Harnessing Big Data Analytics for Advanced Detection of Deepfakes and Cybersecurity Threats Across Industries
Rasheed Afolabi, Rianat Abbas, Rajesh Vayyala, Dorcas Folasade Oyebode, Victoria Abosede Ogunsanya, Adetomiwa Adesokan · International Journal of Scientific and Management Research · 2025
The rise of deepfake technology has introduced a new layer of complexity to cybersecurity, creating opportunities for misuse in areas like misinformation, fraud, and identity theft. These challenges are further amplified by the speed at which deepfakes and other cyber threats evolve, often outpacing traditional detection methods. This study delves into how big data analytics can be harnessed to combat these threats, using advanced machine learning models like gradient boosting to detect malicious patterns in large-scale datasets. Key insights reveal that features such as packet length and flow timing are critical in differentiating between web-based attacks and botnet activities. The model demonstrates strong performance, achieving a high AUC-ROC score of 0.97, showcasing its ability to identify and classify threats effectively. However, the work also highlights challenges, including the need for more computational efficiency, diverse datasets, and adaptability to rapidly changing attack methods. Despite these hurdles, the integration of big data analytics into cybersecurity frameworks shows immense promise, providing scalable and real-time solutions across industries. Moving forward, collaboration across fields and a focus on ethical data practices will be vital to ensuring these technologies are both effective and trustworthy in the fight against emerging cyber risks.