The prediction of network security situation based on deep learning method

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.