CN112350876A - Network flow prediction method based on graph neural network - Google Patents
Network flow prediction method based on graph neural network Download PDFInfo
- Publication number
- CN112350876A CN112350876A CN202110027901.XA CN202110027901A CN112350876A CN 112350876 A CN112350876 A CN 112350876A CN 202110027901 A CN202110027901 A CN 202110027901A CN 112350876 A CN112350876 A CN 112350876A
- Authority
- CN
- China
- Prior art keywords
- gcn
- gru
- model
- layer
- network
- 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.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 33
- 238000013528 artificial neural network Methods 0.000 title claims abstract description 20
- 239000011159 matrix material Substances 0.000 claims description 30
- 238000011156 evaluation Methods 0.000 claims description 15
- 238000012549 training Methods 0.000 claims description 13
- 238000013135 deep learning Methods 0.000 claims description 6
- 238000012545 processing Methods 0.000 claims description 6
- 238000012360 testing method Methods 0.000 claims description 5
- 238000007781 pre-processing Methods 0.000 claims description 3
- 238000005065 mining Methods 0.000 claims description 2
- 238000001228 spectrum Methods 0.000 claims description 2
- 238000004891 communication Methods 0.000 abstract description 14
- 238000013527 convolutional neural network Methods 0.000 abstract description 4
- 230000008859 change Effects 0.000 abstract description 3
- 230000008569 process Effects 0.000 description 9
- 238000002474 experimental method Methods 0.000 description 8
- 230000000694 effects Effects 0.000 description 6
- 238000010586 diagram Methods 0.000 description 5
- 238000004364 calculation method Methods 0.000 description 4
- 230000006870 function Effects 0.000 description 4
- 238000010801 machine learning Methods 0.000 description 4
- 230000007774 longterm Effects 0.000 description 3
- 238000005457 optimization Methods 0.000 description 3
- 238000004458 analytical method Methods 0.000 description 2
- 230000007423 decrease Effects 0.000 description 2
- 230000007547 defect Effects 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 239000000284 extract Substances 0.000 description 2
- 230000000306 recurrent effect Effects 0.000 description 2
- 241001123248 Arma Species 0.000 description 1
- 230000004913 activation Effects 0.000 description 1
- 238000013459 approach Methods 0.000 description 1
- 238000010835 comparative analysis Methods 0.000 description 1
- 230000002354 daily effect Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000008034 disappearance Effects 0.000 description 1
- 230000003203 everyday effect Effects 0.000 description 1
- 238000004880 explosion Methods 0.000 description 1
- 239000002360 explosive Substances 0.000 description 1
- 238000010295 mobile communication Methods 0.000 description 1
- YHXISWVBGDMDLQ-UHFFFAOYSA-N moclobemide Chemical compound C1=CC(Cl)=CC=C1C(=O)NCCN1CCOCC1 YHXISWVBGDMDLQ-UHFFFAOYSA-N 0.000 description 1
- 230000035772 mutation Effects 0.000 description 1
- 238000003062 neural network model Methods 0.000 description 1
- 238000002203 pretreatment Methods 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 230000004044 response Effects 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 238000012706 support-vector machine Methods 0.000 description 1
- 230000002123 temporal effect Effects 0.000 description 1
- 230000036962 time dependent Effects 0.000 description 1
- 230000003313 weakening effect Effects 0.000 description 1
Images
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L41/00—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
- H04L41/14—Network analysis or design
- H04L41/147—Network analysis or design for predicting network behaviour
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- 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
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L41/00—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
- H04L41/14—Network analysis or design
- H04L41/145—Network analysis or design involving simulating, designing, planning or modelling of a network
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- Signal Processing (AREA)
- Computer Networks & Wireless Communication (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Molecular Biology (AREA)
- Health & Medical Sciences (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- General Health & Medical Sciences (AREA)
- Evolutionary Biology (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Data Exchanges In Wide-Area Networks (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a network flow prediction method based on a graph neural network, which comprises the steps of firstly learning a network topological structure of a communication area block by a graph convolutional neural network (GCN), extracting the spatial characteristics of the network flow of the communication area block, then using data with the spatial characteristics as the input of a threshold recursion unit (GRU), learning the time change rule of the attribute of the communication area block, further extracting the time characteristics of the network flow of the communication area block, and finally obtaining a final prediction result through a full connection layer.
