Graduate Theses and Dissertations (2019 - present)

Date of Award

12-2025

Document Type

Thesis

Degree Name

M.S.

Department

Computer and Information Science

Committee Chair

Ryan, G. Benton, Ph.D.

Abstract

Non-linear phase space analysis may be used to represent time-series data as graph data with transitions between states in the time domain. By studying these transitions, we can predict anomalies within the system. Previous research has demonstrated success in learning from phase graphs for malware and seizure detection. These solutions either require extracting global features or converting the graph into an image for convolutional neural networks (CNNs), which adds a layer of complexity and limits the size and potential expressiveness of a graph. To sidestep current limitations, this study proposed Graph Neural Networks (GNNs) for analyzing phase graphs. GNNs do not limit graph complexity, nor require the upfront calculation of global or local features. This study utilizes this approach on two cybersecurity datasets: the well-known Canadian Institute for Cybersecurity Intrusion Detection System (CICIDS) 2017 dataset of network activity and a power usage dataset for rootkit detection. Findings have revealed GNNs can be used successfully with phase space graphs, that the type of GNN does impact the classification accuracy, and that variance in parameters for the phase space graph also may impact the classification accuracy.

Download

Share

COinS