Non-Negative Matrix Factorisation for Feature Selection: A Proposed Approach for the Detection of Multi-Stage Attacks
Moemedi Lefoane, Ibrahim Ghafir, Sohag Kabir, Irfan‐Ullah Awan · 2024
With the emergence of digital technologies like 5G wireless networks, cloud computing, and the Internet of Things (IoT), our daily lives, travel, and work have undergone a transformation. These advancements have led to improved productivity, informed decision-making, and increased profits. However, adversaries have also found lucrative opportunities to launch attacks, which have become more sophisticated and stealthier, making them challenging to detect. Multi-Stage attacks (MSAs), in particular, have gained popularity due to their stealthy nature and the success they have achieved in recent years. To combat these attacks, this paper utilised an optimised Non-Negative Matrix Factorisation (NMF) for feature selection, as part of the Machine Learning (ML) approach to enhance the detection of MSAs.