spatial attention network, and fusing to obtain a space-time correlation matrix; inputting the road network topology matrix into a graph rolling network, combining the road network topology matrix with the space-time correlation matrix, and performing model training; and carrying out speed prediction through the trained model. The time attention network and the space attention network are fused to form a space-time attention network, the relevance of the traffic network in time and space is extracted, meanwhile, the information of adjacent nodes is continuously fused by combining the graph convolution network, so that the traffic speed is predicted, and the accuracy of the traffic speed prediction is improved.

Description

Traffic speed prediction method and device based on space-time attention force diagram convolution network
Technical Field
The invention relates to the field of intelligent traffic and deep learning, in particular to a traffic speed prediction method and a related device based on a space-time attention-seeking convolutional network.
Background
With the continuous improvement of income per capita, urban traffic becomes more and more crowded, traffic prediction becomes an increasingly important research topic, and the accurate and timely prediction of traffic conditions not only provides an effective treatment means for management of traffic management departments, but also provides reasonable planning for travelers. Traffic prediction has important significance for urban planning, traffic management and property safety. However, traffic conditions have been a challenging task in space-time prediction due to their complexity both in time and space.
Early traffic speeds were based primarily on some statistical approach or simple machine learning approach. Representative of these are autoregressive integrated moving average (ARIMA), vector Autoregressive (VAR), K Nearest Neighbor (KNN) and Support Vector Regression (SVR). Although these methods can predict traffic speed, as the spatial complexity of traffic networks increases, and the correlation over time increases, the accuracy of the predictions is lower. Conventional traffic speed predictions are difficult to mine to temporal and spatial correlations, greatly limiting performance in space-time mining.
Therefore, how to design a speed prediction method comprehensively considering space-time correlation under traffic conditions is a technical problem to be solved by those skilled in the art.
Disclosure of Invention
In view of the foregoing, it is an object of the present invention to provide a traffic speed prediction method and related apparatus based on a spatiotemporal attention-seeking convolutional network, so that the prediction of traffic speed is more accurate.
The first aspect of the invention provides a traffic speed prediction method based on a space-time attention-seeking convolutional network, which comprises the following steps:
Sampling the acquired existing speed data set to obtain a component data matrix related to the time sequence;
constructing a time attention network;
inputting the component data matrix into the time attention network to obtain a time correlation matrix;
Constructing a spatial attention network;
inputting the time correlation matrix into the spatial attention network, and fusing to obtain a space-time correlation matrix;
inputting a road network topology matrix into a graph rolling network, combining the road network topology matrix with the space-time correlation matrix, and performing model training;
and carrying out speed prediction through the trained model.
In this solution, the inputting the component data matrix into the time attention network, to obtain the time correlation matrix specifically includes:
Inputting the component data matrix into the time attention network to obtain a time attention matrix T:
Where K e、L1、L2、L3、Ve is a learnable parameter, σ represents a sigmoid activation function, For the matrix of the component data,C r-1 represents the channel number of the input data of the r layer, and T r-1 represents the length of the time dimension of the input data of the r layer;
after normalization, the time attention matrix T captures the association strength between nodes according to the time attention moment matrix:
t i,j reflects the time correlation strength between the times i, j, and multiplies the time correlation strength by the component data matrix to obtain a time correlation matrix
In this solution, the inputting the time correlation matrix into the spatial attention network, and the fusing to obtain the space-time correlation matrix specifically includes:
inputting the time correlation matrix into the spatial attention network to obtain a space-time attention matrix S:
wherein K s、H1、H2、H3、Vs is a learnable parameter;
And calculating a time correlation matrix S' according to the space-time attention matrix S:
In this scheme, the inputting the road network topology matrix into the graph convolution network, and combining with the space-time correlation matrix, performing model training, and obtaining the prediction speed specifically includes:
Inputting the road network topology matrix into a Laplace matrix L:
wherein A represents an input road network topology matrix, D represents a degree matrix, specifically a diagonal matrix, and the elements of the diagonal are A ij represents an element of row i and column j;
The graph rolling network is specifically in the form of chebyshev polynomials, and is expressed as follows:
where G represents the convolution operation of a graph, Is the scaled normalized laplace matrix, λ max is the maximum eigenvalue of L, θ' k (k=0, 1, … K) is the coefficient of the kth term of chebyshev polynomial, which is a learnable parameter;
The definition of the chebyshev polynomial of the K-th order is as follows:
Tk(x)＝2xTk-1(x)-Tk-2(x)；
combining the input road network topology matrix and the space-time correlation matrix to obtain a final graph convolution formula:
And after model training is carried out on the graph convolution formula, speed prediction is carried out through the trained model.
In this scheme, the sampling the existing velocity data set to obtain the component data matrix related to the time sequence specifically includes:
Sampling the existing speed data set in three dimensions to respectively obtain a time division component data matrix, a day component data matrix and a week component data matrix of a related time sequence;
the road network topology matrix is input into a graph convolution network and combined with the space-time correlation matrix, and model training is specifically carried out as follows:
The road network topology matrix is input into a graph rolling network, is combined with the space-time correlation matrix of three dimensions respectively, and adopts the component data matrix of three dimensions to perform model training respectively, and performs fusion output.
In this solution, before the speed prediction is performed by the trained model, the method further includes:
Constructed loss function
Wherein Y t represents the actual traffic speed,Indicating the prediction speed, L reg is the super-parameter for reducing the prediction error in order to avoid over-fitting the parameters.
The second aspect of the present invention also provides a traffic speed prediction system based on a spatiotemporal strive-to-force convolution network, comprising a memory and a processor, wherein the memory comprises a traffic speed prediction method program based on a spatiotemporal strive-to-force convolution network, and the traffic speed prediction method program based on a spatiotemporal strive-to-force convolution network realizes the following steps when executed by the processor:
Sampling the acquired existing speed data set to obtain a component data matrix related to the time sequence;
