The method relies on State of Polarization (SOP) monitoring via digital signal processing in a coherent receiver paired with machine learning for the event classification and enables a proactive detection of a fiber cut. Radiation absorption excites an orbital electron to a higher energy level. Heating the material enables the trapped states to interact with phonons and decay into lower-energy. It is used to certify the performance of new fiber links and monitor the status of existing ones, detecting and locating fault events with advantages including simple operation, rapid response, and cost-effectiveness. However, like any other technology, fiber. We propose a data driven approach for the anomaly detection and faults identification in optical networks to diagnose physical attacks such as fiber breaks and optical tapping.