Description
Technical Field
The invention relates to the fields of electronics, communication and information engineering, in particular to a network flow prediction method based on a graph neural network.
Background
With the rapid development of communication technology, network traffic has a explosive growth trend, and a network traffic prediction technology is developed in response to the purpose of preventing network congestion and improving the utilization rate of network resources. The modeling and prediction of the network flow can know the change trend of the network flow in advance, and a reasonable and effective flow management strategy is formulated according to the predicted value so as to improve the network service quality and the user experience, so that the establishment of a high-precision network flow prediction model has important significance. In recent years, the ground network traffic prediction method is widely applied and paid attention to, and according to the research of scholars at home and abroad, the network traffic prediction method is mainly divided into the following two aspects: (1) a linear prediction model; the linear prediction model is mainly used for processing the short-term prediction problem of the network flow aiming at the short-term correlation characteristic of the network flow, and the related processing model mainly comprises the following steps: most of the traditional linear prediction models are infinite approximation to real network flow data by using a polynomial fitting function, and then the fitting effect is best through optimization of a large number of parameter settings. The ARMA model is simple in principle and calculation process, but cannot process non-stationary sequences. Compared with other network flow prediction methods, the ARIMA model has better effect of processing non-stationary sequences and higher prediction precision. (2) A non-linear predictive model. At present, nonlinear models applied to the aspect of network traffic prediction mainly include machine learning and deep learning related intelligent prediction models. The support vector machine is used as a machine learning algorithm, is mainly based on a statistical theory, and is widely applied to the field of classification prediction. The method has the advantages that the method can be applied to nonlinear separable situations, compared with other algorithms, the method does not need too many samples under the same problem complexity, and meanwhile, the kernel function is introduced to convert nonlinear separable samples into linearly separable high-dimensional space samples, so that better prediction accuracy can be obtained. The disadvantage is that convergence to local optima is achieved and the training data and the parameters of the training data have a great influence on the prediction result. To further address the prediction problem, deep learning approaches have been extensively studied for several years. Compared with machine learning, the deep learning can not only keep learning characteristics, but also ensure the relevance between the deep learning and each task, and can effectively process time series problems. A special model threshold recursion unit of a Recurrent Neural Network (RNN) can solve the long-term dependence capability of the recurrent neural network and the problems of gradient disappearance and gradient explosion in the conventional RNN training process, and is improved in prediction accuracy compared with a machine learning algorithm.
Although the existing network traffic prediction models have good prediction effects, the methods only consider time-series time correlation and ignore the correlation among real communication areas, namely the spatial characteristics of the network traffic, which can cause the higher-dimensional characteristics of the network traffic to be ignored during prediction, so the spatial characteristics of the network traffic need to be considered during prediction.
Disclosure of Invention
The technical problem to be solved by the invention is to overcome the defects of the prior art and provide a network flow prediction method based on a graph neural network, the method firstly learns the network topology structure of a communication area block by a graph convolutional neural network (GCN), extracts the spatial characteristics of the network flow of the communication area block, then uses data with the spatial characteristics as the input of a threshold recursion unit (GRU), learns the time change rule of the attribute of the communication area block, further extracts the time characteristics of the network flow of the communication area block, and finally obtains the final prediction result through a full connection layer, so that the time-space characteristics of the network flow can be better extracted, and better prediction accuracy can be obtained.
The invention adopts the following technical scheme for solving the technical problems:
the invention provides a network flow prediction method based on a graph neural network, which comprises the following steps:
step 1, acquiring network flow data;
inputting the adjacency matrix and the training set into a GCN model, extracting spatial features among the area blocks by the GCN model through frequency spectrum convolution, outputting a network traffic matrix with the spatial features to a first GRU layer by a second GCN layer, iteratively processing the network traffic matrix with the spatial features by the first GRU layer to an Mth GRU layer to extract time features of network traffic, and outputting a network traffic matrix with the time features by the Mth GRU layer; finally, the obtained network traffic matrix with time and space characteristics passes through 1 full connection layer to obtain a final prediction result;
and 6, predicting the network flow by using the GCN-GRU model trained, tested and evaluated in the steps 4 and 5.