constructing a time attention network;
inputting the component data matrix into the time attention network to obtain a time correlation matrix;
Constructing a spatial attention network;
inputting the time correlation matrix into the spatial attention network, and fusing to obtain a space-time correlation matrix;
inputting a road network topology matrix into a graph rolling network, combining the road network topology matrix with the space-time correlation matrix, and performing model training;
and carrying out speed prediction through the trained model.
In this scheme, the sampling the existing velocity data set to obtain the component data matrix related to the time sequence specifically includes:
Sampling the existing speed data set in three dimensions to respectively obtain a time division component data matrix, a day component data matrix and a week component data matrix of a related time sequence;
the road network topology matrix is input into a graph convolution network and combined with the space-time correlation matrix, and model training is specifically carried out as follows:
The road network topology matrix is input into a graph rolling network, is combined with the space-time correlation matrix of three dimensions respectively, and adopts the component data matrix of three dimensions to perform model training respectively, and performs fusion output.
In this scheme, before carrying out speed prediction through the model of training, still include:
built loss function
Wherein Y t represents the actual traffic speed,Indicating the prediction speed, L reg is the super-parameter for reducing the prediction error in order to avoid over-fitting the parameters.
A third aspect of the present invention provides a computer-readable storage medium having embodied therein a traffic speed prediction method program for a machine based on a spatiotemporal strive-graph convolutional network, which when executed by a processor, implements the steps of a traffic speed prediction method for a spatiotemporal strive-graph convolutional network as described in any of the above.
The invention discloses a traffic speed prediction method and a related device based on a space-time attention-seeking convolutional network, wherein the method comprises the following steps: sampling the acquired existing speed data set to obtain a component data matrix related to the time sequence; constructing a time attention network; inputting the component data matrix into a time attention network to obtain a time correlation matrix; constructing a spatial attention network; inputting the time correlation matrix into a spatial attention network, and fusing to obtain a space-time correlation matrix; inputting the road network topology matrix into a graph rolling network, combining the road network topology matrix with the space-time correlation matrix, and performing model training; and carrying out speed prediction through the trained model. The time attention network and the space attention network are fused to form a space-time attention network, the relevance of the traffic network in time and space is extracted, meanwhile, the information of adjacent nodes is continuously fused by combining the graph convolution network, so that the traffic speed is predicted, and the accuracy of the traffic speed prediction is improved.
Drawings
FIG. 1 illustrates a flow chart of a traffic speed prediction method based on a spatiotemporal attention-seeking graph convolutional network in accordance with the present application;
FIG. 2 illustrates a block diagram of a traffic speed prediction system based on a spatiotemporal attention-seeking graph convolutional network in accordance with the present invention.
Detailed Description
In order that the above-recited objects, features and advantages of the present application will be more clearly understood, a more particular description of the application will be rendered by reference to the appended drawings and appended detailed description. It should be noted that, without conflict, the embodiments of the present application and features in the embodiments may be combined with each other.
In the following description, numerous specific details are set forth in order to provide a thorough understanding of the present invention, but the present invention may be practiced in other ways than those described herein, and therefore the scope of the present invention is not limited to the specific embodiments disclosed below.
FIG. 1 illustrates a flow chart of a traffic speed prediction method based on a spatiotemporal attention-seeking graph convolutional network in accordance with the present application.
As shown in fig. 1, the application discloses a traffic speed prediction method based on a space-time attention-seeking convolutional network, which comprises the following steps:
S102, sampling the acquired existing speed data set to obtain a component data matrix of a related time sequence;
s104, constructing a time attention network;
S106, inputting the component data matrix into the time attention network to obtain a time correlation matrix;
s108, constructing a spatial attention network;
s110, inputting the time correlation matrix into the spatial attention network, and fusing to obtain a space-time correlation matrix;
s112, inputting the road network topology matrix into a graph convolution network, and combining the road network topology matrix with the space-time correlation matrix to perform model training;
s114, carrying out speed prediction through a trained model.
Firstly, data sampling is carried out according to the obtained data, data related to a time sequence is obtained, and speed characteristics are extracted from the sampled data; a time attention network is constructed, the sampled data is input into the constructed time attention network, and the time correlation between speeds is extracted. And then constructing a spatial attention network, taking the result obtained after further processing as input, inputting the input into the constructed spatial attention network, and fusing and extracting the space-time relevance.
Then, a space-time attention-seeking convolution model is constructed according to the road network matrix, and model training is performed by using space-time associated data obtained by training the graph rolling network GCN and combining the space-time associated data with the road network topology graph. And finally, carrying out speed prediction according to the trained model.
According to the traffic speed prediction method based on the space-time attention force diagram convolution network, the space-time attention network is formed by fusing the time attention network and the space attention network, the relevance of the traffic network in time and space is extracted, meanwhile, the information of adjacent nodes is continuously fused by combining a graph rolling network (GCN), so that the traffic speed is predicted, and the accuracy of the traffic speed prediction is improved by capturing the space-time relevance among the traffic networks.
According to an embodiment of the present invention, the sampling the existing velocity data set to obtain the component data matrix related to the time sequence specifically includes:
Sampling the existing speed data set in three dimensions to respectively obtain a time division component data matrix, a day component data matrix and a week component data matrix of a related time sequence;
the road network topology matrix is input into a graph convolution network and combined with the space-time correlation matrix, and model training is specifically carried out as follows:
The road network topology matrix is input into a graph rolling network, is combined with the space-time correlation matrix of three dimensions respectively, and adopts the component data matrix of three dimensions to perform model training respectively, and performs fusion output.
