Network traffic matrix prediction method based on self-attention mechanism

Abstract

The invention discloses a network traffic matrix prediction method based on a self-attention mechanism, which relates to the technical field of network traffic prediction and comprises the step of scaling network traffic data to a specified range. And constructing a shortest path tree of all nodes in the target network, coding according to all the shortest path trees to obtain a spatial code, and adding the spatial code to the network flow data. And coding the timestamps of the network flow data according to different granularities to obtain time codes, and adding the time codes to the network flow data. And constructing a traffic prediction model based on a self-attention mechanism by using a trainable position code and the network traffic data, and performing traffic prediction. Through utilization of time and space correlation information of network traffic, a self-attention mechanism can fully mine potential features of the traffic, and the accuracy of long-time prediction of a network traffic matrix is improved.

Description

technical field [0001] The present application relates to the technical field of network traffic prediction, in particular to a network traffic matrix prediction method based on a self-attention mechanism. Background technique [0002] The Internet and its applications have become the primary means of communication for people to carry out their daily activities. The upgrade of network technology has expanded the influence of the Internet on various application fields. The expansion of wireless network and mobile Internet and the improvement of network heterogeneity and complexity have put forward higher requirements for network service quality, and also brought more challenges to network management. The intelligent self-management of the network is an effective solution to the above problems, and its key technology includes network traffic prediction technology. Using historical traffic time series data to predict future traffic changes and reallocate network resources and...

AI Technical Summary

Problems solved by technology

[0007] However, the above methods have obvious shortcomings when predicting complex long-term traffic. For example, the prediction accuracy of traditional machine learning prediction models decreases, and for example, most deep learning models will fail when the prediction time is too long.

Images