The increased integration of renewable energy sources and power electronic converters has transformed conventional power systems into hybrid AC-DC systems, introducing faster and more complex fault dynamics that challenge existing protection schemes. While data-driven approaches have demonstrated high accuracy in fault detection and classification, their lack of prediction validation limits their reliability in critical protection systems. This thesis proposes a real-time data-driven framework for event classification and validation for Hybrid AC-DC systems. A comprehensive dataset of AC and DC faults was generated from an RTDS based grid model, with transient measurements acquired using a Python-based communication interface and processed to extract time domain and dynamic fault features. Independent machine learning classifiers were developed for the AC and DC subsystems. An additional waveform signature validation layer that uses representative transient signatures of event classes and similarity analysis was introduced to verify classifier predictions. Performance evaluation demonstrated that the proposed framework achieves accurate and robust event classification under varying operating conditions and moderate signal degradation, while providing additional confidence assessment for machine learning-based classifier predictions. The proposed approach offers a practical and reliable solution for real-time monitoring, event identification, and supervisory protection in future converter-dominated power systems. ... Read more