DNN BASED INTELLIGENT IDS FOR ANOMALY DETECTION

In today's world there is an increased amount of network attacks such as DoS (Denial of Service), R2L (Remote to Local) attack, U2R (user to root) and probe attack. These network attacks lead to illegal usage of user accounts, stealing software, running code to corrupt systems, perform actions that prevent authorized users from accessing these resources and using data for financial gain etc. To overcome these threats Intrusion Detection Systems were adopted. The earlier Intrusion Detection systems used machine learning algorithms such as Random Forest and Decision Trees require much computational resources and higher time complexity. To overcome these hurdles, a DNN based intrusion detection system is proposed. This improves speed, accuracy and stability of neural networks. These IDS will not only detect the anomalies but also results in higher accuracy compared to existing systems.