Wireless Network Traffic Prediction Method Based on Multi-Graph Convolution

A traffic forecasting and wireless network technology, applied in wireless communication, neural learning methods, biological neural network models, etc., can solve problems such as poor accuracy and efficiency, and achieve reasonable modeling, high prediction accuracy, and good prediction results Effect

CN112911626BActive Publication Date: 2022-05-31FUZHOU UNIV
1 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Publication Date
2022-05-31

Smart Images

  • Figure 1
    Figure 1
  • Figure 2
    Figure 2
  • Figure 3
    Figure 3
Patent Text Reader

Abstract

The present invention proposes a wireless network traffic prediction method based on multi-graph convolution. First, spatially construct adjacent graphs, functional similarity graphs, and spatial traffic correlation graphs, and use graph convolutional networks to extract spatial features; secondly, Construct recent, daily and weekly related flows in the time domain, input the three kinds of time domain flows into the long-term and short-term memory network, and finally add the attention mechanism to judge the importance of the flow at different times. The root mean square, mean absolute error and coefficient of determination are introduced to evaluate the predictive performance of the proposed model. The method of the invention has the characteristics of reasonable modeling, high prediction accuracy and the like.
Need to check novelty before this filing date? Find Prior Art

Description

Wireless network traffic prediction method based on multi-graph convolution technical field The invention belongs to the technical field of wireless communication, graph theory and deep learning, particularly graph convolutional network, long-term and short-term memory memory network, attention mechanism, etc., especially a multi-graph convolution-based wireless network traffic prediction method. Background technique

[0002] With the advent of the 5G era, great changes have taken place in wireless networks, especially a large number of wireless With the dramatic rise in line cellular data, accurate wireless network traffic forecasting is critical for operators; With dense base station distribution, it is an unprecedented challenge to achieve more efficient traffic prediction. To achieve traffic forecasting, the first It is necessary to know the spatiotemporal characteristics of traffic and the distribution of related data, and make reasonable and accurate predictions...

Examples

Embodiment Construction

[0078] 1) The historical flow data used in this embodiment is Milan every ten minutes from 2013.01.11 to 2014.01.01

[0080]

[0082]

[0084]

[0087] Adjacent graph G

[0088]

[0089] Among them, || · || represents the second norm, ||v

[0091]

[0092] represents the inner product, ||·|| represents the two-norm, the interest point matrix P of each grid area, the moment of social activity quantity

[0094]

[0097]

[0099]

[0102]

[0103] T

[0105] 1) Graph Convolutional Network: As shown in FIG. 2, this embodiment uses a graph convolutional network to discover the spatial correlation of traffic. root

[0106] g

[0107]*

[0108]

[0110] Because the Chebyshev polynomial is: T

[0111]

[0113] 2) Long short-term memory network, as shown in Figure 2. This embodiment uses a two-layer long short-term memory network, X

[0114]

[0115] The input gate updates the cell state, first passing the information of the hidden state of the previous layer and the information of the cu...