It should be noted that, according to the rule of data set collection, data of three time dimensions are collected on the data set: the data of three different time dimensions are collected, specifically Y h、Yd and Y w, respectively, where Y h represents the time-division component data matrix of the most recent time, Y d represents the most recent day component data matrix, and Y w represents the most recent week component data matrix, and these three time components share the same network structure. Three sampling batches of different time dimensions can be partitioned according to the regularity of sampling data acquired every 5 minutes. And processing the sampled data, extracting speed characteristics, and constructing a training set and a testing set according to the speed characteristics. In the embodiment of the invention, the traffic speed of one hour in the future can be predicted, so that 12 time windows are taken, and the last 12 time steps are taken as test samples on the basis.
According to an embodiment of the present invention, the inputting the component data matrix into the time attention network to obtain the time correlation matrix specifically includes:
Inputting the component data matrix into the time attention network to obtain a time attention matrix T:
Where K e、L1、L2、L3、Ve is a learnable parameter, σ represents a sigmoid activation function, For the matrix of the component data,C r-1 represents the channel number of the input data of the r layer, and T r-1 represents the length of the time dimension of the input data of the r layer;
after normalization, the time attention matrix T captures the association strength between nodes according to the time attention moment matrix:
t i,j reflects the time correlation strength between the times i, j, and multiplies the time correlation strength by the component data matrix to obtain a time correlation matrix
In the time dimension, there is a correlation between traffic conditions in different time periods, and the correlation also changes in different cases. Likewise, temporal correlation between data is adaptively captured using a temporal attention mechanism.
A time attention network is constructed in the following main modes: inputting the node speed matrix of the time sequence obtained by sampling into a constructed time attention network to obtain
The time attention matrix T is dynamically calculated based on the input of the current layer, where K e、L1、L2、L3、Ve is a learnable parameter. Sigma represents a sigmoid activation function, defined as follows:
Represented is the time correlation module of layer r. C r-1 denotes the number of channels of the r-layer input data, and T r-1 denotes the length of the r-layer input data time dimension.
The T matrix is normalized and then the correlation strength among the nodes is captured according to the calculated time attention matrix:
T i,j dynamically reflects the time correlation strength between the times i and j, and meanwhile, the original sampling matrix is multiplied to obtain a complete time correlation matrix:
according to an embodiment of the present invention, the inputting the time correlation matrix into the spatial attention network, and the fusing to obtain the space-time correlation matrix specifically includes:
inputting the time correlation matrix into the spatial attention network to obtain a space-time attention matrix S:
wherein K s、H1、H2、H3、Vs is a learnable parameter;
And calculating a time correlation matrix S' according to the space-time attention matrix S:
It should be noted that, in the spatial dimension, traffic conditions at different locations affect each other. The embodiment of the invention uses a space attention mechanism to adaptively capture the dynamic phase between nodes in the space dimension, and fuses the matrix with time attention with the space attention network to obtain the space-time attention matrix.
The output (time correlation matrix) obtained by the time attention network is input into the space attention network to be further fused to obtain a space-time attention matrix S:
Based on the input of the time-attention matrix, the spatiotemporal attention S is dynamically calculated, where K s、H1、H2、H3、Vs is a learnable parameter.
Capturing the association strength between the nodes according to the calculated space-time attention matrix:
s' dynamically reflects the space-time correlation strength between the times i and j;
And dynamically adjusting weights according to the obtained space-time correlation matrix S' and the input road network adjacent matrix to generate the attention coefficient of the node. Through space-time correlation, the correlation of the road network speed in time space can be better grasped, and a wider visual field is provided for the speed prediction.
According to the embodiment of the invention, the method for inputting the road network topology matrix into the graph rolling network and combining the road network topology matrix with the space-time correlation matrix to perform model training comprises the following steps:
Inputting the road network topology matrix into a Laplace matrix L:
wherein A represents an input road network topology matrix, D represents a degree matrix, specifically a diagonal matrix, and the elements of the diagonal are A ij represents an element of row i and column j;
The graph rolling network is specifically in the form of chebyshev polynomials, and is expressed as follows:
where G represents the convolution operation of a graph, Is the scaled normalized laplace matrix, λ max is the maximum eigenvalue of L, θ' k (k=0, 1, … K) is the coefficient of the kth term of chebyshev polynomial, which is a learnable parameter;
The definition of the chebyshev polynomial of the K-th order is as follows:
Tk(x)＝2xTk-1(x)-Tk-2(x)；
combining the input road network topology matrix and the space-time correlation matrix to obtain a final graph convolution formula:
And after model training is carried out on the graph convolution formula, speed prediction is carried out through the trained model.
It should be noted that, the traffic speed prediction based on the graph convolution network predicts the future traffic condition according to the historical traffic data, and constructs a space-time attention-seeking graph convolution network (STA-GCN) model through the space-time correlation matrix and the road network topology matrix.
In order to capture the relevance of the road network, the invention adopts the graph volumes of the Chebyshev polynomials to accumulate the information of the neighbor nodes. Using a laplace matrix according to a road network:
A represents an input road network topology matrix, D is a degree matrix, and is a diagonal matrix, and the elements of the diagonal are A ij represents an element of the ith row and j column.
The graph roll-up network in chebyshev polynomial form is expressed as: Where G represents the convolution operation of a graph. Since the convolution operation of the plot signals is equal to the fourier transform of the product plot of these signals that have been converted to the spectral domain, the above formula can be understood as fourier transforming g Θ and x into the spectral domain, respectively, and then re-multiplying them and performing the inverse fourier transform to obtain the final result of the convolution operation. However, it is difficult to directly decompose eigenvalues of the laplace matrix, especially for road networks. The present invention thus approximates, but effectively solves this problem, using chebyshev polynomials, Is a scaled normalized laplace matrix, λ max is the maximum eigenvalue of L, θ' k (k=0, 1, … K) is the coefficient of the kth term of chebyshev polynomial, which is a learnable parameter that is updated iteratively over time by optimizing the loss function during model training. The definition of the chebyshev polynomial of the K-th order is as follows:
