Automatic Intrusion Detection System Using Deep Re-Enforcement Learning With Q-network Algorithm (DQN)
Year: 2023
Type: dissertation
Abstract: IDS is an important part of the cybersecurity and IT space. Network intrusions are most times very hard to detect and group into their right network patterns. Due to the complexity of network intrusion detection, new age machine learning models that could eliminate some problems that an IDS places on a machine learning model e.g. poorly constructed, non-uniform datasets have to be discovered. Reinforcement learning has been applied to varying successful degrees in automation, gaming etc., and this report shows that it can be successfully applied to the field of network intrusion detection as well. Our proposed model uses the algorithm of a Q-network to operate in a reinforcement learning environment. The approach will be evaluated using a reward against episodes learning curve and will be further compared to more traditional machine learning models that use accuracy, recall and precision as evaluation metrics.
Source: NORMA
Author: Nwokedi, Ugochukwu Ikenna
Language: English
Subfield: Computer Networks and Communications
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W7017372835