As a further optimization scheme of the network flow prediction method based on the graph neural network, evaluation indexes comprise root mean square error, average absolute error, accuracy, decision coefficients and interpretable variance score.
As a further optimization scheme of the network flow prediction method based on the graph neural network, M is 100.
Compared with the prior art, the invention adopting the technical scheme has the following technical effects:
the experiment in the invention adopts real network flow data, and after multiple experiments, the accuracy and error analysis comparative analysis is carried out on the prediction result and other models, and the root mean square error of the GCN-GRU model is obtained under the prediction time step lengths of 10min and 20minRMSECompared with a single GRU model, the method reduces the prediction accuracy by 1.7 percent and 1.4 percent respectively, and improves the prediction accuracy by 1.9 percent and 1.1 percent respectively.
Drawings
Fig. 1 is a schematic flow chart of a network traffic prediction method based on a graph convolution neural network according to the present invention.
FIG. 2 is a schematic diagram of a model used in the present invention.
Fig. 3 is a graphical illustration of a trend graph of network traffic data used in the present invention.
FIG. 4 is an error diagram of the number of hidden units of different numbers according to the model of the present invention.
FIG. 5 is a graph of the error comparison of the inventive model to a single GRU model.
FIG. 6 is a graph of the predicted results of the present invention at a time scale of 10 min.
FIG. 7 is a graph of the predicted results of the present invention at a time scale of 20 min.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention will be described in detail with reference to the accompanying drawings and specific embodiments.
The network flow prediction model GCN-GRU based on the combination of the graph convolution neural network (GCN) and the threshold recursion unit (GRU) inherits the advantages of the graph convolution neural network and the single-cycle neural network, simultaneously makes up the defects of the traditional neural network model, captures the spatial characteristics of the network flow, and optimizes the accuracy of network flow prediction.
FIG. 1 is a schematic model flow diagram of a network traffic prediction model GCN-GRU based on the combination of a graph convolutional neural network (GCN) and a threshold recursive unit (GRU) in the invention. Firstly, network traffic data is acquired, and then areas of the acquired network traffic data are divided to form different area blocks. The data used by the invention is network traffic data of a certain mobile communication company in Milan, and the area is divided into grids with the size of M multiplied by N according to the longitude and latitude, wherein M = N =100, and the number of the grids is 10000. Then, network traffic information of different areas is counted, preprocessing is carried out, a network traffic input matrix is constructed, and an adjacency matrix reflecting correlation of different areas is established. And inputting the processed network traffic matrix and the adjacency matrix into a GCN-GRU model, then training the GCN-GRU model to learn and mine the space-time characteristics of the network traffic, and finally obtaining the predicted value of the network traffic.
Fig. 2 is a network traffic trend graph of a week of 9 selected areas according to the present invention. The 9 areas are abstracted into a network topology, and network traffic of each area has certain correlation. Moreover, it can be seen from the figure that the network traffic has a certain daily periodicity, and reaches a peak value every day according to a certain trend, and then decreases, wherein there is also a certain mutation. The non-linear trend and the network traffic use condition are in accordance with the actual condition, and the reliability and the practicability of the selected data set are explained.
FIG. 3 is a schematic diagram of a network traffic prediction model of a convolutional neural network. The GCN principle is that a filter is constructed in a Fourier domain, then the constructed filter is used for processing communication nodes in a communication topological graph and a first-order field of the nodes, so that the spatial characteristics among the nodes in the graph are obtained, and finally a GCN model is established by superposing a plurality of convolution layers. The model GCN uses 2 convolution layers to process the topological structure of the graph, and the calculation formula is as follows:
wherein:Xrepresenting a matrix of characteristics of traffic of the communication network,Awhich represents the adjacency matrix, is,which represents a pre-treatment step, is,a matrix having a self-connecting structure is represented,andrepresenting the weight matrices in the first and second convolutional layers,is a matrix of the degrees, and the degree matrix,,andrepresenting an activation function.
