Enhancing the Classification Accuracy of Intrusion Detection system using Auto-encoder Algorithm

Papaiah, Jonah Abhijit (2022) Enhancing the Classification Accuracy of Intrusion Detection system using Auto-encoder Algorithm. Masters thesis, Dublin, National College of Ireland.

Preview	PDF (Master of Science) Download (1MB) | Preview
Preview	PDF (Configuration manual) Download (1MB) | Preview

Abstract

The development of the internet and commerce has compelled governments and other organizations all over the world to promote new technologies and employ more complex procedures and contemporary networks. For instance, several business organizations allow administrative access to the system via the intranet and actively advocate web services for their partners. However, using such networks brings about a number of security risks that enter the system through connections and these malicious data packets impact the security and integrity of the system as well as the confidential information. Governments and other organizations are being forced to push new technology, use more intricate processes, and utilize cutting-edge networks as a result of the growth of the internet and commerce. For instance, a number of commercial organizations permit partners to use online services and actively promote administrative access to the system via the intranet. But using such networks comes with a variety of security dangers that enter the system through connections and these malicious data packets affect the security and integrity of the system as well as the confidential data. The convincible outrun can be machine learning-based algorithms that are capable of automatically identifying the system invasions such as FTP/ SSH brute force intrusions which have resulted in satisfying outcomes. But nowadays it is observed that generally, deep learning algorithms overcome machine learning algorithms if the data size is large enough therefore four advanced deep learning algorithms are applied in this research work which are CNNs, LSTM, Conv-LSTM, and Auto-Encoders. Since the data source can be different in real-world applications therefore two datasets are accounted for this task. All algorithms are implemented and tested on test data and assessed using various parameters since being developed utilizing KDD CUP-99 IDS dataset and CSE-CIC-IDS2018 datasets. The Auto-Encoder approach, which has been proven to be superior compared to these other algorithms after assessment of parameters like Accuracy, Precision, Recall, and Validation Loss, can be employed to categorize invasion into a variety of forms of obtrusive actions in actual employment.

Item Type:	Thesis (Masters)
Supervisors:	Name Email Nayak, Prashanth UNSPECIFIED
Subjects:	Q Science > QA Mathematics > Electronic computers. Computer science T Technology > T Technology (General) > Information Technology > Electronic computers. Computer science Q Science > QA Mathematics > Algebra > Algorithms > Computer algorithms Q Science > QA Mathematics > Computer software > Computer Security T Technology > T Technology (General) > Information Technology > Computer software > Computer Security Q Science > Q Science (General) > Self-organizing systems. Conscious automata > Machine learning
Divisions:	School of Computing > Master of Science in Data Analytics
Depositing User:	Tamara Malone
Date Deposited:	23 May 2023 16:09
Last Modified:	23 May 2023 16:09
URI:	https://norma.ncirl.ie/id/eprint/6630

Actions (login required)

View Item