Computer Science MRP

Permanent URI for this collectionhttps://hdl.handle.net/10464/14572

Students currently enrolled in the Computer Science graduate program here at Brock University will be required to submit an electronic copy of their final Major Research Paper to this repository as part of graduation requirements. Instructions on how to do this can be found online

Once your MRP has been accepted in the Repository you will receive an email confirmation along with a link to your work

Browse

Recent Submissions

Now showing 1 - 4 of 4
  • Thumbnail Image
    ItemEmbargo
    Edge Communication Efficiency with GNN in the Internet of Vehicles
    Graham, Jessica L.
    Vehicular edge plays a central role in ensuring an effective allocation of resources to provide services and applications. Resource allocation and communication in dynamic vehicular environments face numerous challenges in efficiently managing resources and data sharing, specifically managing the intricate balance of connectivity, storage, energy, computing, and cost of resources. These challenges are also affected by mobility, resulting in the demand for precision in communication range, density, and resource availability. Efficient resource allocation is a critical objective within vehicular networks, and to achieve this, intelligence, prediction, optimization, and incentive modelling are often employed. However, challenges persist, such as sporadic connectivity, transmission delays, and the inherent uncertainty of highly dynamic environments. In response to these challenges, this paper introduces the use of graph neural networks (GNNs) to learn hidden spatial and functional patterns in complex vehicular networks. Combining with clustering-based methodologies. This approach enables the intelligent organization of network nodes, reducing transmission delays and enhancing connectivity in dynamic environments. The resulting framework supports predictions and estimates based on evolving communication and mobility patterns. They are further improving the efficiency of connectivity and communications in vehicular edge networks. Using graph neural networks (GNN) and clustering techniques to address connectivity challenges, reduce transmission latency, and manage the inherent unpredictability of rapidly changing vehicular settings, this study is poised to enhance the delivery of services and applications in vehicular networks. It also lays the foundation for prospective research into resource management.
  • Thumbnail Image
    ItemOpen Access
    A Novel DDoS Detection and Multi-Class Classification Method: A Graph Convolutional Network Approach
    Saunders, Braden
    Distributed Denial of Service (DDoS) is an attack that overwhelms the cyber critical infrastructure system with malicious packets causing it to become unresponsive, which precludes legitimate users from accessing the target system. This work leverages a deep learning method known as Graph Convolutional Network (GCN) to empower DDoS detection systems. The proposed GCN model consists of three hidden layers, each with 128 neurons. Considering the Canadian Institute for Cybersecurity CIC-IDS 2017 dataset, the proposed model achieves an overall accuracy of 99.95%, along with a value of 99.95% for each of the precision, recall, and F1-score metrics for the binary DDoS classification problem. For the multi-class DDoS classification problem, the model scores an overall accuracy of 98.94% and precision, recall, and F1-score values of over 93% for all classes. These results support the use of the proposed GCN DDoS detection method in practice.
  • Thumbnail Image
    ItemOpen Access
    ACS-IoT: A CNN-BiLSTM Model for Anomaly Classification in IoT Networks
    GUAN, YUE
    This work proposes an Anomaly Classification System for IoT (ACS-IoT). The proposed system contains a pipeline of machine learning and deep learning algorithms for the effective classification of anomalies and their sub-types. Machine learning algorithms are adopted to distinguish between normal data and anomaly data. The deep networks, on the other hand, are used to perform anomaly-type classification. We propose the use of the Synthetic Minority Oversampling Technique (SMOTE) to address the data imbalance problem and Particle Swarm Optimization (PSO) as a feature selection mechanism to improve accuracy as well as execution time. The proposed system proved to be accurate as well as precise when tested on a publicly available IoT dataset.
  • Thumbnail Image
    ItemOpen Access
    An Improved Sufficient Condition for Routing on the Hypercube with Blocking Nodes
    Wang, Wenjie
    We study the problem of routing between two nodes in a hypercube with blocking nodes using shortest path. This problem has been previously studied by other researchers, they have proposed a few algorithms to solve the problem. Among the work done, one has found several sufficient conditions for such a path to exist. One such condition states that a shortest path between node 0^n and 1^n exists if the number of blocking nodes is less than n in an n-dimensional hypercube. We improve this condition by proposing the condition that if the size of a SDR (system of distinct representatives) for the blocking nodes is less than n, then a shortest path between the two nodes 0^n and 1^n exists. Since the number of blocking nodes can be greater than or equal to n, while the size of SDR is less than n, thus this result improves the previous sufficient condition.