Title: The prediction of network security situation based on deep learning method

Authors: Zhixing Lin; Jian Yu; Shunfa Liu

Addresses: College of Information Engineering, Sanming University, Fujian Sanming 365004, China ' College of Information Engineering, Sanming University, Fujian Sanming 365004, China ' College of Information Engineering, Sanming University, Fujian Sanming 365004, China

Abstract: Network security situational awareness is one of the important issues in the research of network space security technology. In this paper, deep learning technology is applied to analyse and learn network data, generate counter network by classification for sample amplification, use sparse noise reduction autoencoder for feature selection, and then use LSTM for deep learning model of security situation prediction. After the experiment proved that the proposed model based on sparse noise reduction is not balanced since the encoder-LSTM network security situation prediction model can solve various level attacks against a small number, using the model prediction results accurately in predicting regional security situation has the advantage for a longer time. In order to solve the above problems, the network security management becomes passive to active, adapting measures in advance.

Keywords: computer security; network security situation prediction; deep learning; auto-encoder; generating adversarial networks.

DOI: 10.1504/IJICS.2021.116941

International Journal of Information and Computer Security, 2021 Vol.15 No.4, pp.386 - 399

Received: 25 May 2020
Accepted: 23 Oct 2020

Published online: 28 Jul 2021 *

Full-text access for editors Access for subscribers Purchase this article Comment on this article