Optimization of BPN parameters using PSO for intrusion detection in cloud environment

The usage of internet is getting increased in all aspect, like from building various models that are fully connect with internet to the usage of digital media for 24/7. As this is rising in a way, on the other side, the concern about 'data security' is raising, and everybody's information needs to be protected from any other attacks or can say there should not be no data leakage. So, for detecting these attacks, intrusion detection system is placed. But placing this traditional intrusion detection system (IDS) will increase the concerns of security even more, so to amp the process integration of latest technology such deep learning comes in action. Thereby, this paper proposes an IDS using back propagation network where intrusions are identified based on the system calls that we collected in the dataset KDD cup 99. Also, to increase the optimization of neural network, we used particle swarm optimization thereby increase the accurate detection of the system saying if that is normal or abnormal in behavior. We evaluate our proposed model with other methods such as ANFIS, F-GNP and FCM in which the proposed model gives 96.5% accurate detection.