A Survey on Machine Learning based Advanced Persistent Threat Detection Techniques

Cyber-attacks, fog computing, computer security, and the Internet of Things have all grown significantly and rapidly in the last few years. External threats have the potential to jeopardize information's privacy, reliability, and accessibility (CIA). When an adversary modifies a device's behaviour, applications, and services, and hides the malicious code for a long time while monitoring the target's actions, the outcome is an indication of behavioral deviation from a predefined baseline, especially when manipulating smartphones. Advanced Persistent Threats (APTs) are sophisticated, targeted cyber attacks in which an unauthorized user enters a network and stays hidden for a long time. In this Literature Review (LR) paper, we provide an analysis of Advanced Persistent Threats (APT), including their global market impact, threat models, existing defence mechanisms, research gaps, methodology, and future directions. This study concludes by summarizing numerous cybersecurity initiatives that have been developed to detect APTs or related operations, divided into groups based on the detection strategy applied.