Cybersecurity Fortification through Machine Learning: Predictive Models for Malware Detection in Network Environments

Rajesh, Hudson Paul (2023) Cybersecurity Fortification through Machine Learning: Predictive Models for Malware Detection in Network Environments. Masters thesis, Dublin, National College of Ireland.

Preview	PDF (Master of Science) Download (3MB) | Preview
Preview	PDF (Configuration manual) Download (1MB) | Preview

Abstract

With a particular emphasis on virus detection in network contexts, this research project explores the field of cybersecurity. By using a multimodal approach, we apply well-known machine learning techniques, such as Recurrent Neural Networks (RNN), Artificial Neural Networks (ANN), and Convolutional Neural Networks (CNN), to build resilient models for the detection of harmful activity. The research employs well-known datasets for training and assessing the models’ effectiveness, such as those from the Microsoft malware prediction repository. To ensure the effectiveness of our models, we leverage established datasets, including data from the Microsoft Malware Prediction Database. These datasets serve as a valuable resource for training and evaluating the performance of machine learning models and provide a variety of representative malicious patterns for comprehensive analysis. Our research on the application of RNN, ANN and CNN in malware detection aims to improve the accuracy and effectiveness of cyber security measures. By leveraging the power of these machine learning structures, we aim to strengthen network security, creating proactive defences against cyber threats.

The goal of this research project is to strengthen cybersecurity by employing predictive models to detect malware in network environments. The study carefully uses downsampling methods and investigates how well Convolutional Neural Networks (CNNs) operate in conjunction with conventional machine learning models. The main objective of the three investigations, which involve extensive feature engineering and encoding methodologies, is to improve spatial understanding for more precise virus identification.

It is discovered that the downsampling technique, which reduces the dataset to 100,000 rows, effectively manages computer resources while posing questions about generalisation to a larger dataset. The use of CNNs, particularly in the most recent experiment, provides encouraging new information about the possible benefits of spatial dependency capture in malware detection.

Item Type:	Thesis (Masters)
Supervisors:	Name Email Syed, Muslim Jameel UNSPECIFIED
Uncontrolled Keywords:	Cybersecurity; Predictive Models; Malware Detection; Downsampling; Convolutional Neural Networks (CNNs); Feature Engineering; Network Security
Subjects:	Q Science > QA Mathematics > Electronic computers. Computer science T Technology > T Technology (General) > Information Technology > Electronic computers. Computer science Q Science > QH Natural history > QH301 Biology > Methods of research. Technique. Experimental biology > Data processing. Bioinformatics > Artificial intelligence Q Science > Q Science (General) > Self-organizing systems. Conscious automata > Artificial intelligence 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 Artificial Intelligence
Depositing User:	Tamara Malone
Date Deposited:	07 Apr 2025 10:15
Last Modified:	07 Apr 2025 10:15
URI:	https://norma.ncirl.ie/id/eprint/7373

Actions (login required)

View Item