Network anomaly detection using deep learning techniques

Convolutional neural networks (CNNs) are the specific architecture of feed-forward artificial networks. It is de-facto standard for various operations in machine learning and computer vision. To transform this performance towards task network anomaly detection cyber-security, study proposes a model using one-dimensional CNN architecture. The authors' approach divides traffic data into transmission control protocol (TCP), user datagram (UDP), OTHER categories first phase, then each category treated independently. Before training model, feature selection performed Chi-square technique, then, over-sampling conducted synthetic minority technique to tackle class imbalance problem. method yields weighted average f-score 0.85, 0.97, 0.86, 0.78 TCP, UDP, OTHER, ALL categories, respectively. tested on UNSW-NB15 dataset.