Cyber Security Threats Detection using Deep Learning

The fast growth of computer networks and the proliferation of applications running on them, network security has emerged as a critical concern. The increased interconnectedness of systems has made them vulnerable to security threats, which can have a detrimental impact on the economy. As a result, the accurate and real-time detection of vulnerabilities in network systems has become paramount. This thesis presents a comprehensive study on cyber security threat detection using deep learning techniques, with a focus on utilizing sequential deep learning models to discern attacks from normal network traffic. The study begins by addressing the importance of network security. It highlights the necessity for accurate and efficient detection mechanisms to safeguard network infrastructure. To evaluate the efficacy of deep learning models. The renowned KDD dataset is employed. This dataset, derived from the 1999 DARPA Intrusion Detection Evaluation Program, encompasses various types of attacks and normal network traffic. The dataset's wide coverage enables a comprehensive analysis of network threats and the development of robust detection models. The thesis presents a systematic methodology for processing the KDD dataset, including preprocessing techniques to ensure data quality and model performance. Various data preprocessing steps are undertaken.