Palakkattu East Madom Ramadas, Anusha (2024) Dynamic intrusion detection system for improved cloud security. Masters thesis, Dublin, National College of Ireland.
Preview |
PDF (Master of Science)
Download (1MB) | Preview |
Preview |
PDF (Configuration Manual)
Download (2MB) | Preview |
Abstract
The adoption of cloud computing is accelerating and securing the cloud environments against advancing cyber threats has become a necessity. Traditional intrusion detection systems (IDS) lack the capability of real time detection and have delayed response time. A machine learning (ML) based IDS which detect unknown attacks can help organisations to identify evolving cyber-attacks. This work investigates the use of extreme learning machine (ELM) algorithm in enhancing cloud-based IDS. Dataset leveraged for this work is CSE-CIS-CID2018. The performance of ELM model was evaluated and compared with other ML models like random forest (RF), decision tree (DT), Naive Bayes (NB), Artificial Neural Networks (ANN), and Deep Neural Networks (DNN). The results revealed that ELM model achieved an accuracy of 96.42%, whereas RF achieved an accuracy of 97.12%. ELM consumed less training time but consumed more time to predict compared to other ML models. The key findings from this project is that ML based IDS can enhance cloud security and while ELM may be efficient in certain scenarios, RF can be useful in another set of scenarios.
Actions (login required)
![]() |
View Item |