Two-Stage Deep Learning Framework For NIDS

Deep learning algorithms and frameworks have revolutionized predictive analysis over the past decade. These powerful techniques can be leveraged in the domain of Intrusion Detection to classify and predict cyber-attacks with minimal overhead. The dynamic nature of the problem along with the arise of new network attacks, make this problem highly intricate. In this paper, we aim to study the performance of a unique multi-stage deep learning models on the CICIDS-17 dataset. The dataset is a comprehensive dataset with an amalgam of real, recent and contemporary attacks which also correspond to the footprint of attacks found and encountered in the past five years. We propose a model of an LSTMAutoencoder followed by a classical ANN to classify the attacks in flow-based traffic data.