A Hybrid Approach of ANN-GWO Technique for Intrusion Detection

As informed individuals, while keeping ourselves updated with the whereabouts of the world, we often come across news articles with bold headlines outlining the various cyberattacks happening across the world. In this paper, we've attempted to build our own intrusion detection system (IDS) which we propose as a viable solution for detecting malicious entities in a network. Artificial Neural Networks (ANN) use backpropagation to update their weights which can get stuck in a local minima rather than a global one. This can lead to the weights and biases not reaching the optimal values. We have proposed and built a hybrid model of ANN along with the grey wolf optimization algorithm (GWO), to combine the technological benefits of these two state-of-the-art algorithmic techniques. We have employed the MIT Darpa 1998 intrusion detection dataset in our study, and used four metrics, namely, precision, accuracy, recall, and F1 score to evaluate the performance of our model.