Intrusion Detection and Prevention System Using Deep Learning

As technology is developing day by day and the number of networked digital devices is increasing manifold, the number of intruders and the type of attacks are also increasing. So there arises a need to detect and prevent these attacks. Hence this work is focused to implement an Intrusion detection and prevention system using Deep Learning that can immediately detect the attacks such as DOS, Probe, R2L and U2R and prevent the same. The intrusion when arises is detected using a Deep Learning model called Multi-Layer Perceptron trained by the dataset kddcup99 with high accuracy. Appropriate data from network is captured and it is stored as a csv file and is fed to the implemented Deep learning model to predict the attack in a real time manner, thus detection is achieved. In second phase, the intrusion is prevented using a script that runs in the background. The script is developed to perform the prevention phase by taking appropriate decision on the different prevention function to be performed for different types of attacks. The decision can be made by using the data from the classification part achieved through the Multi-Layer Perceptron model. In this paper both the separate Intrusion Detection System and the Intrusion Prevention system are combined as a single system to achieve the aim of intrusion detection and prevention tasks in a faster and efficient manner.