The calculation process of each parameter in the GCN-GRU model is as follows,the graph convolution process is represented and defined in equation 2. W and b represent the weight and bias of the model during training, respectively.
Overall, the GCN-GRU model can handle complex spatial and temporal dependencies. And (3) obtaining the time-space correlation characteristics of the network flow through the formulas (2) to (5), and then realizing the final prediction of the network flow according to the extracted time-space characteristics.
The process is concretely realized as follows:
(1) the front 806 data was used as training set and the back 202 data was used as test set, for a total of 1008 data.
(2) In order to eliminate the dimension influence among indexes, the indexes of the subsequent result analysis are in the same range, the data are normalized, the MinMaxScaler () function is used for normalizing the data to be (-1, 1), and the denormalization operation is carried out after the result is output, wherein the formula is as follows:
wherein,is a normalized standard value;is the maximum value in the experimental data;is the minimum value in the experimental data.
(3) And inputting the processed data into a GCN model, wherein the GCN can acquire the relation between nodes in a topological graph, namely the spatial characteristics of network traffic, 2 convolutional layers are used in the GCN model to process the topological structure of the graph, and an adjacency matrix is acquired by using the Pearson correlation principle. And then inputting the data with the spatial characteristics into a GRU unit to capture the time correlation of the time sequence, further extracting the time characteristics of the network flow, and finally obtaining a final prediction result by the obtained time-space correlation characteristics through a full connection layer.
In order to verify that different areas on the data set have spatial correlation, the invention selects data flow of 9 areas as experimental data, and calculates the Pearson correlation coefficient for the selected 9 areas. The correlation formula is as follows:
wherein,Xandrepresents a random variable, i.e. a region in the data set,andthe size of the grid of the representative region,representsXAndthe covariance of (a) of (b),andis the product of the respective standard deviations.
FIG. 4 is an error diagram of the number of hidden units in different models. The hyper-parameters of the GCN-GRU model mainly comprise: learning rate, batch, iteration number and hidden layer number. In this experiment, a learning rate of 0.001, a batch value of 64, and a number of iterations of 3000 were set. In order to ensure the prediction accuracy, the number of hidden units of the model needs to be reasonably set in the experiment, so that different numbers of hidden units need to be selected for carrying out a transverse comparison experiment, and then the most appropriate value is selected. We set the number of concealment layers to [8, 16, 32, 64, 100, 128], respectively, then perform model training, and finally perform comparison. As shown in fig. 4, the horizontal axis represents the number of hidden units and the vertical axis represents the variation of different metrics, showing the RMSE and MAE results for different hidden units. It can be seen that when the number is 100, the error is the smallest and the prediction result is the best. When the number of hidden units increases, the prediction accuracy increases first and then decreases. This is mainly because when the hidden unit is larger than a certain degree, the complexity of the model and the calculation difficulty are greatly increased, thereby reducing the prediction accuracy. Therefore, we set the number of hidden units to 100 in all experiments.
FIG. 5 is a graph of model versus single GRU model error. To verify whether the GCN-GRU model has the ability to characterize spatio-temporal features from network traffic data, the GCN-GRU model was compared to the GRU model. As shown in fig. 5, it can be clearly seen that the RMSE of the method GCN-GRU based on the spatio-temporal characteristics is lower than that of the model based on the single factor (GRU), which can indicate that the GCN-GRU model can obtain the spatial characteristics from the network traffic data, and optimize the problem of the time-dependent dependency reduction of the GRU model during the long-term prediction, compared with the GRU model only considering the time characteristics, the RMSE of the GCN-GRU model is reduced by about 1.7% and 1.4% in the predictions of 10min and 20min, and the prediction accuracy is improved by 1.9% and 1.1% respectively, which indicates that the GCN-GRU model can better capture the spatio-temporal correlations, and predict the network traffic from the multi-dimensional characteristics.
