Neural networks have been used in situation awareness and anomaly detection. However, most existing applications are based on fully connected neural network that is not suitable enough for processing raw situation data generated by networks. Generally, situation data can be organized and comprehended by time series because network attacks and states change over time. It\'s a gradual process and there are logical associations between different stages. Considering above factors, we propose a neural network structure based on LSTM to establish time correlation between situation data. Improve LSTM with cross entropy function, rectified linear unit and appropriate layer stacking. Experiments prove that the improved LSTM neural network has better discriminative performance for current network security situation. The LSTM-based method enables an efficient comprehension and evaluation of network security situation.
Software And Hardware
• Hardware: Processor: i3 ,i5 RAM: 4GB Hard disk: 16 GB • Software: operating System : Windws2000/XP/7/8/10 Anaconda,jupyter,spyder,flask Frontend :-python Backend:- MYSQL