Wireless service traffic prediction method and device based on self-attention convolutional network, and medium

A technology for wireless business and traffic prediction, applied in wireless communication, data exchange network, neural learning methods, etc., can solve the problems of difficult parallel implementation of prediction algorithms and low training efficiency, so as to improve traffic prediction performance, improve training efficiency, The effect of reasonable resource scheduling and management

Active Publication Date: 2021-06-04
SHANDONG UNIV
View PDF8 Cites 6 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0006] The present invention proposes a wireless service traffic prediction method based on self-attention convolution network, which is used to solve the problem that the deep learning method based on RNN structure cannot capture the complex correlation of traffic data in a long-term sequence, and the prediction algorithm is difficult to parallelize implementation, the problem of low training efficiency

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Wireless service traffic prediction method and device based on self-attention convolutional network, and medium
  • Wireless service traffic prediction method and device based on self-attention convolutional network, and medium
  • Wireless service traffic prediction method and device based on self-attention convolutional network, and medium

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0070] A wireless service traffic prediction method based on a self-attention convolutional network, which refers to: preprocessing the original wireless service traffic to be predicted and inputting it into a trained traffic prediction model to obtain predicted traffic data, the traffic data refers to traffic value;

[0071] The process of preprocessing the original wireless service traffic to be predicted includes:

[0072] Taking hours as the time granularity unit, using sliding time windows to divide the original wireless service traffic, so that the time span of each group of traffic data is T hours; The regional flow data are stored in the database in the form of a data matrix, and each set of processed flow data D′={D 1 ',D 2 ',...,D t '...,D T ′}, where the matrix H represents the number of grid rows, W represents the number of grid columns, the center of the grid is taken as the origin, and the matrix D t element d in ' t (h,w)′ Indicates the flow value of th...

Embodiment 2

[0075] According to a method for predicting wireless service traffic based on a self-attention convolutional network described in Embodiment 1, the difference is that:

[0076] The traffic prediction model includes temporal encoding network, self-attention machine convolutional network, feature embedding network and convolutional residual network; such as figure 1 As shown, the time coding network extracts the features in the timestamp information, and fuses the obtained features with the traffic data with a certain time span; the self-attention machine convolution network performs correlation analysis and feature analysis on the wireless traffic data at different time nodes. Representation; the feature embedding network extracts other relevant features; the convolutional residual network fuses the output of the self-attention machine convolution network with the output representation of the feature embedding network to obtain the predicted traffic Y' for the next period.

Embodiment 3

[0078] According to a method for predicting wireless service traffic based on a self-attention convolutional network described in Embodiment 1, the difference is that:

[0079] In order to improve the overall performance of wireless service traffic prediction, the training process of the traffic prediction model is as follows:

[0080] (1) Using hours as the time granularity unit, the original wireless service traffic is divided by using the sliding time window, so that the time span of each group of traffic data is T hours;

[0081] In the spatial dimension, different regions are spliced ​​into a grid, so that the flow data of different regions in the same period of time are stored in the database in the form of a data matrix, and each group of processed flow data D={D 1 ,D 2 ,...,D t ...,D T}, where the matrix H represents the number of grid rows, W represents the number of grid columns, the center of the grid is taken as the origin, and the matrix D t element d in t ...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

PUM

No PUM Login to view more

Abstract

The invention relates to a wireless service traffic prediction method and device based on a self-attention convolutional network, and a medium, and the method comprises the steps: carrying out the preprocessing of a to-be-predicted original wireless service traffic, inputting the to-be-predicted original wireless service traffic into a trained traffic prediction model, and obtaining predicted traffic data; and utilizing the self-attention convolutional network to learn and fuse the historical data and time characteristics of the wireless service traffic, so that the dynamic characteristics of the traffic data under a large time scale can be effectively extracted. According to the algorithm, the accuracy of traffic prediction can be effectively improved, parallel processing can be carried out, the training efficiency of the neural network is improved, and the overall performance of the traffic prediction algorithm is further improved.

Description

technical field [0001] The invention belongs to the technical field of communication network and artificial intelligence, and relates to a wireless service flow prediction method, equipment and medium based on a self-attention mechanism, which can be used for network management and planning in a communication system. Background technique [0002] Efficient and accurate wireless service traffic prediction is of great significance for realizing communication network automation and intelligent resource allocation. However, the communication behavior of end users is highly dynamic, especially in the long-term scale, the complexity and time-varying nature of user communication behavior will increase, which increases the difficulty of wireless service traffic prediction. [0003] The wireless traffic forecasting problem can be modeled as a time series forecasting problem. Traditional statistical learning algorithms cannot model the high dynamic characteristics of traffic data in ...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

Application Information

Patent Timeline
no application Login to view more
Patent Type & Authority Applications(China)
IPC IPC(8): H04L12/24H04W24/06G06N3/04G06N3/08
CPCH04L41/147H04W24/06G06N3/08G06N3/045
Inventor 张海霞沈文鑫郭帅帅袁东风
Owner SHANDONG UNIV
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap