Bordeanu, Octavian Ciprian; (2024) From Data to Insights: Unraveling Spatio-Temporal Patterns of Cybercrime using NLP and Deep Learning. Doctoral thesis (Ph.D), UCL (University College London).

Preview

Text Octavian_Bordeanu_thesis.pdf - Accepted Version Download (17MB) | Preview

Abstract

Measuring and mapping cybercrime is crucial for informing crime reduction initiatives, identifying gaps in strategies, and guiding further research. This thesis aims to develop and test approaches to detect, prevent, and understand spatio-temporal patterns of cybercrime, specifically focusing on malicious download events and IoT botnet attacks. By leveraging natural language processing (NLP) methods, such as word embeddings, and deep learning techniques for time series analysis, like long short-term memory networks, alongside principles from financial trading technical analysis, I aim to provide insights into the detection and prevention of cybercrime. Defining and understanding cyberspace is a challenging aspect of studying cybercrime. While cybercrime classification frameworks are well-established and applicable in the long term, establishing a unified definition and taxonomy for cyberspace itself is difficult due to its volatile nature and the multitude of possible frameworks. To address this challenge, I adopt a bottom-up design approach, defining cyberspace based on the type of cybercrime being investigated, data availability, research objectives, and my own creativity. This approach allows for a more practical and contextual understanding of cyberspace within the specific domain of cybercrime. The thesis structure includes a comprehensive literature review that explores common behavioral features of physical and online crimes and highlights the limitations of a single cyberspace taxonomy. Additionally, specific case studies and methodologies are presented, including the analysis of malicious download events using word embeddings and the exploration of evolving patterns of the Mirai botnet using deep learning techniques. Overall, this thesis contributes to the field of cybercrime research by leveraging advanced techniques to understand and combat cybercrime, while acknowledging the challenges in defining cyberspace itself. The findings and methodologies presented aim to inform crime reduction strategies and stimulate further research in this evolving field.

Download activity - last month

Export as XMLCSVJSON

Download activity - last 12 months

Export as XMLCSVJSON

Downloads by country - last 12 months

Export as JSONXMLCSV

Archive Staff Only

View Item