Design and Deployment of a Network Anomaly Detection System.

Abstract

Computer security plays an important role in everybody’s life, especially when it comes to connected applications and services. Therefore, securing the network connectivity from being compromised by untrusted parties and malicious usage has a great significance in maintaining their efficient functioning. Considered as one of the primary defense lines of network security and expected to be adapting to the dynamically changing threat patterns, Intrusion detection systems have been developed with different techniques by researchers from various disciplines like mathematics, machine learning, and data mining in order to achieve a good immunity against attacks and reliable detection of anomalies. In this thesis, we propose an Artificial Neural Network-based anomaly detection application. In addition to the benign class, the proposed method deals with fourteen attack classes and unlike most methods, our work has been trained using an up to date dataset. As a result, the proposed system is capable of achieving an accuracy of 99% and a false positive rate of 1%.