Tk(x)＝2xTk-1(x)-Tk-2(x)；
solving this equation using the approximate expansion of the chebyshev polynomial corresponds to extracting information for each node around.
Combining the input road network adjacency matrix with the obtained space-time correlation matrix to obtain a final graph convolution formula:
The relevance of the road network in time and space is extracted through a time-space attention mechanism, modeling is carried out by combining a road network matrix, and the network with three different time dimensions is trained.
After the same network learning according to three different time dimensions, the output is fused:
The predicted node's dependence on each dimension is different for the three time dimensions, so the three components of each node have different impact weights and should be learned from historical data. Wherein, the ". Is Hadamard product. W h、Wd and W w are learning parameters reflecting the extent to which the three-dimensional time component affects the prediction target. And obtaining the predicted speed according to the final fusion output.
According to an embodiment of the present invention, before the speed prediction by the trained model, the method further includes:
Constructed loss function
Wherein Y t represents the actual traffic speed,Indicating the prediction speed, L reg is the super-parameter for reducing the prediction error in order to avoid over-fitting the parameters.
It should be noted that, according to the difference comparison between the predicted speed and the actual speed of the space-time attention-seeking convolution model, a new loss function calculation method is provided, which is continuously updated during each iteration, so as to accelerate the model training speed, and the method can make the networks of three different time dimensions converge at the same time.
The constructed loss function is defined as follows:
the constructed loss function is
Where Y t represents the actual traffic speed,The predicted speed is indicated as such,For minimizing the error between the actual traffic speed and the prediction. L reg helps to avoid the problem of overfitting, whereas ω is an over-parameter that minimizes the error prediction between actual traffic speed and actual traffic speed.
The space-time attention mechanism can learn the implicit space-time correlation among the nodes according to the characteristics of each node in the input data, and the attention score among the nodes is dynamically calculated according to the input, so that the attention score can be captured when the topological structure of the road network changes; in addition, since spatial self-attention gathers information of all nodes, it can also capture spatial correlation of road network from the global.
FIG. 2 illustrates a block diagram of a traffic speed prediction system based on a spatiotemporal attention-seeking graph convolutional network in accordance with the present invention.
As shown in fig. 2, the invention discloses a traffic speed prediction system 2 based on a space-time attention-seeking convolutional network, which comprises a memory 21 and a processor 22, wherein the memory 21 comprises a traffic speed prediction method program based on the space-time attention-seeking convolutional network, and the traffic speed prediction method program based on the space-time attention-seeking convolutional network realizes the following steps when being executed by the processor 22:
Sampling the acquired existing speed data set to obtain a component data matrix related to the time sequence;
constructing a time attention network;
inputting the component data matrix into the time attention network to obtain a time correlation matrix;
Constructing a spatial attention network;
inputting the time correlation matrix into the spatial attention network, and fusing to obtain a space-time correlation matrix;
inputting a road network topology matrix into a graph rolling network, combining the road network topology matrix with the space-time correlation matrix, and performing model training;
and carrying out speed prediction through the trained model.
Firstly, data sampling is carried out according to the obtained data, data related to a time sequence is obtained, and speed characteristics are extracted from the sampled data; a time attention network is constructed, the sampled data is input into the constructed time attention network, and the time correlation between speeds is extracted. And then constructing a spatial attention network, taking the result obtained after further processing as input, inputting the input into the constructed spatial attention network, and fusing and extracting the space-time relevance.
Then, a space-time attention-seeking convolution model is constructed according to the road network matrix, and model training is performed by using space-time associated data obtained by training the graph rolling network GCN and combining the space-time associated data with the road network topology graph. And finally, carrying out speed prediction according to the trained model.
According to the traffic speed prediction method based on the space-time attention force diagram convolution network, the space-time attention network is formed by fusing the time attention network and the space attention network, the relevance of the traffic network in time and space is extracted, meanwhile, the information of adjacent nodes is continuously fused by combining a graph rolling network (GCN), so that the traffic speed is predicted, and the accuracy of the traffic speed prediction is improved by capturing the space-time relevance among the traffic networks.
According to an embodiment of the present invention, the sampling the existing velocity data set to obtain the component data matrix related to the time sequence specifically includes:
Sampling the existing speed data set in three dimensions to respectively obtain a time division component data matrix, a day component data matrix and a week component data matrix of a related time sequence;
the road network topology matrix is input into a graph convolution network and combined with the space-time correlation matrix, and model training is specifically carried out as follows:
The road network topology matrix is input into a graph rolling network, is combined with the space-time correlation matrix of three dimensions respectively, and adopts the component data matrix of three dimensions to perform model training respectively, and performs fusion output.
It should be noted that, according to the rule of data set collection, data of three time dimensions are collected on the data set: the data of three different time dimensions are collected, specifically Y h、Yd and Y w, respectively, where Y h represents the time-division component data matrix of the most recent time, Y d represents the most recent day component data matrix, and Y w represents the most recent week component data matrix, and these three time components share the same network structure. Three sampling batches of different time dimensions can be partitioned according to the regularity of sampling data acquired every 5 minutes. And processing the sampled data, extracting speed characteristics, and constructing a training set and a testing set according to the speed characteristics. In the embodiment of the invention, the traffic speed of one hour in the future can be predicted, so that 12 time windows are taken, and the last 12 time steps are taken as test samples on the basis.
According to an embodiment of the present invention, the inputting the component data matrix into the time attention network to obtain the time correlation matrix specifically includes:
