Abstract
The increasing reliance on technology for various tasks has led to a surge in computational demands, thereby significantly boosting the use of computer networks in recent decades. This increased need for computation and storage presents a lucrative business opportunity for many companies; however, it also attracts considerable attention from cyber attackers. To counter these threats, numerous researchers have developed various models aimed at detecting and preventing attacks. This paper proposes a novel intrusion detection model that operates in two phases: the first phase involves creating a feature ontology to train a LSTM model, while the second phase focuses on testing the trained LSTM model. For feature selection, the proposed model utilizes an Artificial Immune System-based genetic algorithm, which effectively identifies a robust set of features for classifying Nnetwork session types. The experiments were conducted using a real Nnetwork dataset, and the proposed model demonstrated the capability to detect multiple types of attacks within normal sessions. The results indicate that the proposed model improves accuracy and other performance metrics compared to existing models.
Published Version
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call