[en] Due to the promise of deterministic Ethernet networking, Time Sensitive Network (TSN) standards are gaining popularity in the vehicle on-board networks sector. Among these, Generalized Precision Time Protocol (gPTP) allows network devices to be synchronized with a greater degree of precision than other synchronization protocols, such as Network Time Protocol (NTP). However, gPTP was developed without security measures, making it susceptible to a variety of attacks. Adding security controls is the initial step in securing the protocol. However, due to current gPTP design limitations, this countermeasure is insufficient to protect against all types of threats. In this paper, we present a novel supervised Machine Learning (ML)-based pipeline for the detection of high-risk rogue master attacks.