Inputting the component data matrix into the time attention network to obtain a time attention matrix T:
Where K e、L1、L2、L3、Ve is a learnable parameter, σ represents a sigmoid activation function, For the matrix of the component data,C r-1 represents the channel number of the input data of the r layer, and T r-1 represents the length of the time dimension of the input data of the r layer;
after normalization, the time attention matrix T captures the association strength between nodes according to the time attention moment matrix:
t i,j reflects the time correlation strength between the times i, j, and multiplies the time correlation strength by the component data matrix to obtain a time correlation matrix
In the time dimension, there is a correlation between traffic conditions in different time periods, and the correlation also changes in different cases. Likewise, temporal correlation between data is adaptively captured using a temporal attention mechanism.
A time attention network is constructed in the following main modes: inputting the node speed matrix of the time sequence obtained by sampling into a constructed time attention network to obtain
The time attention matrix T is dynamically calculated based on the input of the current layer, where K e、L1、L2、L3、Ve is a learnable parameter. Sigma represents a sigmoid activation function, defined as follows:
Represented is the time correlation module of layer r. C r-1 denotes the number of channels of the r-layer input data, and T r-1 denotes the length of the r-layer input data time dimension.
The T matrix is normalized and then the correlation strength among the nodes is captured according to the calculated time attention matrix:
T i,j dynamically reflects the time correlation strength between the times i and j, and meanwhile, the original sampling matrix is multiplied to obtain a complete time correlation matrix:
according to an embodiment of the present invention, the inputting the time correlation matrix into the spatial attention network, and the fusing to obtain the space-time correlation matrix specifically includes:
inputting the time correlation matrix into the spatial attention network to obtain a space-time attention matrix S:
wherein K s、H1、H2、H3、Vs is a learnable parameter;
And calculating a time correlation matrix S' according to the space-time attention matrix S:
It should be noted that, in the spatial dimension, traffic conditions at different locations affect each other. The embodiment of the invention uses a space attention mechanism to adaptively capture the dynamic phase between nodes in the space dimension, and fuses the matrix with time attention with the space attention network to obtain the space-time attention matrix.
The output (time correlation matrix) obtained by the time attention network is input into the space attention network to be further fused to obtain a space-time attention matrix S:
Based on the input of the time-attention matrix, the spatiotemporal attention S is dynamically calculated, where K s、H1、H2、H3、Vs is a learnable parameter.
Capturing the association strength between the nodes according to the calculated space-time attention matrix:
s' dynamically reflects the space-time correlation strength between the times i and j;
And dynamically adjusting weights according to the obtained space-time correlation matrix S' and the input road network adjacent matrix to generate the attention coefficient of the node. Through space-time correlation, the correlation of the road network speed in time space can be better grasped, and a wider visual field is provided for the speed prediction.
According to the embodiment of the invention, the method for inputting the road network topology matrix into the graph rolling network and combining the road network topology matrix with the space-time correlation matrix to perform model training comprises the following steps:
Inputting the road network topology matrix into a Laplace matrix L:
wherein A represents an input road network topology matrix, D represents a degree matrix, specifically a diagonal matrix, and the elements of the diagonal are A ij represents an element of row i and column j;
The graph rolling network is specifically in the form of chebyshev polynomials, and is expressed as follows:
where G represents the convolution operation of a graph, Is the scaled normalized laplace matrix, λ max is the maximum eigenvalue of L, θ' k (k=0, 1, … K) is the coefficient of the kth term of chebyshev polynomial, which is a learnable parameter;
The definition of the chebyshev polynomial of the K-th order is as follows:
Tk(x)＝2xTk-1(x)-Tk-2(x)；
combining the input road network topology matrix and the space-time correlation matrix to obtain a final graph convolution formula:
And after model training is carried out on the graph convolution formula, speed prediction is carried out through the trained model.
It should be noted that, the traffic speed prediction based on the graph rolling network predicts the future traffic condition according to the historical traffic data, and constructs the STA-GCN model through the space-time correlation matrix and the road network topology matrix.
In order to capture the relevance of the road network, the invention adopts the graph volumes of the Chebyshev polynomials to accumulate the information of the neighbor nodes. Using a laplace matrix according to a road network:
A represents an input road network topology matrix, D is a degree matrix, and is a diagonal matrix, and the elements of the diagonal are A ij represents an element of the ith row and j column.
The graph roll-up network in chebyshev polynomial form is expressed as: Where G represents the convolution operation of a graph. Since the convolution operation of the plot signals is equal to the fourier transform of the product plot of these signals that have been converted to the spectral domain, the above formula can be understood as fourier transforming g Θ and x into the spectral domain, respectively, and then re-multiplying them and performing the inverse fourier transform to obtain the final result of the convolution operation. However, it is difficult to directly decompose eigenvalues of the laplace matrix, especially for road networks. The present invention thus approximates, but effectively solves this problem, using chebyshev polynomials, Is a scaled normalized laplace matrix, λ max is the maximum eigenvalue of L, θ' k (k=0, 1, … K) is the coefficient of the kth term of chebyshev polynomial, which is a learnable parameter that is updated iteratively over time by optimizing the loss function during model training. The definition of the chebyshev polynomial of the K-th order is as follows:
Tk(x)＝2xTk-1(x)-Tk-2(x)；
solving this equation using the approximate expansion of the chebyshev polynomial corresponds to extracting information for each node around.
Combining the input road network adjacency matrix with the obtained space-time correlation matrix to obtain a final graph convolution formula:
The relevance of the road network in time and space is extracted through a time-space attention mechanism, modeling is carried out by combining a road network matrix, and the network with three different time dimensions is trained.
After the same network learning according to three different time dimensions, the output is fused:
The predicted node's dependence on each dimension is different for the three time dimensions, so the three components of each node have different impact weights and should be learned from historical data. Wherein, the ". Is Hadamard product. W h、Wd and W w are learning parameters reflecting the extent to which the three-dimensional time component affects the prediction target. And obtaining the predicted speed according to the final fusion output.
According to an embodiment of the present invention, before the speed prediction by the trained model, the method further includes:
