Neural Nets to Detect Abnormal Traffic in Communication Networks
Ilmensky Mikhail, Kireev S. Kh, A.I. Klimenko, Elena Georgievna Balenko · 2019 International Multi-Conference on Industrial Engineering and Modern Technologies (FarEastCon) · 2019
The paper describes using a multilayer perceptron to detect abnormal traffic in communication networks; the perceptron is based on Adam, a first-order algorithm for gradient optimization of stochastic objective functions. The traffic it detects pertains to four attack types: Probe, DoS, R2L, U2R. The applied ANN model shows 92% cross-validation accuracy on NSL-KDD dataset. Assumingly, an optimal algorithm for weight optimization and attribute selection will help improve the accuracy.