Network flow prediction method based on LSTM

Abstract

The invention discloses a network flow prediction method based on LSTM, and the method comprises the steps: obtaining a burst pulse string according to a flow signal, enabling the burst pulse string to be an impact factor signal for simulating an emergency, and inputting the flow signal and the burst pulse string into a network flow parallel LSTM predictor for flow prediction, wherein the networkflow parallel LSTM predictor comprises two LSTM predictors, coefficients of all layers of neural networks in the two LSTM predictors are the same, internal state information can be exchanged, one predictor is a main predictor, the other predictor is used for detecting a burst moment, internal state exchange is carried out between the two predictors, and the main predictor carries out multivariableprediction by utilizing the information obtained by the burst predictor, so that the main predictor can adapt to the change of a flow mode caused by burst flow, and the prediction accuracy is improved. Simulation experiments show that the network flow parallel LSTM predictor can adapt to flow changes of different intensities, and compared with a traditional single-variable LSTM predictor, the prediction accuracy of the network flow parallel LSTM predictor is improved by about 10%.

Description

Technical field [0001] The present invention relates to the field of network technology, in particular to a method for predicting network traffic based on LSTM. Background technique [0002] As the scale of the Internet continues to expand, there are more and more types of network traffic data and services, and the contradiction between supply and demand between network resources and network demand is becoming increasingly acute. Network traffic prediction is helpful to analyze network security, manage the network scientifically and prevent improper network behavior. Therefore, the research and implementation of network traffic prediction is of great significance. The network traffic prediction method based on artificial intelligence neural network has the characteristics of nonlinearity and self-adaptation, and has high prediction accuracy. [0003] In recent years, the use of deep neural networks to predict time series has become an important research direction. LSTM (LongShort...

AI Technical Summary

Problems solved by technology

[0005] The purpose of the present invention is to overcome the problem in the prior art that the prediction error of the burst time and the subsequent traffic volume is relatively large due to the burst traffic flow, and to provide a network traffic prediction method based on LSTM to improve the prediction of traffic flow. Prediction of Bursty Traffic Flow

Images