Constructed loss function
Wherein Y t represents the actual traffic speed,Indicating the prediction speed, L reg is the super-parameter for reducing the prediction error in order to avoid over-fitting the parameters.
It should be noted that, according to the difference comparison between the predicted speed and the actual speed of the space-time attention-seeking convolution model, a new loss function calculation method is provided, which is continuously updated during each iteration, so as to accelerate the model training speed, and the method can make the networks of three different time dimensions converge at the same time.
The constructed loss function is defined as follows:
the constructed loss function is
Where Y t represents the actual traffic speed,The predicted speed is indicated as such,For minimizing the error between the actual traffic speed and the prediction. L reg helps to avoid the problem of overfitting, whereas ω is an over-parameter that minimizes the error prediction between actual traffic speed and actual traffic speed.
The space-time attention mechanism can learn the implicit space-time correlation among the nodes according to the characteristics of each node in the input data, and the attention score among the nodes is dynamically calculated according to the input, so that the attention score can be captured when the topological structure of the road network changes; in addition, since spatial self-attention gathers information of all nodes, it can also capture spatial correlation of road network from the global.
A third aspect of the present invention provides a computer-readable storage medium having embodied therein a traffic speed prediction method program for a machine based on a spatiotemporal strive-graph convolutional network, which when executed by a processor, implements the steps of a traffic speed prediction method for a spatiotemporal strive-graph convolutional network as described in any of the above.
The invention discloses a traffic speed prediction method and a related device based on a space-time attention-seeking convolutional network, wherein the method comprises the following steps: sampling the acquired existing speed data set to obtain a component data matrix related to the time sequence; constructing a time attention network; inputting the component data matrix into a time attention network to obtain a time correlation matrix; constructing a spatial attention network; inputting the time correlation matrix into a spatial attention network, and fusing to obtain a space-time correlation matrix; inputting the road network topology matrix into a graph rolling network, combining the road network topology matrix with the space-time correlation matrix, and performing model training; and carrying out speed prediction through the trained model. The time attention network and the space attention network are fused to form a space-time attention network, the relevance of the traffic network in time and space is extracted, meanwhile, the information of adjacent nodes is continuously fused by combining a graph rolling network (GCN), so that the traffic speed is predicted, and the accuracy of the traffic speed prediction is improved by capturing the space-time relevance among the traffic networks.
The following is an application example of a traffic speed prediction method based on a space-time attention-seeking convolutional network, which is provided by the invention:
1. Preprocessing of data
The present invention was tested on PEMSD and PEMSD8 datasets. The PEMSD data set contains traffic data (including flow, speed, lane occupancy) for 307 ring detectors in san francisco bay area from 1 st 2018, 1 st to 2 nd 2018. The PEMSD data set contains traffic data (including traffic, speed, lane occupancy) collected by 170 ring detectors in the san Bei Nadi nuo region from 7.1 in 2016 to 8.31 in 2016. The raw data includes two parts, one part is traffic data and the other part is the distance between the individual sensors. The data preprocessing comprises the steps of segmenting a data set to manufacture training and testing samples.
2. Model training
And loading the processed data set into the formulas (1) - (5) in the embodiment to obtain the space-time correlation matrix S' among the nodes.
And (3) combining the input road gateway joint matrix A, and inputting the space-time correlation matrix S' into a formula (6) to perform model training. Then 2 performance indexes of average absolute error MAE and root mean square error RMSE are calculated, and the 2 performance indexes are defined as follows:
Wherein, X i is a total number of the components, Representing the i-th element in the true value and the predicted value, respectively, and n represents the total number of elements.
Table 1 comparison of the invention with HA, ARIMA, VAR on PEMSD data
Table 2 comparison of the invention with HA, ARIMA, VAR on PEMSD8 data
It can be found that the prediction of the invention achieves better effect to various degrees. The graph network fused with the time-space correlation has good effect on the accuracy of prediction, and has certain reference value and practical economic benefit.
In the several embodiments provided by the present application, it should be understood that the disclosed apparatus and method may be implemented in other ways. The above described device embodiments are only illustrative, e.g. the division of the units is only one logical function division, and there may be other divisions in practice, such as: multiple units or components may be combined or may be integrated into another system, or some features may be omitted, or not performed. In addition, the various components shown or discussed may be coupled or directly coupled or communicatively coupled to each other via some interface, whether indirectly coupled or communicatively coupled to devices or units, whether electrically, mechanically, or otherwise.
The units described above as separate components may or may not be physically separate, and components shown as units may or may not be physical units; can be located in one place or distributed to a plurality of network units; some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.
In addition, each functional unit in each embodiment of the present invention may be integrated in one processing unit, or each unit may be separately used as one unit, or two or more units may be integrated in one unit; the integrated units may be implemented in hardware or in hardware plus software functional units.
Those of ordinary skill in the art will appreciate that: all or part of the steps for implementing the above method embodiments may be implemented by hardware related to program instructions, and the foregoing program may be stored in a computer readable storage medium, where the program, when executed, performs steps including the above method embodiments; and the aforementioned storage medium includes: a mobile storage device, a Read-Only Memory (ROM), a random access Memory (RAM, random Access Memory), a magnetic disk or optical disk, or the like, which can store program codes.
Or the above-described integrated units of the invention may be stored in a computer-readable storage medium if implemented in the form of software functional modules and sold or used as separate products. Based on such understanding, the technical solutions of the embodiments of the present invention may be embodied in essence or a part contributing to the prior art in the form of a software product stored in a storage medium, including several instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute all or part of the methods described in the embodiments of the present invention. And the aforementioned storage medium includes: a removable storage device, ROM, RAM, magnetic or optical disk, or other medium capable of storing program code.