The experimental environment parameters and model parameters related to the network flow prediction model provided by the invention are as follows:
TABLE 1 Experimental hardware parameters
For the estimation of the prediction result of the network model, the following 5 evaluation indexes are adopted in the experiment:
(1) root Mean Square Error (RMSE)
(2) Mean Absolute Error (MAE)
(3) Precision (Accuracy)
(5) Explained Variance Score (var)
Table 2 compares the data of the conventional network traffic prediction method and the method proposed by the present invention under 5 evaluation indexes.
TABLE 2 evaluation index results of different models
Fig. 6 and 7 are prediction result graphs of the GCN-GRU model under the prediction spans of 10min and 20min, and it can be seen that the predicted value of the GCN-GRU model under the prediction span of 10min changes more smoothly relative to the actual value, and the fitting effect is better. The model provided by the invention is optimal in all indexes, integrates the advantages of GCN and GRU, and optimizes the long-term dependence problem of GRU in prediction by using the network flow spatial characteristics extracted by GCN in consideration of multiple characteristics of network flow, thereby improving the prediction accuracy.
Experiments show that a single model cannot well predict network traffic, the problem of characteristic weakening in single model prediction is solved from the perspective of multiple characteristics of the network traffic, and then prediction is performed. The prediction result can show that the network flow prediction model GCN-GRU based on the combination of the graph convolution neural network (GCN) and the threshold recursion unit (GRU) has better prediction effect, the RMSE of the GCN-GRU model is respectively reduced by about 1.7% and 1.4% in the prediction of 10min and 20min, and the prediction accuracy is respectively improved by 1.9% and 1.1%.
The above description is only for the specific embodiment of the present invention, but the scope of the present invention is not limited thereto, and any changes or substitutions that can be easily conceived by those skilled in the art within the technical scope of the present invention are included in the scope of the present invention.
Claims (3)
1. A network flow prediction method based on a graph neural network is characterized by comprising the following steps:
step 1, acquiring network flow data;
step 2, dividing the areas of the network traffic data obtained in the step 1 to form different area blocks, counting the network traffic of the different area blocks in different time periods, constructing a network traffic matrix, and constructing an adjacent matrix between the different area blocks according to the Pearson correlation coefficient; then preprocessing and normalizing the network traffic matrix, and finally dividing the normalized network traffic matrix into a training set and a test set;
step 3, constructing a deep learning GCN-GRU model combining a graph convolution neural network GCN and a gating recursion unit GRU, and initializing the GCN-GRU model; the GCN-GRU model comprises a graph convolution neural network GCN model, first to M gate control recursion unit GRU layers and 1 full connection layer, wherein the GCN model comprises a first GCN layer and a second GCN layer, the first GCN layer is connected with the second GCN layer, the second GCN layer is connected with the first GRU layer, the first to M gate control recursion unit GRU layers are sequentially connected, and the M gate control recursion unit GRU layer is connected with the 1 full connection layer; wherein, the value range of M is between 32 and 128;
step 4, inputting the adjacency matrix and the training set into the GCN-GRU model constructed in the step 3, and learning and mining the time characteristics and the space characteristics of the network flow to obtain a trained GCN-GRU model; wherein,
inputting the adjacency matrix and the training set into a GCN model, extracting spatial features among the area blocks by the GCN model through frequency spectrum convolution, outputting a network traffic matrix with the spatial features to a first GRU layer by a second GCN layer, iteratively processing the network traffic matrix with the spatial features by the first GRU layer to an Mth GRU layer to extract time features of network traffic, and outputting a network traffic matrix with the time features by the Mth GRU layer; finally, the obtained network traffic matrix with time and space characteristics passes through 1 full connection layer to obtain a final prediction result;
step 5, inputting the test set into the trained GCN-GRU model, evaluating the GCN-GRU model by utilizing the evaluation index, changing the value of M if the evaluation index of the GCN-GRU model does not accord with the preset evaluation index, and then continuously executing the steps 3 to 4 until the evaluation index of the trained GCN-GRU model meets the preset evaluation index;
and 6, predicting the network flow by using the GCN-GRU model trained, tested and evaluated in the steps 4 and 5.
