Aiming at the problems of scattered data, difficulty in tracing risk sources and lagging prevention and control of aviation safety incidents, this paper designs a big data traceability and risk prevention and control system for aviation safety incidents. The system integrates flight operation parameters, safety incident reports, maintenance records, meteorological information, airport operation logs and air traffic control interaction data. Through data cleaning, time alignment, semantic coding and correlation graph modeling, it realizes event element extraction, risk chain tracking and level assessment. The experimental results show that the event recognition Accuracy of the proposed method reaches 95.1%, Macro-F1 reaches 94.6%, AUC reaches 97.1%, traceability hit rate reaches 92.6%, Top-3 risk source coverage reaches 96.4%, and risk level consistency rate reaches 93.1%. The end-to-end response time is 154 ms. The research shows that the system can provide data support for active early warning, cause review and hierarchical prevention and control of aviation safety events.