Claims (3)

1. A traffic speed prediction method based on a spatiotemporal attention-seeking graph convolutional network, comprising the steps of:
Sampling the acquired existing speed data set to obtain a component data matrix related to the time sequence;
constructing a time attention network;
inputting the component data matrix into the time attention network to obtain a time correlation matrix;
Constructing a spatial attention network;
inputting the time correlation matrix into the spatial attention network, and fusing to obtain a space-time correlation matrix;
inputting a road network topology matrix into a graph rolling network, combining the road network topology matrix with the space-time correlation matrix, and performing model training;
Carrying out speed prediction through a trained model;
Inputting the component data matrix into the time attention network to obtain a time correlation matrix specifically comprises:
Inputting the component data matrix into the time attention network to obtain a time attention matrix T:
Where K e、L1、L2、L3、Ve is a learnable parameter, σ represents a sigmoid activation function, For the matrix of the component data,C r-1 represents the channel number of the input data of the r layer, and T r-1 represents the length of the time dimension of the input data of the r layer;
after normalization, the time attention matrix T captures the association strength between nodes according to the time attention moment matrix:
T i,j reflects the time correlation strength between the times i, j, and multiplies the time correlation strength by the component data matrix to obtain a time correlation matrix
Inputting the time correlation matrix into the spatial attention network, and fusing to obtain a space-time correlation matrix specifically comprises the following steps:
inputting the time correlation matrix into the spatial attention network to obtain a space-time attention matrix S:
wherein K s、H1、H2、H3、Vs is a learnable parameter;
And calculating a time correlation matrix S' according to the space-time attention matrix S:
Inputting the road network topology matrix into a graph convolution network, combining the road network topology matrix with the space-time correlation matrix, and performing model training, wherein the obtaining of the prediction speed specifically comprises the following steps:
Inputting the road network topology matrix into a Laplace matrix L:
wherein A represents an input road network topology matrix, D represents a degree matrix, specifically a diagonal matrix, and the elements of the diagonal are A ij represents an element of row i and column j;
The graph rolling network is specifically in the form of chebyshev polynomials, and is expressed as follows:
where G represents the convolution operation of a graph, Is the scaled normalized laplace matrix, λmax is the maximum eigenvalue of L, θ' k (k=0, 1, … K) is the coefficient of the kth term of chebyshev polynomial, which is a learnable parameter;
The definition of the chebyshev polynomial of the K-th order is as follows:
Tk(x)＝2xTk-1(x)-Tk-2(x)；
combining the input road network topology matrix and the space-time correlation matrix to obtain a final graph convolution formula:
After model training is carried out on the graph convolution formula, speed prediction is carried out through a trained model;
The sampling the existing speed data set to obtain a component data matrix related to the time sequence specifically comprises the following steps:
Sampling the existing speed data set in three dimensions to respectively obtain a time division component data matrix, a day component data matrix and a week component data matrix of a related time sequence;
the road network topology matrix is input into a graph convolution network and combined with the space-time correlation matrix, and model training is specifically carried out as follows:
Inputting a road network topology matrix into a graph rolling network, respectively combining the road network topology matrix with the space-time correlation matrix of three dimensions, respectively performing model training by adopting the component data matrix of three dimensions, and performing fusion output;
before the speed prediction is performed through the trained model, the method further comprises the following steps:
Constructed loss function
Wherein Y t represents the actual traffic speed,Indicating the prediction speed, L reg is the super-parameter for reducing the prediction error in order to avoid over-fitting the parameters.
2. A traffic speed prediction system based on a spatiotemporal strikings-striving-network comprising a memory and a processor, said memory comprising a traffic speed prediction method program based on a spatiotemporal strikings-convolving-network, said traffic speed prediction method program based on a spatiotemporal strikings-convolving-network implementing the steps when executed by said processor:
Sampling the acquired existing speed data set to obtain a component data matrix related to the time sequence;
constructing a time attention network;
inputting the component data matrix into the time attention network to obtain a time correlation matrix;
Constructing a spatial attention network;
inputting the time correlation matrix into the spatial attention network, and fusing to obtain a space-time correlation matrix;
inputting a road network topology matrix into a graph rolling network, combining the road network topology matrix with the space-time correlation matrix, and performing model training;
Carrying out speed prediction through a trained model;
Inputting the component data matrix into the time attention network to obtain a time correlation matrix specifically comprises:
Inputting the component data matrix into the time attention network to obtain a time attention matrix T:
Where K e、L1、L2、L3、Ve is a learnable parameter, σ represents a sigmoid activation function, For the matrix of the component data,C r-1 represents the channel number of the input data of the r layer, and T r-1 represents the length of the time dimension of the input data of the r layer;
after normalization, the time attention matrix T captures the association strength between nodes according to the time attention moment matrix:
T i,j reflects the time correlation strength between the times i, j, and multiplies the time correlation strength by the component data matrix to obtain a time correlation matrix
Inputting the time correlation matrix into the spatial attention network, and fusing to obtain a space-time correlation matrix specifically comprises the following steps:
inputting the time correlation matrix into the spatial attention network to obtain a space-time attention matrix S:
wherein K s、H1、H2、H3、Vs is a learnable parameter;
And calculating a time correlation matrix S' according to the space-time attention matrix S:
Inputting the road network topology matrix into a graph convolution network, combining the road network topology matrix with the space-time correlation matrix, and performing model training, wherein the obtaining of the prediction speed specifically comprises the following steps:
Inputting the road network topology matrix into a Laplace matrix L:
wherein A represents an input road network topology matrix, D represents a degree matrix, specifically a diagonal matrix, and the elements of the diagonal are A ij represents an element of row i and column j;
The graph rolling network is specifically in the form of chebyshev polynomials, and is expressed as follows:
where G represents the convolution operation of a graph, Is the scaled normalized laplace matrix, λ max is the maximum eigenvalue of L, θ' k (k=0, 1, … K) is the coefficient of the kth term of chebyshev polynomial, which is a learnable parameter;
The definition of the chebyshev polynomial of the K-th order is as follows:
Tk(x)＝2xTk-1(x)-Tk-2(x)；
combining the input road network topology matrix and the space-time correlation matrix to obtain a final graph convolution formula:
After model training is carried out on the graph convolution formula, speed prediction is carried out through a trained model;
The sampling the existing speed data set to obtain a component data matrix related to the time sequence specifically comprises the following steps:
Sampling the existing speed data set in three dimensions to respectively obtain a time division component data matrix, a day component data matrix and a week component data matrix of a related time sequence;
the road network topology matrix is input into a graph convolution network and combined with the space-time correlation matrix, and model training is specifically carried out as follows:
Inputting a road network topology matrix into a graph rolling network, respectively combining the road network topology matrix with the space-time correlation matrix of three dimensions, respectively performing model training by adopting the component data matrix of three dimensions, and performing fusion output;
before the speed prediction is performed through the trained model, the method further comprises the following steps:
Constructed loss function
Wherein Y t represents the actual traffic speed,Indicating the prediction speed, L reg is the super-parameter for reducing the prediction error in order to avoid over-fitting the parameters.
3. A computer readable storage medium, comprising a traffic speed prediction method program based on a spatiotemporal strikings force diagram convolutional network, which when executed by a processor, implements the steps of a traffic speed prediction method based on a spatiotemporal strikings force diagram convolutional network as claimed in claim 1.
CN202110980592.8A 2021-08-25 2021-08-25 Traffic speed prediction method and device based on space-time attention force diagram convolution network Active CN113705880B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202110980592.8A CN113705880B (en) 2021-08-25 2021-08-25 Traffic speed prediction method and device based on space-time attention force diagram convolution network