2. The method of claim 1, wherein the evaluation metrics comprise root mean square error, mean absolute error, accuracy, decision coefficient, and interpretable variance score.
3. The method of claim 1, wherein M is 100.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202110027901.XA CN112350876A (en) | 2021-01-11 | 2021-01-11 | Network flow prediction method based on graph neural network |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202110027901.XA CN112350876A (en) | 2021-01-11 | 2021-01-11 | Network flow prediction method based on graph neural network |
Publications (1)
Publication Number | Publication Date |
---|---|
CN112350876A true CN112350876A (en) | 2021-02-09 |
Family
ID=74428201
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202110027901.XA Pending CN112350876A (en) | 2021-01-11 | 2021-01-11 | Network flow prediction method based on graph neural network |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN112350876A (en) |
Cited By (19)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN112906982A (en) * | 2021-03-22 | 2021-06-04 | 哈尔滨理工大学 | GNN-LSTM combination-based network flow prediction method |
CN113011501A (en) * | 2021-03-22 | 2021-06-22 | 广东海启星海洋科技有限公司 | Method and device for predicting typhoon water level based on graph convolution neural network |
CN113094860A (en) * | 2021-04-29 | 2021-07-09 | 北京邮电大学 | Industrial control network flow modeling method based on attention mechanism |
CN113112004A (en) * | 2021-04-21 | 2021-07-13 | 上海电力大学 | Time-space prediction method based on Hox process and graph neural network |
CN113393017A (en) * | 2021-05-07 | 2021-09-14 | 浙江警官职业学院 | Individual complex relationship-based supervision place safety early warning method and device |
CN113744889A (en) * | 2021-09-08 | 2021-12-03 | 平安科技(深圳)有限公司 | Infectious disease prediction method, system, device and storage medium based on neural network |
CN113743594A (en) * | 2021-09-30 | 2021-12-03 | 中国联合网络通信集团有限公司 | Network flow prediction model establishing method and device, electronic equipment and storage medium |
CN113852492A (en) * | 2021-09-01 | 2021-12-28 | 南京信息工程大学 | Network flow prediction method based on attention mechanism and graph convolution neural network |
CN114124734A (en) * | 2021-11-20 | 2022-03-01 | 四川大学 | Network traffic prediction method based on GCN-Transformer integration model |
CN114205251A (en) * | 2021-12-09 | 2022-03-18 | 西安电子科技大学 | Switch link resource prediction method based on space-time characteristics |
CN114239959A (en) * | 2021-12-15 | 2022-03-25 | 中国电信股份有限公司 | Hotspot prediction method and device, storage medium and electronic equipment |
CN114374655A (en) * | 2021-12-22 | 2022-04-19 | 中国电信股份有限公司 | Network flow characteristic extraction method, device, equipment and storage medium |
CN114501518A (en) * | 2022-04-01 | 2022-05-13 | 香港中文大学(深圳) | Flow prediction method, flow prediction device, flow prediction apparatus, flow prediction medium, and program product |
CN114826948A (en) * | 2022-04-26 | 2022-07-29 | 电子科技大学 | SDN network flow prediction method based on graph convolution network |
CN115359350A (en) * | 2022-07-28 | 2022-11-18 | 北京空间飞行器总体设计部 | Group target formation change identification method based on graph model |
CN115442253A (en) * | 2022-09-15 | 2022-12-06 | 西安电子科技大学 | Network flow prediction method using attention mechanism |
CN115564114A (en) * | 2022-10-10 | 2023-01-03 | 南京航空航天大学 | Short-term prediction method and system for airspace carbon emission based on graph neural network |
CN116094758A (en) * | 2022-12-02 | 2023-05-09 | 肇庆医学高等专科学校 | Large-scale network flow acquisition method and system |
CN116633803A (en) * | 2023-07-24 | 2023-08-22 | 四川众力佳华信息技术有限公司 | Novel network flow combined prediction model |
Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20180137412A1 (en) * | 2016-11-16 | 2018-05-17 | Cisco Technology, Inc. | Network traffic prediction using long short term memory neural networks |
CN109754605A (en) * | 2019-02-27 | 2019-05-14 | 中南大学 | A kind of traffic forecast method based on attention temporal diagram convolutional network |
