International Journal of Computer Science Issues

More than a traditional journal...

As attacks became more complicated, the traditional and contemporary methods such as firewalls were not successful and suitable in exact diagnosis. This caused Intrusion detection system (IDS) to finally the strictly centralized role in network security. first is misuse and second is abnormal detection. Misuse detection compares data to well-known attack signature so it cannot diagnose unknown attacks. Abnormal detection has better performance to detect new attacks by modeling. In most cases, Attacks has been centralized into four groups: DoS, Probe, U2R, and R2L. There are many approaches have been used to identify attacks in Intrusion detection system. One of them is artificial neural network who called (ANN). This paper strictly centralized approach to implement a hybrid Artificial Neural Network in IDS based on RBF. This paper investigates the effectiveness we shall explor our results by compared to (SVM) .