- College Credit
- Beginner
About this Course
In this chapter, machine learning techniques are applied to the cybersecurity domain with a focus on ransomware detection and classification, where a multi-tiered streaming analytics model leverages 24 static and dynamic traits to distinguish between various ransomware families and versions. Experimental evaluations demonstrate that the proposed hybrid machine learner outperforms state-of-the-art methods in accuracy, speed, and resource efficiency, addressing critical challenges such as family attribution, multi-descent fusion, and imbalanced datasets.
Machine Learning in Cybersecurity
In this stage, the chapter introduces the use of machine learning for ransomware analysis by outlining the evolution of ransomware families, the "recipe-to-success" attack strategy, and the need for robust detection models that can adapt to rapidly evolving threats.
4 steps18.4. Experiments and Results
In this stage, the chapter presents a comprehensive evaluation of the proposed ransomware streaming analytics model and hybrid machine learner (HML) using both controlled comparative experiments and a month-long realistic test. These experiments demonstrate that the proposed solution outperforms state-of-the-art machine learners and anti-ransomware tools in detecting and classifying diverse and evolving ransomware threats, achieving high accuracy, efficient throughput, and robust performance on imbalanced real-world data.
8 steps