Network traffic prediction method of LSTM model based on attention mechanism

Abstract

The invention provides a network traffic prediction method for an LSTM model based on an attention mechanism, and the method comprises the following steps of: 1, data preprocessing: carrying out the standardization of network traffic data, and dividing the network traffic data into training data and test data; the method comprises the following steps of: step 1, constructing an attention mechanism-based LSTM model, step 2, constructing a model, and constructing an attention mechanism-based LSTM model, step 3, performing model training, inputting training data into the attention mechanism-basedLSTM model, and performing iterative training based on an Adam optimization algorithm to obtain a trained model; and 4, predicting the network traffic. According to the network traffic prediction method, when the output state of the current moment is calculated, the hidden states of the previous multiple moments are comprehensively considered, so that the generation probability of each item in the output sequence is influenced by the hidden states of the input multiple historical sequences, and the traffic prediction is more accurate.

Description

technical field [0001] The invention belongs to the technical field of network traffic forecasting, and in particular relates to a network traffic forecasting method based on an attention mechanism-based LSTM model. Background technique [0002] With the rapid development of Internet technology, electronic products such as mobile phones and tablet computers have gradually penetrated into people's lives, and various network applications are more widely used by everyone. At the same time, the scale of network data is also growing, which puts forward higher requirements for the security, efficiency, and stability of network equipment. How to do a good job in network planning and resource allocation has become a difficult problem for network administrators. Network traffic forecasting can predict the network traffic value in the future based on historical network traffic data, which can effectively help network administrators deal with network congestion, properly configure net...

AI Technical Summary

Problems solved by technology

Unable to capture the importance of different historical time series for the current time series output

[0006] 2. When the length of the historical sequence is long, the earlier processed historical sequence in LSTM is more likely to be forgotten, resulting in information loss and affecting model performance

Images