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202110980592.8A CN113705880B (en) 2021-08-25 2021-08-25 Traffic speed prediction method and device based on space-time attention force diagram convolution network

Publications (2)

Publication Number Publication Date
CN113705880A CN113705880A (en) 2021-11-26
CN113705880B true CN113705880B (en) 2024-09-20

Family

ID=78654602

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202110980592.8A Active CN113705880B (en) 2021-08-25 2021-08-25 Traffic speed prediction method and device based on space-time attention force diagram convolution network

Country Status (1)

Country Link
CN (1) CN113705880B (en)

Families Citing this family (15)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114038200B (en) * 2021-11-29 2022-09-20 东北大学 Attention mechanism-based time-space synchronization map convolutional network traffic flow prediction method
CN113962495A (en) * 2021-12-21 2022-01-21 华控清交信息科技(北京)有限公司 Wind power plant power prediction method and device and electronic equipment
CN114529081B (en) * 2022-02-18 2024-06-11 哈尔滨工程大学 Space-time combined traffic flow prediction method and device
CN115169426B (en) * 2022-02-22 2023-10-13 国网山东省电力公司信息通信公司 Anomaly detection method and system based on similarity learning fusion model
CN114511154B (en) * 2022-02-22 2024-06-14 南通大学 Traffic flow prediction method based on space-time complex graph convolutional network
CN114743374B (en) * 2022-04-01 2023-04-07 清华大学深圳国际研究生院 Multi-element traffic flow prediction method based on graph network
CN114900357A (en) * 2022-05-07 2022-08-12 长三角信息智能创新研究院 Method, equipment and storage medium for detecting flooding attack of time-space diagram neural network link
CN115063972B (en) * 2022-05-18 2024-06-11 浙江工业大学 Traffic speed prediction method and system based on graph convolution and gating circulation unit
CN115081717B (en) * 2022-06-27 2023-03-24 北京建筑大学 Rail transit passenger flow prediction method integrating attention mechanism and graph neural network
CN115307780B (en) * 2022-09-29 2023-01-06 中国海洋大学 Sea surface temperature prediction method, system and application based on time-space information interaction fusion
KR20240050822A (en) * 2022-10-12 2024-04-19 현대자동차주식회사 Apparatus for predicting traffic speed and method thereof
CN115758173B (en) * 2022-11-16 2023-09-26 西华大学 Cloud platform system anomaly detection method and device based on parallel graph attention network
CN116109021B (en) * 2023-04-13 2023-08-15 中国科学院大学 Travel time prediction method, device, equipment and medium based on multitask learning
CN117133129B (en) * 2023-10-26 2024-01-30 山东科技大学 Traffic speed prediction method based on multi-component attention-seeking neural network
CN118171938B (en) * 2024-05-13 2024-08-13 深圳市城市交通规划设计研究中心股份有限公司 Traffic fusion prediction method

Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111368815A (en) * 2020-05-28 2020-07-03 之江实验室 Pedestrian re-identification method based on multi-component self-attention mechanism
CN112071065A (en) * 2020-09-16 2020-12-11 山东理工大学 Traffic flow prediction method based on global diffusion convolution residual error network

Family Cites Families (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112183862A (en) * 2020-09-29 2021-01-05 长春理工大学 Traffic flow prediction method and system for urban road network
CN112241814B (en) * 2020-10-20 2022-12-02 河南大学 Traffic prediction method based on reinforced space-time diagram neural network
CN112801404B (en) * 2021-02-14 2024-03-22 北京工业大学 Traffic prediction method based on self-adaptive space self-attention force diagram convolution
CN112905379B (en) * 2021-03-10 2023-07-18 南京理工大学 Traffic big data restoration method of graph self-encoder based on self-attention mechanism
CN113128237B (en) * 2021-04-09 2023-04-07 青岛海大新星软件咨询有限公司 Semantic representation model construction method for service resources

Patent Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111368815A (en) * 2020-05-28 2020-07-03 之江实验室 Pedestrian re-identification method based on multi-component self-attention mechanism
CN112071065A (en) * 2020-09-16 2020-12-11 山东理工大学 Traffic flow prediction method based on global diffusion convolution residual error network

Also Published As

Publication number Publication date
CN113705880A (en) 2021-11-26

Similar Documents

Publication Publication Date Title
CN113705880B (en) Traffic speed prediction method and device based on space-time attention force diagram convolution network
CN112801404B (en) Traffic prediction method based on self-adaptive space self-attention force diagram convolution
CN113450568B (en) Convolutional network traffic flow prediction model based on space-time attention mechanism
CN110675623B (en) Short-term traffic flow prediction method, system and device based on hybrid deep learning
CN112350899B (en) Network flow prediction method based on graph convolution network fusion multi-feature input
CN113313947B (en) Road condition evaluation method of short-term traffic prediction graph convolution network
CN112988723B (en) Traffic data restoration method based on space self-attention force diagram convolution cyclic neural network
CN114944053B (en) Traffic flow prediction method based on space-time hypergraph neural network
CN113240182A (en) Short-term traffic flow prediction method, storage medium and system under complex road network
CN114120652A (en) Traffic flow prediction method based on dynamic graph neural network
Cheng et al. Deep belief network for meteorological time series prediction in the Internet of Things
WO2021129668A1 (en) Neural network training method and device
CN115578851A (en) Traffic prediction method based on MGCN
CN112437451B (en) Wireless network flow prediction method and device based on generation countermeasure network
CN115376317B (en) Traffic flow prediction method based on dynamic graph convolution and time sequence convolution network
CN113808396B (en) Traffic speed prediction method and system based on traffic flow data fusion
CN113988263A (en) Knowledge distillation-based space-time prediction method in industrial Internet of things edge equipment
CN115828990A (en) Time-space diagram node attribute prediction method for fused adaptive graph diffusion convolution network
CN117156442A (en) Cloud data security protection method and system based on 5G network
CN116434569A (en) Traffic flow prediction method and system based on STNR model
CN115691129A (en) Traffic flow prediction method of depth residual error space-time diagram convolution network based on attention
CN115423162A (en) Traffic flow prediction method and device, electronic equipment and storage medium
Zhang et al. Ultra-short-term prediction of regional photovoltaic power based on dynamic graph convolutional neural network
CN117392846A (en) Traffic flow prediction method for space-time self-adaptive graph learning fusion dynamic graph convolution
CN117290706A (en) Traffic flow prediction method based on space-time convolution fusion probability sparse attention mechanism

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant