A REVIEW OF THE APPLICATIONS OF MACHINE LEARNING IN CYBERSECURITY AND ITS CHALLENGES
Aishwarya Ulhas Desai · Journal of Digital Security and Forensics · 2024
Our study aims to identify use of Machine Learning (ML) to solve cybersecurity problems and its challenges. ML algorithms are used for malicious network traffic detection, phishing Universal Resource Locator (URLs) detection, malware analysis, and more. Based on the review of selected articles, it is inferred that high-quality data sets, efficient feature extraction, proper training, and testing of ML mode can deliver accurate detection model. Even though the integration of ML in cybersecurity can improve threat detection capabilities, there are some challenges, including limited availability of real-world cyber datasets to train the ML models, the vulnerability of the ML model to data poisoning, imbalanced data sets causing biased detection, and data protection laws making hard to collect necessary data to train and test the ML model efficiently.