-
2021
- 2021-01-11 CN CN202110027901.XA patent/CN112350876A/en active Pending
Patent Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20180137412A1 (en) * | 2016-11-16 | 2018-05-17 | Cisco Technology, Inc. | Network traffic prediction using long short term memory neural networks |
CN109754605A (en) * | 2019-02-27 | 2019-05-14 | 中南大学 | A kind of traffic forecast method based on attention temporal diagram convolutional network |
Non-Patent Citations (1)
Title |
---|
LI YANG等: ""A Noval Satellite Network Traffic Prediction Method Based on GCN-GRU"", 《2020 INTERNATIONAL CONFERENCE ON WIRELESS COMMUNICATIONS AND SIGNAL PROCESSING (WCSP)》 * |
Cited By (29)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN113011501B (en) * | 2021-03-22 | 2022-05-24 | 广东海启星海洋科技有限公司 | Method and device for predicting typhoon water level based on graph convolution neural network |
CN113011501A (en) * | 2021-03-22 | 2021-06-22 | 广东海启星海洋科技有限公司 | Method and device for predicting typhoon water level based on graph convolution neural network |
CN112906982A (en) * | 2021-03-22 | 2021-06-04 | 哈尔滨理工大学 | GNN-LSTM combination-based network flow prediction method |
CN113112004B (en) * | 2021-04-21 | 2022-12-06 | 上海电力大学 | Time-space prediction method based on Hox process and graph neural network |
CN113112004A (en) * | 2021-04-21 | 2021-07-13 | 上海电力大学 | Time-space prediction method based on Hox process and graph neural network |
CN113094860B (en) * | 2021-04-29 | 2023-09-01 | 北京邮电大学 | Industrial control network flow modeling method based on attention mechanism |
CN113094860A (en) * | 2021-04-29 | 2021-07-09 | 北京邮电大学 | Industrial control network flow modeling method based on attention mechanism |
CN113393017A (en) * | 2021-05-07 | 2021-09-14 | 浙江警官职业学院 | Individual complex relationship-based supervision place safety early warning method and device |
CN113393017B (en) * | 2021-05-07 | 2022-08-23 | 浙江警官职业学院 | Individual complex relationship-based supervision place safety early warning method and device |
CN113852492A (en) * | 2021-09-01 | 2021-12-28 | 南京信息工程大学 | Network flow prediction method based on attention mechanism and graph convolution neural network |
CN113744889A (en) * | 2021-09-08 | 2021-12-03 | 平安科技(深圳)有限公司 | Infectious disease prediction method, system, device and storage medium based on neural network |
CN113743594A (en) * | 2021-09-30 | 2021-12-03 | 中国联合网络通信集团有限公司 | Network flow prediction model establishing method and device, electronic equipment and storage medium |
CN113743594B (en) * | 2021-09-30 | 2023-12-05 | 中国联合网络通信集团有限公司 | Network traffic prediction model establishment method and device, electronic equipment and storage medium |
CN114124734B (en) * | 2021-11-20 | 2023-04-07 | 四川大学 | Network traffic prediction method based on GCN-Transformer integration model |
CN114124734A (en) * | 2021-11-20 | 2022-03-01 | 四川大学 | Network traffic prediction method based on GCN-Transformer integration model |
CN114205251B (en) * | 2021-12-09 | 2022-12-02 | 西安电子科技大学 | Switch link resource prediction method based on space-time characteristics |
CN114205251A (en) * | 2021-12-09 | 2022-03-18 | 西安电子科技大学 | Switch link resource prediction method based on space-time characteristics |
CN114239959A (en) * | 2021-12-15 | 2022-03-25 | 中国电信股份有限公司 | Hotspot prediction method and device, storage medium and electronic equipment |
CN114374655A (en) * | 2021-12-22 | 2022-04-19 | 中国电信股份有限公司 | Network flow characteristic extraction method, device, equipment and storage medium |
CN114501518A (en) * | 2022-04-01 | 2022-05-13 | 香港中文大学(深圳) | Flow prediction method, flow prediction device, flow prediction apparatus, flow prediction medium, and program product |
CN114826948A (en) * | 2022-04-26 | 2022-07-29 | 电子科技大学 | SDN network flow prediction method based on graph convolution network |
CN115359350A (en) * | 2022-07-28 | 2022-11-18 | 北京空间飞行器总体设计部 | Group target formation change identification method based on graph model |
CN115442253B (en) * | 2022-09-15 | 2023-07-18 | 西安电子科技大学 | Network flow prediction method utilizing attention mechanism |
CN115442253A (en) * | 2022-09-15 | 2022-12-06 | 西安电子科技大学 | Network flow prediction method using attention mechanism |
CN115564114A (en) * | 2022-10-10 | 2023-01-03 | 南京航空航天大学 | Short-term prediction method and system for airspace carbon emission based on graph neural network |
CN116094758A (en) * | 2022-12-02 | 2023-05-09 | 肇庆医学高等专科学校 | Large-scale network flow acquisition method and system |
CN116094758B (en) * | 2022-12-02 | 2023-07-21 | 肇庆医学高等专科学校 | Large-scale network flow acquisition method and system |
CN116633803A (en) * | 2023-07-24 | 2023-08-22 | 四川众力佳华信息技术有限公司 | Novel network flow combined prediction model |
CN116633803B (en) * | 2023-07-24 | 2023-10-20 | 四川众力佳华信息技术有限公司 | Novel network flow combined prediction model |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN112350876A (en) | Network flow prediction method based on graph neural network | |
CN112988723B (en) | Traffic data restoration method based on space self-attention force diagram convolution cyclic neural network | |
Wang et al. | A compound framework for wind speed forecasting based on comprehensive feature selection, quantile regression incorporated into convolutional simplified long short-term memory network and residual error correction | |
CN113852492A (en) | Network flow prediction method based on attention mechanism and graph convolution neural network | |
CN112906982A (en) | GNN-LSTM combination-based network flow prediction method | |
CN112085124A (en) | Complex network node classification method based on graph attention network | |
CN109635928A (en) | A kind of voltage sag reason recognition methods based on deep learning Model Fusion | |
Hou et al. | Review of load forecasting based on artificial intelligence methodologies, models, and challenges | |
CN114422381A (en) | Communication network flow prediction method, system, storage medium and computer equipment | |
CN111832825B (en) | Wind power prediction method and system integrating long-term memory network and extreme learning machine | |
CN112187554B (en) | Operation and maintenance system fault positioning method and system based on Monte Carlo tree search | |
CN114428803B (en) | Air compression station operation optimization method, system, storage medium and terminal | |
CN112396234A (en) | User side load probability prediction method based on time domain convolutional neural network | |
CN113780684A (en) | Intelligent building user energy consumption behavior prediction method based on LSTM neural network | |
CN109886464A (en) | The low information loss short-term wind speed forecasting method of feature set is generated based on optimization singular value decomposition | |
CN115828990A (en) | Time-space diagram node attribute prediction method for fused adaptive graph diffusion convolution network | |
CN109583588A (en) | A kind of short-term wind speed forecasting method and system | |
CN115694985A (en) | TMB-based hybrid network traffic attack prediction method | |
CN114694379B (en) | Traffic flow prediction method and system based on self-adaptive dynamic graph convolution | |
CN110516792A (en) | Non-stable time series forecasting method based on wavelet decomposition and shallow-layer neural network | |
CN112562312B (en) | GraphSAGE traffic network data prediction method based on fusion features | |
CN117354172A (en) | Network traffic prediction method and system | |
CN117332227A (en) | Environment variable prediction method and equipment for multi-time space scale attention mechanism | |
CN116578858A (en) | Air compressor fault prediction and health degree evaluation method and system based on graphic neural network | |
CN113240904B (en) | Traffic flow prediction method based on feature fusion |
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 | ||
RJ01 | Rejection of invention patent application after publication | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20210209 |